#bioc_git
2017-08-12
Sean Davis (15:14:49): > @Sean Davis has joined the channel
Sean Davis (15:14:49): > set the channel description: Discuss care and feeding of git in the context of Bioc
Nitesh Turaga (15:17:45): > @Nitesh Turaga has joined the channel
Kevin Rue-Albrecht (15:18:01): > @Kevin Rue-Albrecht has joined the channel
Aaron Lun (15:18:08): > @Aaron Lun has joined the channel
Levi Waldron (15:18:16): > @Levi Waldron has joined the channel
Kasper D. Hansen (15:18:22): > @Kasper D. Hansen has joined the channel
Davide Risso (15:26:55): > @Davide Risso has joined the channel
2017-08-13
Peter Hickey (09:20:22): > @Peter Hickey has joined the channel
2017-08-14
Mike Smith (05:01:53): > @Mike Smith has joined the channel
Lorena Pantano (10:06:13): > @Lorena Pantano has joined the channel
Leonardo Collado Torres (11:16:55): > @Leonardo Collado Torres has joined the channel
Martin Aryee (13:55:45): > @Martin Aryee has joined the channel
2017-08-15
Nitesh Turaga (11:53:47): > <!channel>We are on schedule. The experiment-data package SVN commits have been disabled. The git transition for these packages is underway now with the latest snapshot.
Kevin Rue-Albrecht (12:08:44): > git cherry-pick bioc-devel-email?:stuck_out_tongue_winking_eye:
Nitesh Turaga (12:09:12): > Ha!
2017-08-16
Nitesh Turaga (09:15:37): > <!channel>It’s August 16th 9am. The SVN commits to the software packages are blocked. There will NOT be any SVN commits now onwards to the software and data-experiment packages. > > NOTE: Only the workflow packages are still on SVN until further notice about their transition.
Nitesh Turaga (17:02:42): > It’s ALIVE!
Nitesh Turaga (17:02:51): > The git server is alive!!
Sean Davis (17:03:19): > AWESOME work! I felt the Earth shake in Bethesda.
Ju Yeong Kim (18:54:39): > @Ju Yeong Kim has joined the channel
2017-08-17
Martin Morgan (08:53:50): > @Martin Morgan has joined the channel
Martin Morgan (08:54:05): > A couple of things from the#singlecellexperimentchannel
Martin Morgan (08:54:42): > An alternative to archiving a previous, very divergent ‘old’ repository is the old switcheroo (rename branches) described athttp://bioconductor.org/developers/how-to/git/abandon-changes/
Martin Morgan (08:56:50): > The motivation for the phrase@Kevin Rue-Albrechtquotes about all authors having write access is meant to convey that the authors are in this together, and that Bioc will not play moderator between disagreements – the authors will have to work things out between themselves, we will not refuse an author the ability to commit to a repository that they are on. This has consequences for who becomes a package author.
Martin Morgan (08:58:10): > I would recommend that the maintainer establishes a public git(hub) repo that mirrors bioc, and that collaborates contribute to that single repo via pull requests; it will be confusing to both users and developers to maintain multiple public repos that are not clearly marked as ‘forks’. A looser model is for the maintainer to invite collaborators; this invites conflicts at the public repo stage, and this will be problematic / frustrating when packages are under active development.
Kasper D. Hansen (08:59:08): > I can see the sense in avoiding being a moderator in feuds. In can also see the sense in potentially having multiple people with commit right. But let me point out that is is yet another interpretation of the author field in the DESCRIPTION file (author = commit rights), which different people already interpret differently
Martin Morgan (09:03:19): > Yes, Bioconductor interprets it in the manner described in Kevin’s quote:slightly_smiling_face:. I think of these as ‘living’ scientific publications, which requiring ongoing consent of all authors with respect to content. I think Bioconductor is more closely tied to the academic endeavor than a typical CRAN package, and so the interpretation of authorship is made in that context.
Martin Morgan (09:04:20): > Practically, disagreements occur very rarely (a couple of instances that I remember).
Kasper D. Hansen (09:23:07): > I prefer the Bioconductor interpretation. Although I am sure that is still interpreted differently between groups.
Nitesh Turaga (10:16:21): > @Davide RissoReplying to your comment on#singlecellexperiment, Nope, unfortunately there isn’t. What would you like to see there?https://git.bioconductor.org/Do you mean like a git-web?
Nitesh Turaga (10:16:45): > <thread: is there a way to “see” the git server… something like the old https://hedgehog.fhcrc.org/bioconductor/trunk/madman/Rpacks>
Davide Risso (10:17:19): > I just used to go tohttps://hedgehog.fhcrc.org/bioconductor/trunk/madman/Rpacksto make sure my push was reflected
Davide Risso (10:17:55): > I guess there is the github mirror but that’s not updated immediately
Kasper D. Hansen (10:17:57): > Might be worth thinking about@Nitesh Turagabut perhaps one should start by figuring out demand
Nitesh Turaga (10:18:37): > @Davide Rissothe github-mirror doesn’t do anything anymore, as an FYI. It’ll be stale. Specificallyhttps://github.com/Bioconductor-mirror
Davide Risso (10:19:05): > oh ok
Peter Hickey (10:19:21): > i would certainly appreciate something like a web viewable thing. i used the bioc github mirrorheaps
Aaron Lun (10:19:33): > I will say that I often enjoyed staring at other people’s code on Bioc mirror. Or even my own.
Kasper D. Hansen (10:20:09): > I would suggest aggressively removing stuff that is not intended to be used in future
Kasper D. Hansen (10:20:13): > like the github mirror
Leonardo Collado Torres (10:20:13): > I also looked at the Bioc mirror, specially when I didn’t understand the documentation of a function and wanted to figure out some details
Davide Risso (10:20:23): > I thought that the github mirror could be a read-only mirror of Bioconductor, much like metacran. It’s useful to be able to look up the code of all bioc packages directly in one place
Kasper D. Hansen (10:21:21): > Hmm, I think for looking at code we should rather think about also linking to “official” devel repositories with issues etc.
Nitesh Turaga (10:22:02): > Unfortunately, people don’t use the bioconductor-mirror with that sense, they fork and develop, and with this new git transition, it’ll cause a lot of problems when merging i.eunrelated-histories
Nitesh Turaga (10:22:27): > @Kasper D. HansenThere is a move to get everything “core” team maintained into github in the next few days.
Aaron Lun (10:22:50): > Any searchable code base on a browser would be satisfactory, avoid us having to clone the repository every time we want to inspect the code.
Nitesh Turaga (10:23:18): > if that’s what you mean. The core -team will also be doing the same workflows as maintainers, develop github –> sync with bioc-git, use issues + all other social coding attributes on github.
Davide Risso (10:25:20): > Another suggestion would be to add a more obvious link to the “official” devel repo in the package landing pages
Nitesh Turaga (10:25:59): > Give me an example.
Davide Risso (10:26:21): > There is already a “git source” link on the landing page of each package
Davide Risso (10:26:51): > but it’s at the bottom of the page and if you don’t know that it’s there you won’t see it
Davide Risso (10:27:15): > moving that link towards the top (perhaps with the badges)
Davide Risso (10:28:09): > (assuming that every developer has a github repo which might not be the case since many of those links point to the mirror now)
Davide Risso (10:29:54): > or I guess all of them link to the mirror?
Leonardo Collado Torres (10:30:03): > @Leonardo Collado Torresuploaded a file:Screen Shot 2017-08-17 at 10.27.39 AM.pngand commented: You and Kasper are saying that “git source” should point to the developer’s github page. In my case I have that under “URL” right now. But just making things uniform would be nice. > > An added advantage of the Bioc-mirror was that I could always “type” the url without finding it elsewhere. But given that you want to deprecate the mirror, I agree with Davide and Kasper that having a standard place to look for the git code would be nice. - File (PNG): Screen Shot 2017-08-17 at 10.27.39 AM.png
Marcel Ramos Pérez (10:31:16): > @Marcel Ramos Pérez has joined the channel
Kasper D. Hansen (10:31:25): > Is search facility across packages important?
Leonardo Collado Torres (10:39:57): > I haven’t had to do that. Though I don’t know what Herve does when he fixes tons of packages that depend on his code (like with the IRanges to S4Vectors changes from a while ago). Well, maybe he looks at the “reserve depends” page on his package. > > In my case, once I find the github source page of a package that I’m looking into, I use the “search this repository” GitHub functionality the most. I might find that a package uses code from another one, so then I go and search that other one. > > For example, I was looking at some args inclusterProfiler::enrichGOwhich lead me tohttps://github.com/Bioconductor-mirror/clusterProfiler/blob/2ca4d55e0234605830aa5781cfee7b0cb725b11f/R/utilities.R#L129that readsenricher_internal <- DOSE:::enricher_internal. With the mirror, all it took me was typing the url for DOSE’s github code, then using “search this repo” to findhttps://github.com/Bioconductor-mirror/DOSE/blob/master/R/enricher_internal.R. If the github repo of the package is visible atbioconductor.org/packages/pkgNamethen it’ll be almost the same process. - Attachment (GitHub): Bioconductor-mirror/clusterProfiler > This is a read-only mirror of the Bioconductor SVN repository. Package Homepage: http://bioconductor.org/packages/devel/bioc/html/clusterProfiler.html Contributions: https://guangchuangyu.github.io… - Attachment (GitHub): Bioconductor-mirror/DOSE > This is a read-only mirror of the Bioconductor SVN repository. Package Homepage: http://bioconductor.org/packages/devel/bioc/html/DOSE.html Contributions: https://guangchuangyu.github.io/DOSE. Bug …
> ./DESeq/vignettes/DESeq.Rnw:\date{\Rpackage{DESeq} version \Sexpr{packageDescription("DESeq")$Version} (Last revision \StrMid{$Date: 2016-01-12 19:18:09 -0500 (Tue, 12 Jan 2016) $}{8}{18})} > ./DEXSeq/vignettes/DEXSeq.Rnw:\date{Last revision of this document: \StrMid{$Date: 2016-11-29 09:04:25 -0500 (Tue, 29 Nov 2016) $}{8}{18}} > ./genefilter/vignettes/independent_filtering_plots.Rnw:\date{\Rpackage{genefilter} version \Sexpr{packageDescription("genefilter")$Version} (Last revision \StrMid{$Date: 2016-03-22 18:37:15 -0400 (Tue, 22 Mar 2016) $}{8}{18})} > ./genefilter/vignettes/independent_filtering.Rnw:\date{\Rpackage{genefilter} version \Sexpr{packageDescription("genefilter")$Version} (Last revision \StrMid{$Date: 2016-03-22 18:37:15 -0400 (Tue, 22 Mar 2016) $}{8}{18})} > ./mosaics/vignettes/mosaics-example.Rnw:%%\markboth{$Date$\hfil$Revision$}{\thepage} > ./DBChIP/vignettes/DBChIP.Rnw:%%\markboth{$Date$\hfil$Revision$}{\thepage} > > git issues threadhttps://github.com/Bioconductor/bioc_git_transition/issues/32 - Attachment (GitHub): packages using svn properties to stamp date in vignettes · Issue #32 · Bioconductor/bioc_git_transition > ./DESeq/vignettes/DESeq.Rnw:
./DEXS…
Kasper D. Hansen (10:57:16): > @Nitesh Turagaperhaps send that repos with issues to email list
Nitesh Turaga (10:57:36): > Yep, doing it now. That was just a notification for people on this channel.
Lorena Pantano (12:24:35): > stupid question, now the git.bioconductor will sync with the github mirror or not anymore?
Kevin Rue-Albrecht (12:27:32): > @Lorena Pantanohere’s the plan for the github-mirror - Attachment: Attachment > @Davide Risso the github-mirror doesn’t do anything anymore, as an FYI. It’ll be stale. Specifically https://github.com/Bioconductor-mirror
Lorena Pantano (12:29:07): > thanks!
Aaron Lun (14:03:23): > A bit off-topic; but when I commit a version bump tobeachmat, will the build system automatically flag all downstream packages (i.e., depending onbeachmat) for rebuilding? Otherwise any modifications tobeachmatC++ code will result in segmentation faults in downstream packages trying to link to old versions of the shared library.
Aaron Lun (14:03:59): > More pernicious is the fact thatbiocLitemight pull down updated versions ofbeachmat, but it won’t rebuild the local (i.e., on the user’s computer) versions of downstream packages, resulting in the aforementioned problem.
Kasper D. Hansen (14:05:36): > It is my understanding that everything gets rebuild in the daily build on devel. Not sure what happens with dependences if you update release
Kasper D. Hansen (14:06:32): > The second issue might happen in the wild, definately. I am not sure whether this is handled automatically; I don’t think so
Aaron Lun (14:12:25): > I vaguely recall someone (maybe Martin?) bumping the version numbers of packages dependent onRhtslibwheneverRhtslibgot updated. So maybe we’ll have to do something like that as well forbeachmatandRhdf5lib. Perhaps they could set up an automated system on the git server to do this.
2017-08-18
Martin Morgan (05:10:06): > @Aaron Lunif you introduce an API change, in R or in C / C++, then already installed downstream packages generally break on user computers. The build system builds nightly from scratch, so packages available via biocLite() are not affected (well, not quite, there is a chicken and egg situation where if beachmat installs but does not check, then the downstream packages are shipped based on the new beachmat but the new beachmat itself is not available). It’s generally the developer responsibility to identify downstream packages, since ‘we’ generally have no way of knowing whether a commit breaks the API.
Mike Smith (05:56:39): > I’ve been thinking about tools for searching across packages too. I quite often work in the manner@Leonardo Collado Torresdescribed, using the github interface to find and view how other people have implemented things. I always thought this was a massive improvement over checking out the entire SVN and grepping which is what I did before. I’m not sure if there are any existing tools that aren’t also overkill, but I’ve put together a proof of principle implementation of Hound (https://github.com/etsy/hound) that indexes the first 50 (alphabetically) BioC packages and lets you search the code. Not sure it totally does what I’m looking for, but maybe it’s something to foster a discussion. You can give it a try athttps://bioc-git-search.sloppy.zone
Aaron Lun (10:46:54) (in thread): > While I’m not an expert on this, one of the problems specific to C++ is its use of name mangling. From what I understand, the exact mangling is context dependent, so the final mangled name in the shared library will depend on the names of the classes and of the overloaded function, the number and arguments, etc. This means that - even if the API does not change - the mangled name might change in thebeachmatshared library, resulting in package libraries trying to link to a mangled name that no longer exists inbeachmat. > > The solution is to force a recompilation of the package library on user computers, and this would require a version bump to the git repositories. I’m just wondering whether it’s worth putting in an automated system to bump the version numbers of downstream packages whenbeachmat’s C++ code changes. The alternative would be for users to see some very obscure and interesting errors , e.g., memory not allocated, nil addresses.
Aedin Culhane (10:49:15): > @Aedin Culhane has joined the channel
Aaron Lun (10:54:00) (in thread): > I say this because I personally have run afoul of this problem, and have been confused more than once. (“Huh? Why isscranbreaking? Why isscaterbreaking? What’s going on?”) And I should know better, because I manage the C++ code for all of them.
Mike Smith (11:13:44) (in thread): > Maybe I’m missing something, but isn’t this one of the reasons why we have the concept of Release and Devel? If I choose to use developmental packages, then I just accept that there will be unexpected errors, and I might need to reinstall things.
Aaron Lun (12:08:27) (in thread): > That’s fair enough, though there may be many things that need re-installing if - as we hope - many packages end up linking to our libraries. We can pre-empt the confusion and extra work to some extent with an automated version bump; it could be argued that’s too heavy a touch, to which I am sympathetic. > > Bug fixes to the release version of the API, though, should probably trigger a bump to all downstream packages.
2017-08-19
Martin Morgan (08:03:57): > A couple of minor updates on the transition
Martin Morgan (08:04:42): > Fromhttps://stat.ethz.ch/pipermail/bioc-devel/2017-August/011340.htmla few repos will be regenerated, invalidating any clones.
Martin Morgan (08:06:59): > The build system was sometimes using older commits to build packages (the ‘Last Commit’ number on a package landing page likehttp://bioconductor.org/checkResults/3.6/bioc-LATEST/CEMiTool/malbec1-buildsrc.htmlis reported correctly, but is not the last commit before the ‘Snapshot Date’). This should be corrected in the builds that are produced today. The older builds would not propagate, so no harm to the end user.
2017-08-21
Kasper D. Hansen (11:59:53): > I think@Nitesh Turagajust said on bioc-devel that only maintainers will have git access. Seems like a policy (whatever it is) should be adapted
Martin Morgan (12:36:40): > Please seehttps://stat.ethz.ch/pipermail/bioc-devel/2017-August/011354.htmlfor plans to regenerate at least some repositories
Aaron Lun (17:06:01): > I didn’t noticebeachmatandSingleCellExperimenton that list of non-regenerated repositories; both of these packages should already have committed changes, and there are no unknown users. Similarly,scranalso has no unknown users. Your follow-up email indicates that repositories with no unknown users will not be regenerated; is this the case?
2017-08-22
Aaron Lun (13:37:31): > @Nitesh TuragaI don’t know if you’ve already mentioned this in the various BioC-devel threads, but if I update my Github SSH keys, will the updated keys automatically be respected bygit.bioconductor.org? I’m going on holiday in a couple of weeks and I’ll be using several different computers.
Nitesh Turaga (14:06:26): > @Aaron LunYou’ll have to wait 24 hours after updating your keys on github.
Nitesh Turaga (14:06:35): > It runs in a 24 hour cycle.
Nitesh Turaga (14:06:56): > (depending on when you update. It runs at noon EST)
Kasper D. Hansen (15:13:44): > @Nitesh Turagahow about multiple keys - it respects all of them?
Nitesh Turaga (17:11:14): > yes@Kasper D. Hansenif you have multiple ssh keys on your github it should work for all of them.
2017-08-23
Aaron Lun (04:23:22): > Thanks@Nitesh Turaga.
2017-08-25
Lorena Pantano (09:51:13): > @Martin Morgan, can I ask if the read-only git repository (httpsprotocol) is updated at the same time I push to thegitprotocol. These to repos for my isomiRs package are not pointing to the same commit.
Lorena Pantano (09:52:33): > I did a commit 24 hours ago, to bump the version to try to fix the build that is failing right now
Kevin Rue-Albrecht (09:52:33): > I don’t think they update the github-mirror anymore. They’ve mentioned that it will be deprecated and removed at some point.
Lorena Pantano (09:52:46): > I meant thehttps://git.bioconductor.org/packages/isomiRs
Lorena Pantano (09:53:11): > when I clone that one, is one commit behind mygit@git.bioconductor.org:packages/isomiRs.git
Kevin Rue-Albrecht (09:54:11): > Oops, that part I don’t know yet.
Kevin Rue-Albrecht (10:03:39): > I just checked my last commit since the transition (18th August) and cloning from both git and https gives me the same log/history. Might be something related to the ‘regeneration’ of packages 4 days ago?
Kevin Rue-Albrecht (10:04:47): > (isomiRs was one of the packages in the listnotregenerated - not sure if that’s relevant)
Nitesh Turaga (10:05:35): > @Lorena PantanoI’ll inspect real quick and get back to you.
Nitesh Turaga (10:05:51): > The access protocol shouldn’t have anything to with how the repo is updated.
Lorena Pantano (10:06:22): > thanks@Nitesh Turaga, I wasn’t sure if you have two repos, or only one
Nitesh Turaga (10:06:28): > just 1
Lorena Pantano (10:06:50): > I just did pull , and I got it
Lorena Pantano (10:06:55): > did you do something?
Nitesh Turaga (10:06:57): > Nope.
Lorena Pantano (10:07:16): > lol, I was doing pull all the time since yesterday and nothing…not sure what is going on
Nitesh Turaga (10:07:20): > It is impossible that you have two different versions, its just one repo being served .
Nitesh Turaga (10:07:49): > :smile:Glad you have it resolved.
Lorena Pantano (10:07:59): > I should have done screenshot. I know sounds crazy.
Nitesh Turaga (10:08:19): > Haha, next time it happens, be sure to.
Lorena Pantano (10:08:37): > I’ll put timestamp in my prompt:slightly_smiling_face:
Nitesh Turaga (10:08:51): > :slightly_smiling_face:
Nitesh Turaga (10:09:31) (in thread): > @Kevin Rue-AlbrechtNot relevant to this problem i’d say.
Nitesh Turaga (10:09:52) (in thread): > Thanks for the quick check and response.:smile:
Kevin Rue-Albrecht (10:10:32) (in thread): > no worries - thanks for catching up my misleading comment
Kevin Rue-Albrecht (10:11:54) (in thread): > i actually thought that git and https are just protocol for a single repo, but I immediately looked for convoluted possibilities because I was not expecting such a simple conclusion:stuck_out_tongue_closed_eyes:
Nitesh Turaga (10:14:34) (in thread): > They are just protocols:smile:
Kevin Rue-Albrecht (10:17:39): > While I’m thinking git and Bioconductor, should some general guidance be written out about merging or rebasing local master branches on the upstream(bioc)/master when the release/version bump will happen in October ?
Kevin Rue-Albrecht (10:26:49): > I feel like this should be plain straightforward, but somehow I still doubt exactly how to do that cleanly
Martin Morgan (10:29:33): > maybe your institution does a local cache of https requests, so you’re getting (their) cache?
Lorena Pantano (12:29:37): > I could look at that, but that was at home and work.
Lorena Pantano (12:30:03): > although that would explain I am not crazy
2017-08-26
Aaron Lun (06:37:33): > @Nitesh TuragaCan I get write access to theSingleCellExperimentrepository?
Martin Morgan (09:15:42): > done
Aaron Lun (09:17:25): > Thanks
Aaron Lun (09:39:16): > How long does it usually take for the updated access rights to become active?
Martin Morgan (09:42:51): > If you submitted your public key previously, they should be now. Tryssh -vgit@git.bioconductor.orgto see repos you have access to, and also to see the negotiation. But oops, I committed but forgot to push the change,, so try again
Aaron Lun (09:45:48): > Works now, thanks.
Martin Morgan (09:46:15) (in thread): > if you’ve got github & bioc repos then it’shttp://bioconductor.org/developers/how-to/git/sync-existing-repositories/, from step 3 onward if already configured. If only a clone of git, then it’s just a git pull or git fetch + git merge
2017-08-30
Lorena Pantano (09:21:29): > a question about the test coverage badget, that points to bioconductor-mirror, is there any plan for this to migrate to the new git repo?
Martin Morgan (09:46:25): > yes, we’re aware of this and it’ll transition ‘eventually’; it’s a relatively low priority
Lorena Pantano (10:16:50): > thanks!
Sean Davis (10:52:42): > Random KUDOS. Thanks@Martin Morganand@Nitesh Turagafor the continued work on the git transition and for being so incredibly responsive.
2017-09-01
Kasper D. Hansen (15:17:38): > @Nitesh Turagaare you going to remove the github mirrors. Thats what I would do.
Nitesh Turaga (15:20:07): > @Kasper D. HansenWe will be soon, we just need to figure out a few issues with the code coverage. As soon as that is done, we’ll be removing it.
Kasper D. Hansen (15:24:41): > ok, just wanted to make sure it was on the radar.
Kasper D. Hansen (15:25:20): > Im pretty impressed at the transition so far, judging from reactions. Still have to try to make my own commit though:slightly_smiling_face:
Nitesh Turaga (15:26:01): > Yep, on the radar:smile:I’m glad its gone as smoothly as it has actually.
2017-10-03
Lori Shepherd (10:44:04): > @Lori Shepherd has joined the channel
2017-10-06
David Jenkins (16:33:12): > @David Jenkins has joined the channel
2017-10-27
Stian Lågstad (13:35:51): > @Stian Lågstad has joined the channel
2017-10-31
Guangchuang Yu (22:40:36): > @Guangchuang Yu has joined the channel
Guangchuang Yu (22:44:04): > please add the write permission of seqcombo to g.yu (guangchuangyu@gmail.com).
2017-11-01
Martin Morgan (06:55:09): > @Guangchuang Yudone.
Guangchuang Yu (06:57:06): > still throw error: FATAL: W any packages/seqcombo g.yu DENIED by fallthru
Martin Morgan (06:58:17): > oops, forgot to push (stuck in svn land!) try again…
Guangchuang Yu (07:01:04): > thanks@Martin Morgan, it works fine now.
2017-11-28
Stephanie Hicks (14:19:27): > @Stephanie Hicks has joined the channel
2017-11-29
Matthew McCall (09:32:03): > @Matthew McCall has joined the channel
2018-01-03
Nitesh Turaga (09:33:42): > Hello Maintainers, > > Happy new year! > > The GIT logs are available as RSS feeds, or as markdown on the website athttp://bioconductor.org/developers/gitlog/This is across all the packages on the bioconductor git server. It allows developers to keep track of their package development, and for the community to keep track of their favorite packages. The website shows only the most recent 200 commits. > > Please note that the website rebuilds every 20 mins, so it will take 20 mins for your commit to show up on the website.
2018-01-19
Aaron Lun (11:43:26): > @Nitesh TuragaCan I get write access to thescatergit repo?@Davis McCarthycan authorize this if necessary.
Aaron Lun (11:44:12): > I mean push access.
Davis McCarthy (11:46:09): > @Davis McCarthy has joined the channel
Davis McCarthy (11:46:55): > Yes, I definitely authorize Aaron for push access to scater git repo
Nitesh Turaga (12:25:55): > Done.
Nitesh Turaga (12:28:13): > Both@Davis McCarthyand@Aaron Lunshould have push access to scater now. It’s up to you guys to update the DESCRIPTION file to reflect Maintainers / Authors / contributors.
Aaron Lun (12:43:06): > Thanks
Kasper D. Hansen (15:19:50): > In a few years I fear Aaron might be responsible for 87% of the packages in Bioc.
2018-01-20
Davis McCarthy (08:45:28): > That’s his masterplan, I’m sure!
2018-01-26
Keegan Korthauer (11:27:29): > @Keegan Korthauer has joined the channel
2018-02-12
Aaron Lun (10:14:54): > In lieu of a better channel for this question: Part of my localbeachmattesting involves checking that a small test package (inst/testpkg) can be INSTALLed and CHECKed. This is necessary to ensure that we can correctly link tobeachmatfrom another package. Is it possible to make the CHECK forinst/testpkgrun as part of the CHECK forbeachmatitself? Preferably without any side-effects, such as leavingtestpkgpermanently installed.
Sean Davis (10:18:41): > With no deep thought (so could be very off), it seems like you should be able to implement a test for beachmat that executes the install and check?
Sean Davis (10:19:51): > Or at least check.
Aaron Lun (10:19:55): > Unfortunately?checkdoesn’t provide much insight into how CHECK can be run from within R.
Sean Davis (10:21:35): > devtools::check?
Sean Davis (10:21:59): > Returns an object with errors, warnings, and notes, according to the docs.
Aaron Lun (10:22:07): > Oh sweet
Sean Davis (10:22:08): > Not ideal, perhaps.
Sean Davis (10:22:30): > https://www.rdocumentation.org/packages/devtools/versions/1.13.3/topics/check
Aaron Lun (11:09:00): > I remember@Martin Morgansending an email out about the differences between devtools and bioC check - anything that might derail the use of devtools::check here?
Aaron Lun (11:14:19): > There is also the interesting point of whether I can assume thatsystem.file("testpkg", package="beachmat")exists during CHECKing ofbeachmat. I would guess so, as WRE 1.3.1 states that the package is installed first thing in CHECK.
Sean Davis (11:48:34): > Ondevtools::check, there are differences. How those will impact the integrity of the checking, I do not know. I’m also not sure how to replicated BiocCheck in code without using something like asystem()call.
Aaron Lun (12:02:37): > I suspect that the finer points would not be relevant to my test package - I don’t even have documentation for the functions in there, as they would never be run by any user.
Martin Morgan (12:37:33): > https://github.com/Bioconductor/AnnotationDbi/commit/546eec8d909bdf5b181091607b3e70bbe964409eandhttps://github.com/Bioconductor/AnnotationDbi/commit/648d44c809ae228577f34598a88b29180594b24dshow one attempt at a similar activity – create a temporary directory for library installation, install into the temporary location, remove. - Attachment (GitHub): use temporary directory for org.testing.db, part II · Bioconductor/AnnotationDbi@546eec8 > AnnotationDbi - Annotation Database Interface - Attachment (GitHub): use temporary directory for org.testing.db · Bioconductor/AnnotationDbi@648d44c > AnnotationDbi - Annotation Database Interface
Aaron Lun (12:42:26): > Interesting. Though it seemsdevtools::checkmight be a bit easier, as I’ll have to CHECK anyway.
Aaron Lun (13:01:06): > Yep, looks like thesystem.filecall should be all right.
2018-02-23
Leonardo Collado Torres (13:17:14): > Hi! > > Is there a way to get Slack notifications whenever a specific package repo gets updated atgit.bioconductor.org? > > For example, I have Slack setup to show notifications in a channel wheneverhttps://github.com/lcolladotor/derfindergets updated. But a long time might pass before I realize thatgit@git.bioconductor.org:packages/derfinder.git was updated (say by the bioc-core team). While fixing a unit test forderfindertoday I noticed that I was missing some commits (thx again forhttps://bioconductor.org/developers/how-to/git/pull-upstream-changes/). This prompted me to run: > > git checkout master > git fetch upstream > git merge upstream/master > git push origin master > > in all my repos. > > Best, > Leo - Attachment (GitHub): lcolladotor/derfinder > derfinder - Annotation-agnostic differential expression analysis of RNA-seq data via expressed regions-level or single base-level approaches
Sean Davis (13:18:39): > In short, no.
Sean Davis (13:20:19): > There is an RSS feed of package commits, but it contains all commits to any package. That may help, but you’ll probably need to check that outside of slack.
2018-02-26
Nitesh Turaga (09:17:33): > Hi@Leonardo Collado Torres, Yes Sean is right. We don’t have a way to track a specific package through the Bioc-gir RSS feed. But, depending on the RSS feed reader you use, you can filter on the package name.
Leonardo Collado Torres (10:08:01): > Thanks Sean and Nitesh! Which rss feed reader(s) that supports this do you know?
Sean Davis (10:16:45): > RSS feeds are just XML. You can use the xml2 package to read and parse using a cron job. You’ll end up with a running record.
Nitesh Turaga (10:17:13): > I know@Martin Morganuses his thunderbird email as feed reader.
Nitesh Turaga (10:17:34): > Maybe he knows more about feed-readers than I do. He does the filtering within thunderbird
Leonardo Collado Torres (10:18:11): > ok, thanks again!
Sean Davis (10:20:12): > There are some really nice tools out there for working with feeds. See this, for example:https://ifttt.com/feed - Attachment (ifttt.com): IFTTT > Connect RSS Feed to Notifications, feed, IFTTT, Email, Pocket, Telegram, Gmail, Twitter, Email Digest and more.
Leonardo Collado Torres (11:04:04): > Thanks, I usedhttps://ifttt.com/applets/326276p-if-rss-filter-match-then-email - Attachment (ifttt.com): If RSS filter match, Then email > If rss feed entry contains some word (Keyword or simple phrase), then email it
2018-02-27
Aaron Lun (09:33:06): > @Nitesh TuragaAny news on the workflow git transition?
Nitesh Turaga (09:51:09): > @Aaron LunWe are working on getting the build system ready for them.
Aaron Lun (09:51:18): > Sweet, thankx.
2018-02-28
Daniel Van Twisk (15:40:08): > @Daniel Van Twisk has joined the channel
2018-03-16
Aaron Lun (17:11:34): > Probably not entirely the right place to ask this, but I recently committed a change to thebeachmatvignettes, splitting it up into three shorter vignettes namedlinking.Rmd,input.Rmdandoutput.Rmd. How is the order in which they show up determined on the BioC website? I would like linking to be first, followed by input and output; but it seems that the website shows linking > output > input, which doesn’t follow any obvious order.
2018-03-17
Martin Morgan (17:41:19): > I believe they are alphabetical by title; see beachmat and e.g.,http://bioconductor.org/packages/devel/bioc/html/GenomicRanges.html - Attachment (Bioconductor): GenomicRanges (development version) > The ability to efficiently represent and manipulate genomic annotations and alignments is playing a central role when it comes to analyzing high-throughput sequencing data (a.k.a. NGS data). The GenomicRanges package defines general purpose containers for storing and manipulating genomic intervals and variables defined along a genome. More specialized containers for representing and manipulating short alignments against a reference genome, or a matrix-like summarization of an experiment, are defined in the GenomicAlignments and SummarizedExperiment packages, respectively. Both packages build on top of the GenomicRanges infrastructure.
2018-03-19
Kasper D. Hansen (05:13:29): > @Aaron LunOnce you hit 3 vignettes, perhaps it is time to have an “Overview” for a really quick introduction to what the package provides?
Aaron Lun (05:29:54): > Possibly, yes.
Aaron Lun (05:33:48): > I realized I was also missing another vignette, so I better add that as well.
2018-03-20
Dario Righelli (13:14:44): > @Dario Righelli has joined the channel
Dario Righelli (13:38:36): > hi, > my package has been accepted last week, and even if my pubkey is freely available at “this address”https://github.com/
Nitesh Turaga (13:39:17): > http://bioconductor.org/developers/how-to/git/faq/
Dario Righelli (13:40:18): > yeah, I’m stucked at the second step!
Dario Righelli (13:40:37): > second FAQ… sorry!
Nitesh Turaga (13:41:01): > Which part on the second step?
Nitesh Turaga (13:41:35): > There isn’t information in your question for any of us to help. Take a look at #14 on the FAQ
Dario Righelli (13:42:52): > thanks
Dario Righelli (13:43:07): > using ssh -Tgit@git.bioconductor.orggives me Permission denied (publickey).
Dario Righelli (13:43:37): > even if my key has been added to the git repository…
Dario Righelli (13:44:10): > Indeed I reveived this email, > > Hello Maintainer > > Thank you for submitting your public key to Bioconductor. > Your key has been added to our git server, and you may now > access your package. > > Thank you, > Bioconductor team
Nitesh Turaga (13:45:08): > Ok, #14 asks to you give the output ofgit remote -v.
Dario Righelli (13:45:38): > I already configured the upstream…
Dario Righelli (13:45:48): > origin git@github.com:drighelli/DEScan2.git (fetch) > origin git@github.com:drighelli/DEScan2.git (push) > upstreamgit@git.bioconductor.org:packages/DEScan2.git (fetch) > upstreamgit@git.bioconductor.org:packages/DEScan2.git (push)
Nitesh Turaga (13:49:01): > The reason we ask for the remote info is to check if you’ve configured it properly, and also for the package name. Without the package name, there isn’t much I can do. You have two keys in our system, > > ssh-rsa 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 > > ssh-rsa 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 > > Are you sure you are using private keys which correspond to these public keys? If not check #15 in the FAQ.
Dario Righelli (13:49:43): > yes i’m sure, the second one is the one i’m using from this PC
Dario Righelli (13:49:59): > I’m using it also for other private repos
Dario Righelli (13:50:32): > thanks again for your time
Nitesh Turaga (13:50:48): > Ok, i’m assuming it working now?
Dario Righelli (13:51:12): > yes
Dario Righelli (13:51:18): > what happened?
Dario Righelli (13:54:34): > thanks
2018-04-02
Aaron Lun (11:30:10): > @Nitesh TuragaIt seems that mychipseqDBrepository has a bunch of large datafiles in thevignettes/subdirectory that were not previously there (everything butchipseq_db.Rmdandref.bib, basically). I assume that these were inadvertently added during the transition to git. I could delete them, but they would still be present in the repository history. Could you force-erase them from the bioconductor repo? It’s a little off-putting to have to clone a 100 MB repo when the actual package size is very small.
Nitesh Turaga (11:34:05): > There is a51M ./vignettes/cbp_objects.Rdafile.
Nitesh Turaga (11:34:08): > Is that needed?
Nitesh Turaga (11:35:07): > @Aaron Lun
Aaron Lun (11:35:24): > No, none of the****.Rda, ****.rds,****.bed or ****.bed.gz files are needed.
Aaron Lun (11:35:53): > Hm, meant to put a star there, slack just converted it to bold.
Aaron Lun (11:36:22): > All of those are either output files, or in the case of the blacklist, downloaded from an online source.
Nitesh Turaga (11:46:27): > Try now.
Nitesh Turaga (11:46:56): > I added a few commits to remove those files.
Aaron Lun (11:49:30): > Worked like a charm, thanks for the quick response.
2018-04-03
Aaron Lun (07:57:28): > I’ve also noticed that I am listed as the maintainer forTENxBrainData. I don’t have much to do with it anymore, nor do I see further involvement from my part - the package is essentially finished, and any changes would only reflect updates toHDF5ArrayorAnnotationHub. Indeed, I don’t think I even have write access to the GH or BioC repos, and@Martin Morganhas made more commits than me recently. Can I suggest changing the maintainer, e.g., to the Bioconductor Package Maintainer, or Martin?
Kasper D. Hansen (08:01:17): > @Aaron LunDoesn’t this request go against your “embrace, extend, extinguish” goal of replacing all of Bioconductor with yourself by 2020?
Martin Morgan (08:08:56): > I updated the DESCRIPTION file (by deleting the Maintainer: field, so Authors@R, which lists Bioc Pkg Maintainer as the maintainer, rules)
Aaron Lun (08:09:35): > Thanks Martin.
Aaron Lun (08:44:53): > I’m currently reading the “Creating An ExperimentHub Package” vignette - does themake-data.Rfile need to have that exact name? I’m putting together a package for mychipseqDBworkflow data (for release 3.8, as I’ve missed the 3.7 new package deadline). I’d like to havemake-data-h3k9ac.Rmdandmake-data-cbp.Rmddocuments describing the generation of BAM files for the two datasets that I’ve been using. These will serve the same purpose asmake-data.R, but hopefully by breaking it into two files it should be a lot more readable.@Valerie Obenchain@Lori Shepherdany thoughts?
Aaron Lun (08:45:30): > Ah, forgot what channel I was on. Probably should have put this in general.
Lori Shepherd (09:32:52): > That is okay. We just want some record of what sort of processing happened somewhere for the data that is up on the hub.
2018-04-12
Aaron Lun (11:19:22) (in thread): > I wonder if it is possible to make the ordering on the website consistent with the ordering in which R builds the vignettes. It’s quite awkward to have to keep track of the alphanumeric ordering ofboththe vignette file names as well as the vignette titles. I don’t mind doing one or the other, but doing both seems unnecessary to me.
Aaron Lun (11:23:36): > For example, the display order of the vignettes inhttps://bioconductor.org/packages/devel/workflows/html/simpleSingleCell.htmlis pretty chaotic right now, despite being built in a very strictly controlled order. Yes, I could modify the titles, but it’s hard to do anything when you need a word that sorts above “Analyzing”. And prefixing the titles with numbers (e.g., “1. Introduction”) seems redundant to me when the numbers are already present in the vignette file names; especially as it means I have to rename two things when adding new workflow files to the start/middle of the sequence. - Attachment (Bioconductor): simpleSingleCell (development version) > This workflow implements a low-level analysis pipeline for scRNA-seq data using scran, scater and other Bioconductor packages. It describes how to perform quality control on the libraries, normalization of cell-specific biases, basic data exploration and cell cycle phase identification. Procedures to detect highly variable genes, significantly correlated genes and subpopulation-specific marker genes are also shown. These analyses are demonstrated on a range of publicly available scRNA-seq data sets.
2018-04-13
Martin Morgan (20:49:05) (in thread): > bit of a legacy thing going on, since the workflows are managed the same as packages, and package have been done that way for along time.
2018-05-03
Mike Smith (11:17:27): > In trying to remove duplicate commits I’ve managed to triplicate them insteadhttps://github.com/grimbough/IONiseR/commits/masterI was trying to use thegit reset --hard``git cherry-pickparadigm but seem to have totally cocked it up. Does anyone have any suggestions, since I clearly can’t follow instructions properly? - Attachment (GitHub): grimbough/IONiseR > IONiseR - Quality Assessment Tools for Oxford Nanopore MinION data
Mike Smith (11:22:29): > Also, out of interest, why does removing the duplicate commits matter? In the case above they look like a bi-product of the git-svn bridge and have stopped since switching fully to git..
Stephanie Hicks (12:29:37): > this is a nice git guide in this subjecthttp://justinhileman.info/article/git-pretty/
Nitesh Turaga (12:30:56): > Hi@Mike Smithduplicate commits don’t matter in terms of “going forward” in development. They also do not matter for the build-system. But they do matter, when you have to bug-fix or go back in time to fix a bug.
Nitesh Turaga (12:32:29): > These duplicate artifacts happen when you try to “merge” totally different commit histories. I’ll try to write something up today which is a more active example of clearing duplicate commits, and post the document to thehttp://bioconductor.org/developers/how-to/git/page.
Kasper D. Hansen (13:37:41): > Well@Mike Smithyou no longer have duplicate commits!
Nitesh Turaga (13:49:25): > @Kasper D. HansenIONiseR still has duplicate commits. Can you point me to the version without duplicates?
Kasper D. Hansen (13:49:47): > He has triplicate commits I think
Kasper D. Hansen (13:50:03): > So technically he solved the issue
Nitesh Turaga (13:50:13): > haha! did not catch that:smile:
Mike Smith (15:33:55): > Yes, I tried doing thegit cherry-pickapproach suggested in your email, but now seem to have all the original commits, plus some sort of mega commit that recommitted all the ones I selected. I’ve no idea what I managed to do - I suspect there’s some weird merge going on that I don’t understand. Whengit cherry-pickasks me to resolve a conflict from work I did in 2015 it’s pretty much a guess and I imagine I’m guessing incorrectly.
Nitesh Turaga (15:47:34): > Hi@Mike Smith, i’ll get to this by end of the day tomorrow. I took at look at your package, and i’m just trying to make this as easy as possible and not ver laborious. But branch –> reset –>cherry-pickonto branch is the way to go. It’s the most obvious at least.
2018-05-19
Rhonda Bacher (13:33:16): > @Rhonda Bacher has joined the channel
2018-06-26
Elana Fertig (08:37:03): > @Elana Fertig has joined the channel
Federico Marini (09:53:00): > @Federico Marini has joined the channel
2018-07-30
Mike Smith (07:10:02): > @Nitesh Turagado you have time to take a look at this query on the mailing listhttps://stat.ethz.ch/pipermail/bioc-devel/2018-July/013806.html? It looks like the maintainer managed to accidentally downgraded their version number to lower than what is in the RELEASE_3_7 branch (1.5.2 vs 1.6.0) and now can’t commit a change that reverts it to 1.7.z I’ve tested with one of my packages and confirmed that I encounter the same block when trying to ‘double-bump’ the devel version number so I’m not sure what advice to offer them.
Nitesh Turaga (10:24:27): > Hi@Mike Smith, I’ll take a look at this now.
Mike Smith (10:25:27): > Awesome, thanks.
Nitesh Turaga (10:25:42): > What I don’t understand is why would Shana try to go back in time to begin with.
Nitesh Turaga (10:26:17): > I also tested it, and it’s not possible to just revert versions. I’m actually intrigued in how Shana managed to do that at all.
Nitesh Turaga (10:28:58): > The only possible reasoning I could come up with is, when Shana merged, the conflict was resolved to a odd version number and it got accepted. And then she tried to jump minor versions and it’s not allowed.
Mike Smith (10:29:29): > My assumption was that they didn’t really try to go back in time, but didn’t properly pull and merge from the git.bioc after a release cycle, and then used the existing version on Github. but then I would have expected an ‘changes in the remote you don’t have error’
Nitesh Turaga (10:30:34): > So would you like me to fix themastertree or just update the version to 1.7.1 leaving the bad versions in place?
Nitesh Turaga (10:32:25): > I’ll just reset to6a3e2a79cbde27493dfd21872393d3531ada65ddand push. You can sync your github from there.
Nitesh Turaga (10:33:58): > @Mike Smith
Nitesh Turaga (10:33:59): > ?
Mike Smith (10:35:59): > I guess I should point out that I have no connection to the package in question. I just tried to answer the initial query on the email list.
Nitesh Turaga (10:36:09): > Aaah, ok.
Nitesh Turaga (10:36:15): > I’ll reach out to Shana.
Mike Smith (10:36:46): > I am a bit confused that6a3e2a79cbde27493dfd21872393d3531ada65ddseems to have the correct version number. Why did they change it again?
Nitesh Turaga (10:37:09): > Not a clue.
Mike Smith (10:39:14): > Contacting Shana seems like the best bet, I expect reverting to that commit will be fine. I guess the earlier commits added the bugfix they wanted and it was the the version number to propagate them that caused the issues.
Mike Smith (10:39:49): > Out of interest, how does the git server check the versioning?
Nitesh Turaga (10:39:56): > Thanks. I thought you were related to the package somehow.
Nitesh Turaga (10:40:09): > But, we have a pre-receive hook which checks version numbers.
Nitesh Turaga (10:40:28): > https://github.com/Bioconductor/bioc_git_transition/blob/master/hooks/repo-specific/prevent_bad_version_numbers.py - Attachment (GitHub): Bioconductor/bioc_git_transition > bioc_git_transition - This python package is for transitioning bioconductor from SVN to git
Mike Smith (11:04:07): > Cool thanks. My guess this is a situation where the commit history can be a bit misleading. I presume many of the commits in their log where attempts to correct things, but were rejected but the hook. However they will still appear as commits once a ‘valid’ version number is supplied in the HEAD commit.
Mike Smith (11:05:08): > It looks to me like there isn’t a check in thecheck_version_in_releasethat the Y version remains the same. Maybe they managed to introduce some inconsistency there?
2018-08-22
Mike Smith (09:27:46): > Is it possible to change the file size limits on specific packages? I’m trying to update the pre-compiledlibhdf5.ainRhdf5libbut it’s 6.7MB and so the commit gets rejected.
Martin Morgan (13:19:20): > @Mike SmithI’ll respond to your post on the mailing list
Martin Morgan (16:25:14): > @Mike Smithnitesh gave me some instructions and you should be good to go
2018-08-23
Martin Morgan (06:53:35): > Also@Mike Smith@Hervé Pagèsmentions that from his experience with Rhtslib it was not at the end of the daythatdifficult get the windows DLL to build; but I’m not sure whether he’s saying he’ll actually help here…:smile:
Mike Smith (06:59:27): > My problem has always been that it relies on CMake to actually run the compilation process, and I’m not sure it’s possible to require that on users’ machines. I did manage to get a compilation to complete using the Rtools toolchain once, but the resulting libraries didn’t actually work for reasons that go beyond my understanding. If it’s possible to compile at package installation on Windows I’m all for it, but I admitted defeat a while back.
Mike Smith (07:03:55) (in thread): > Great, that worked for me. Many thanks Martin and@Nitesh Turaga
Aaron Lun (08:38:59): > @Mike SmithDoesbeachmatwork okay with the newRhdf5lib? I will try it out later, but I don’t imagine that the switch to 64 bit would cause any problems… If you could give it a spin as well, that would be nice. (Note that the full suite of tests require--test-dir=longtestsinR CMD checkto execute.)
Martin Morgan (08:40:35) (in thread): > the update looks great, thanks for the tremendous effort!
Aaron Lun (09:11:45): > CHECKing now on Linux. I’ll need you to check on a windows machine, if you can.
Mike Smith (09:14:08): > Looks good to me on Linux > > * checking tests in 'longtests' ... > Running 'testthat.R' > OK >
Aaron Lun (09:16:46): > Oh, sweet. Yeah, I would have hoped it would be okay, C++ wouldn’t care. And if it did, thetypedeforauto’s should have taken care of any type changes.
Aaron Lun (09:17:23): > I’ll probably have to bump versions of all packages linking toRhdf5lib, though, to trigger recompilation. It’s a shame we can’t see the reverseLinkingTodependencies on the landing page.
Aaron Lun (09:30:22): > Yep, works fine on my linux machine as well.
Martin Morgan (12:57:17) (in thread): > We’ll try to make this (reverse LinkingTo) happen
Aaron Lun (12:59:07) (in thread): > Cool. I think I had a similar request forpkgDepToolsa while back.
2018-09-04
Lori Shepherd (09:01:01) (in thread): > Links To Me is now added to the package landing pages as the Reverse LinkingTo
Aaron Lun (09:01:41) (in thread): > :+1:
2018-09-11
Hervé Pagès (15:01:28): > @Hervé Pagès has joined the channel
Hervé Pagès (15:03:33) (in thread): > @Mike SmithNote that we have CMake on the Windows build machines and we could always add other tools if you need them. Our Windows users don’t need to have anything on their machine because we distribute the package binary.
2018-09-17
Mike Smith (03:48:40) (in thread): > Out of interest, so you also have Visual Studio? I’ll take a look at making aconfigure.winthat can build from the distributed source tarball using CMake.
2018-09-19
Hervé Pagès (02:16:28) (in thread): > We don’t have Visual Studio on the Windows build machines. I don’t know the details of what hdf5 needs to compile on Windows but it could even be that you can avoid using aconfigure.winand that aMakevars.winis enough. (FWIW Rhtslib uses aMakevars.wininsrc/and a customMakefile.Rhtslib.wininsrc/htslib-1.7/:https://bioconductor.org/checkResults/3.8/bioc-LATEST/Rhtslib/tokay1-install.html)).
2018-11-01
Aaron Lun (10:02:16): > Did the workflow packages get an automatic version bump? I see thatsimpleSingleCellhas the same version number for both release and devel landing pages.
Martin Morgan (10:05:59): > There was an issue with the workflow packages; I think they are correct in git, but@Nitesh Turagacan shed some light…
Nitesh Turaga (10:34:31): > I have corrected them. They should build today and propagate by this evening.
Nitesh Turaga (10:34:47): > @Aaron Lun
Aaron Lun (10:39:21): > :+1:
2018-11-03
Aaron Lun (09:51:13): > Looks likedevtoolsisn’t installed on the macs for some reason:http://bioconductor.org/checkResults/devel/bioc-LATEST/beachmat/merida2-checksrc.html
Aaron Lun (09:52:02): > ah … and that’s probably because they haven’t rebuilt the mac binaries on CRAN.
2018-11-18
Thomas Girke (14:29:29): > @Thomas Girke has joined the channel
2018-11-23
Aaron Lun (15:59:58): > @Lori ShepherdWould it be much work to get build badges for the longtests as well?
2018-11-26
Lori Shepherd (09:59:46): > I can look into it but there are other projects that are higher priority right now - If you could post in as an issue at thebioconductor.orgrepo so I don’t lose track of it -https://github.com/Bioconductor/bioconductor.org - Attachment (GitHub): Bioconductor/bioconductor.org > Website for bioconductor.org. Contribute to Bioconductor/bioconductor.org development by creating an account on GitHub.
Lori Shepherd (10:00:05): > off hand - I’m not sure how much work it would be
2018-11-27
Aaron Lun (11:17:49): > okay.
2018-12-06
Stevie Pederson (10:26:53): > @Stevie Pederson has joined the channel
Kevin Rue-Albrecht (17:18:30): > Not exactly “git”-related, but not « new package » either: I was just wondering about timing aspects of the nightly build. What is the “TIMEOUT” limit again? > All I could find was “The package should require less than 5 minutes to run R CMD check –no-build-vignettes.” (http://bioconductor.org/developers/package-guidelines/) but when I look at packages affected by « TIMEOUT » in the build report, it seems to be 40 min (for CHECK), is that right?
Martin Morgan (18:32:29): > the guideline is 5 minutes, and we hold packages to this as much as possible; the build system is more generous to allow for (a) legacy and (b) transient, usually internet related, long times.
2018-12-07
Kevin Rue-Albrecht (01:54:15): > Thanks! > Then my second, related, question is whether anExperimentHubcache may be used to speed up subsequent builds, using an internet connection only the first time (?). Does a cache “survive” between two builds? Or are data packages (e.g.scRNAseq) still the #1 recommended practice?
Martin Morgan (02:12:00): > yes, the cache survives between builds; I like ExperimentHub (for the data) with a light-weight experiment data package wrapper for the metadata / user access.
Kevin Rue-Albrecht (02:30:18): > Ok. Thanks again. My underlying motivation being to have a few examples including both dataandmetadata to use as vignette material (for the Hancock package, in development). > A bit more searching and re-discovering on my part pointed out that the ExperimentData packages like « airway » could do just fine.
2018-12-30
Evan Biederstedt (14:39:55): > @Evan Biederstedt has joined the channel
2019-01-17
Kayla Interdonato (08:27:03): > @Kayla Interdonato has joined the channel
2019-01-24
Ming Tang (19:39:37): > @Ming Tang has joined the channel
2019-02-07
Hirak (14:18:37): > @Hirak has joined the channel
2019-02-20
Kellie Kravarik (11:37:31): > @Kellie Kravarik has joined the channel
2019-03-17
gamzeaydilek (07:17:38): > @gamzeaydilek has joined the channel
2019-04-26
Almut (09:45:11): > @Almut has joined the channel
2019-05-06
Firas (10:31:42): > @Firas has joined the channel
2019-05-12
Aaron Lun (15:59:53): > Watchinghttps://bioconductor.org/developers/gitlog/can be quite entertaining.
Aaron Lun (16:00:06): > Find some kindred spirits with nothing else to do on the weekend.
Aaron Lun (21:57:18): > liking@Hervé Pagès’ epic GenomicRanges commit message f988a5a9ae9885aa500766930a6ff68ae606a906
2019-05-13
Hervé Pagès (12:00:21): > Glad you enjoy my ASCII art@Aaron LunBTW this little reorg of the GenomicRangesList hierarchy is causing some breakage on the latest BioC 3.10 build report. Still working on it with the hope that tomorrow’s report will start to look a little bit better:crossed_fingers:
Aaron Lun (12:11:23): > :+1:
2019-05-24
Simina Boca (15:46:25): > @Simina Boca has joined the channel
Nitesh Turaga (15:46:51): > Hi@Simina BocaIf you have any git questions, feel free to ask.
Simina Boca (15:46:56): > hi!
Simina Boca (15:47:07): > thank you for pointing this out and for all your help so far
Simina Boca (15:47:43): > I’m fine with the previous package but having some other issues with another one - still trying to do a bit of trouble-shooting first, since I think part of the issues are git-related
Nitesh Turaga (15:49:05): > okay
Simina Boca (15:56:41): > So I’m getting this error message: > $ ssh -Tgit@git.bioconductor.orggit@git.bioconductor.org: Permission denied (publickey).
Simina Boca (15:56:58): > Despite the fact that I did upload the key to GitHub
Nitesh Turaga (15:57:34): > Right, is it the same key which is available on the BiocCredentials app as well?https://git.bioconductor.org/BiocCredentials/
Nitesh Turaga (15:57:56): > The public key should match the one you have on github.
Simina Boca (15:59:06): > yeah, I think they all match
Nitesh Turaga (15:59:06): > If you are not sure you are using the correct key, you could try, > > ssh -i </path/private_key_name> -T[git@git.bioconductor.org](mailto:git@git.bioconductor.org) > > The default key is usually the one labeledid_rsa.
Nitesh Turaga (16:00:12): > your SSH keys should be in your.sshfolder usually.
Simina Boca (16:03:55): > they are there but I still get the same error…
Simina Boca (16:04:02): > Simina@INFORMA-JQWJBH2 MINGW64 ~/Documents/Github/swfdr (master) > $ ls ~/.ssh > id_rsa id_rsa.pub known_hosts > > Simina@INFORMA-JQWJBH2 MINGW64 ~/Documents/Github/swfdr (master) > $ ssh -i ~/.ssh/id_rsa -Tgit@git.bioconductor.orggit@git.bioconductor.org: Permission denied (publickey).
Nitesh Turaga (16:10:58): > Can you check one more thing, that if you try adiffbetween the output of > > ssh-keygen -y -f ~/.ssh/id_rsa > temp.pub > > > > diff temp.pub id_rsa.pub > > This is to make sure that the private key and public key are actually corresponding key pairs.
Nitesh Turaga (16:12:52): > Everything on the Bioconductor end of things is ok as far as the set up for your packageswfdrgoes.
Simina Boca (16:15:04): > what should I see if they’re key pairs?
Nitesh Turaga (16:15:28): > Ideally, there should be no difference between the temp.pub and the id_rsa.pub.
Simina Boca (16:15:50): > got it
Simina Boca (16:16:01): > the only difference is that temp.pub has no email at the end
Nitesh Turaga (16:16:23): > It means that yourid_rsa(private key) andid_rsa.pubare genuine key pairs. The email doesn’t make a difference.
Nitesh Turaga (16:16:48): > ok, back to square one.
Nitesh Turaga (16:17:45): > Can you send me the output of thegit remote -vin your swfdr directory?
Simina Boca (16:18:47): > $ git remote -v > originhttps://github.com/leekgroup/swfdr.git(fetch) > originhttps://github.com/leekgroup/swfdr.git(push)
2019-05-26
Jayaram Kancherla (09:08:11): > @Jayaram Kancherla has joined the channel
2019-06-25
Daniela Cassol (15:40:25): > @Daniela Cassol has joined the channel
2019-06-26
Junhao Li (13:25:27): > @Junhao Li has joined the channel
2019-11-08
Mike Jiang (17:45:41): > @Mike Jiang has joined the channel
Mike Jiang (17:45:59): > > RProtoBufLib$ git push bioc trunk:master --force > Counting objects: 199, done. > Delta compression using up to 4 threads. > Compressing objects: 100% (185/185), done. > Writing objects: 100% (199/199), 2.23 MiB | 4.64 MiB/s, done. > Total 199 (delta 113), reused 0 (delta 0) > remote: Resolving deltas: 100% (113/113), completed with 12 local objects. > remote: FATAL: + refs/heads/master packages/RProtoBufLib m.jiang DENIED by fallthru > remote: error: hook declined to update refs/heads/master > To[git.bioconductor.org:packages/RProtoBufLib.git](http://git.bioconductor.org:packages/RProtoBufLib.git)! [remote rejected] trunk -> master (hook declined) > error: failed to push some refs to '[git@git.bioconductor.org](mailto:git@git.bioconductor.org):packages/RProtoBufLib.git' >
Mike Jiang (17:49:21): > I have a local branch that diverts from the existing bioc/devel quite a lot, so how do I force it to be pushed bioc devel without having to resolve all the conflicts(which will cost a lot efforts and error-prone since the code bases are significantly different)
Mike Jiang (18:10:52): > looks likegit merge bioc_devel -s ourspaired withgit merge trunk -x theirsworks!
2019-11-12
Nitesh Turaga (04:45:27): > Hi@Mike JiangYou cannot “force” push.
Nitesh Turaga (04:46:30): > I’d have to force it for you, if that’s what you wish. But merging is the best option from your end to maintain continuity.
2019-11-20
Russ Bainer (12:03:30): > @Russ Bainer has joined the channel
2019-12-04
Jonathan Carroll (17:37:45): > @Jonathan Carroll has joined the channel
2019-12-09
Jovana Maksimovic (23:44:22): > @Jovana Maksimovic has joined the channel
2019-12-11
Aaron Lun (11:38:12): > @Aaron Lun has left the channel
Christine Choirat (12:08:24): > @Christine Choirat has joined the channel
2020-03-19
Stian Lågstad (15:02:56): > @Stian Lågstad has left the channel
2020-03-20
Aaron Lun (03:44:15): > @Aaron Lun has joined the channel
Aaron Lun (03:44:29): > @Nitesh TuragaDoes the BioC git server support submodules?
Nitesh Turaga (08:30:08): > Nope
Nitesh Turaga (08:30:55): > The bioconductor package structure doesn’t support submodules so there was never a need
Nitesh Turaga (08:30:59): > @Aaron Lun
Aaron Lun (15:13:34): > Okay.
Aaron Lun (15:16:06): > I’ve got some screenshots that I’d like to be able to include into the vignettes. It may be that the best solution is just to commit the PNGs directly to the git repo, as distateful as that seems.
Nitesh Turaga (15:16:40): > Yep
Aaron Lun (15:18:53): > Hm. Maybe I can set up partial clone to avoid pulling that particular subdirectory unless specified.
Sean Davis (20:29:56): > You can also use solutions likehttps://imgur.com/and then use links to images. - Attachment (imgur.com): Imgur > Imgur: The magic of the Internet
Aaron Lun (20:37:27): > That was considered, and in fact was my first approach (hosted in a separate github repository, to be precise). But the problem is that it introduces a dependency on another system, which is almost as undesirable as the big binary blobs.
2020-03-21
Aaron Lun (00:52:04): > Well. I can’t say I’m all too happy about it. But I’ve just stuffed the screenshots into the package repo.
2020-03-27
ceclaurent (06:09:59): > @ceclaurent has joined the channel
2020-03-30
Minoo (11:15:33): > @Minoo has joined the channel
Minoo (11:18:54): > Hi@Nitesh Turaga, Thanks for all your helps.
2020-05-07
Aaron Lun (17:13:51): > What is in scope for the Bioconductor GitHub organization? Can I move some of the packages out of my user account into there?
Nitesh Turaga (17:14:36): > Depends@Aaron Lun. The packages in the Bioconductor Github org are maintained by the “core team” or rather “maintainer@bioconductor.org”.
Nitesh Turaga (17:15:17): > i.e folks on the team is responsible for fixing bugs etc and regular package maintenance.
Nitesh Turaga (17:15:31): > i’m not sure if that answers your question though.
Aaron Lun (17:16:44): > For example, what reasoning do we have for the OSCA book content to live there? It’s a good place, but if I wanted to add more book content, can I?
Nitesh Turaga (17:18:02): > Are you talking about thishttps://github.com/Bioconductor/OrchestratingSingleCellAnalysis-devel?
Aaron Lun (17:18:12): > Yes.
Nitesh Turaga (17:18:23): > Do you not have certain permissions?
Nitesh Turaga (17:18:28): > to add more content?
Aaron Lun (17:18:38): > Sorry, I meant “add more content” in terms of “add a different book”.
Nitesh Turaga (17:23:46): > So a different github repo?
Aaron Lun (17:24:02): > Yes.
Nitesh Turaga (17:25:05): > I guess so..It seems you don’t have permissions for that, so you’ll have to have someone on the core team open the repo and add you as a collaborator.
Aaron Lun (17:26:04): > And that’s fine, but the question is, should they? How do we decide whether a repo is “worthy” of being in the Bioconductor org?
Nitesh Turaga (17:26:24): > @Martin MorganDecides that I guess.
Aaron Lun (17:26:25): > If you say it’s just the core team-maintained packages, that’s fine, but then it doesn’t make sense that the OSCA repo is in there.
Leonardo Collado Torres (17:26:56) (in thread): > not using threads eh:stuck_out_tongue:
Aaron Lun (17:26:59): > “Core team repo and books with theBioconductor.orgdomain name” is also okay but not super elegant.
Nitesh Turaga (17:27:05) (in thread): > I take that back, after looking at our org some more, there are more nuanced reasons.
Nitesh Turaga (17:27:59): > So, what other book did you want to add I guess? Maybe that’s the better question.
Nitesh Turaga (17:28:18): > Instead of trying to generalize the organization details…
Aaron Lun (17:29:20): > It’s mostly hypothetical. There’s a proposed spatial book from#spatial, and a proposed mass cytometry book floating around somewhere.
Aaron Lun (17:30:09): > I have a book for a specific package athttps://ltla.github.io/SingleRBook, but that doesn’t seem as official without a nice “bioconductor.org” at the end. - Attachment (ltla.github.io): Assigning cell types with SingleR > The SingleR book. Because sometimes, a vignette just isn’t enough.
Hervé Pagès (17:30:16): > I don’t think we have an established policy yet so it’s been more on a case by case so far. Send us a check and we can look at your request:wink:
Martin Morgan (18:29:33) (in thread): > it seemed like OSCA was a highly valuable Bioconductor resource, and it would benefit both the project and the contributors to have it published there. And we were asked if we would host it. In particularosca.bioconductor.orgseemed like good ‘branding’ compared torando.github.io/osca. I wouldn’t host a one-package book (but well-organized vignettes are a great resource for users, available in the package as well as at standard locations on the web site)
2020-05-08
Leonardo Collado Torres (15:55:19) (in thread): > it’s always about $$$ hehe:stuck_out_tongue:
Hervé Pagès (18:38:01): > BTW I see that there are now 3 OrchestratingSingleCellAnalysis-* repos (base, release, and devel) in the Bioconductor organization on GitHub. Couldn’t branches by used instead?
Aaron Lun (18:38:29): > I couldn’t figure out how to deploy different GitHub Pages off different branches.
Martin Morgan (18:39:45): > I agree that branches somehow seem more appropriate here; maybe others have an understanding of ‘advanced’ github pages (or maybe the project has outgrown github pages?)
Marcel Ramos Pérez (18:44:30): > I don’t know how they each are being deployed but couldn’t each branch have it’s own_bookdown.ymlfile andoutput_dir?
Aaron Lun (18:48:21): > The problem is precisely how they are being deployed.
Aaron Lun (18:49:10): > GitHub Pages just doesn’t support taking things from other branches beyondgh-pages.
Aaron Lun (19:56:33): > There is the additional complication that -release and -devel consist purely of rendering artifacts; all of the real content is hosted in *-base, which is used to dynamically regenerate the Rmarkdown files that are then used for bookdown.
2020-05-09
Sean Davis (15:19:02): > @Aaron Lun, I think github pages can use a different branch besides gh-pages, but only one such branch. That doesn’t address your problem, but just wanted to point that out.
Sean Davis (15:20:59): > In practice, though, to use gh-pages for multiple “versions”, you could build each version and place each in a subdirectory and then have a single “index” page that links to each version subdirectory. Then push that entire set of files togh-pages.
2020-05-13
Nitesh Turaga (12:30:00): > > Dear Developers, > > Every time there is a release for Bioconductor we (the core team), face an issue where developers end up pushing code with the "merge conflicts". One of the reasons why we get the issue "Bad description" file in the build report is because of a merge which has gone bad and produced conflicts. > > To avoid this issue, we have taken a step to include a pre-receive hook in git which checks for the merge conflict markers and rejects your push. > > Please note that, although this will keep our end of things (the Bioconductor git repo) clean, your local branch will need to be fixed before you try and push to Bioconductor again. To avoid having your push to Bioconductor rejected, it's easy to employ some good practices, > > 1. Check your code and DESCRIPTION file for merge conflict markers, i.e for anything that looks like > > ++<<<<<<< HEAD > +Version: 1.1.2 > ++======= > + Version: 1.0.1 > ++>>>>>>> origin/RELEASE_3_11 > > Notice the <<<<<<< HEAD, =======, >>>>>>> markers in the files. > > 2. Before a commit, you can use "git diff --check", and it will show you the issues with your code. > > git diff --check > > DESCRIPTION:4: leftover conflict marker > DESCRIPTION:6: leftover conflict marker > DESCRIPTION:8: leftover conflict marker > > 3. If you use an IDE there are certain "extensions" you can add to your IDE to make it detect merge conflicts for you automatically. Most editors have them (emacs, vim, visual-studio, atom, sublime etc). Please use these if they are your development environments. You can even get built in ones for bash or zsh. > > If you need help resolving merge conflicts, please check,[https://help.github.com/en/github/collaborating-with-issues-and-pull-requests/addressing-merge-conflicts](https://help.github.com/en/github/collaborating-with-issues-and-pull-requests/addressing-merge-conflicts)This is mostly an effort to avoid packages from breaking for needless reasons. As usual, thanks to everyone for the contributions. > > > Best regards, >
2020-05-22
Charlotte Soneson (08:00:55): > @Charlotte Soneson has joined the channel
2020-06-06
Olagunju Abdulrahman (19:56:54): > @Olagunju Abdulrahman has joined the channel
2020-07-14
Luke Zappia (06:08:52): > @Luke Zappia has joined the channel
2020-07-30
Ayush Raman (12:44:18): > @Ayush Raman has joined the channel
Rene Welch (12:57:13): > @Rene Welch has joined the channel
2020-07-31
Dr Awala Fortune O. (16:14:06): > @Dr Awala Fortune O. has joined the channel
2020-08-10
Aaron Lun (01:09:46): > @Nitesh TuragaWhat do you think about having some repo/user-specific SSH keys for machine accounts to automate pushes to the BioC Git servers? I’m thinking in the context of workflows from GHA, where I would like to get automatic propagation to BioC git but would rather not have my private keys distributed throughout all participating repos.
Nitesh Turaga (08:20:46): > You can do this now@Aaron Lun. Just create a new key and put it on your Bio credentials and make sure it’s the one you are using on your GitHub actions.
Nitesh Turaga (08:20:54): > I’m not sure why you can’t do this now??
Aaron Lun (11:13:17): > The last I remember, the SSH keys were linked to my entire GitHub account; I didn’t want to give GHA for a given repo access to all my Bioconductor repos.
Nitesh Turaga (11:14:05): > Then, you can add SSH keys to your BiocCredentials account which isn’t on your GIthub
Nitesh Turaga (11:14:46): > That way, you’ll be able to provide a “limited” key…use the same key pair to push to your Bioconductor Git repo.
Aaron Lun (11:15:02): > TBH I don’t think I’ve ever used BiocCredentials when it came out, I must have set up my SSH keys way earlier.
Aaron Lun (11:15:06): > I’ll give it a go.
Nitesh Turaga (11:15:12): > Yep, that’s the best way
Nitesh Turaga (11:15:18): > Activate your account and you’ll be set
2020-08-11
Aaron Lun (00:00:46): > @Nitesh TuragaIt keeps on telling me thatinfinite.monkeys.with.keyboards@gmail.comis not associated with a BioC maintainer
Aaron Lun (02:42:21): > Is the system still listening to public keys posted on GitHub, or is everything managed via BiocCredentials now?
Nitesh Turaga (06:53:22): > We can solve this easily over a 5 min call during the day@Aaron Lun… How about around 1pm est.
Aaron Lun (11:08:44): > sure.
Aaron Lun (11:08:54): > hold on, 1 pm est… that’s…
Aaron Lun (11:08:58): > 10 am.
Aaron Lun (11:09:01): > okay.
Nitesh Turaga (11:09:03): > Yeah…
Nitesh Turaga (11:09:05): > :smile:
Aaron Lun (13:39:59): > While you’re here, I also got a fun message from the pre-commit hooks: > > Counting objects: 51, done. > Delta compression using up to 8 threads. > Compressing objects: 100% (21/21), done. > Writing objects: 100% (51/51), 5.90 KiB | 5.90 MiB/s, done. > Total 51 (delta 29), reused 38 (delta 18) > remote: Error: Illegal version bump from '1.0.0' to '1.1.0'. > remote: > remote: Check[http://bioconductor.org/developers/how-to/version-numbering/](http://bioconductor.org/developers/how-to/version-numbering/)remote: for details. > 2 Package: OrchestratingSingleCellAnalysis > remote: > To git.bioconductor.org:packages/simpleSingleCell > ! [remote rejected] master -> master (pre-receive hook declined) > error: failed to push some refs to 'git@git.bioconductor.org:packages/simpleSingleCell' > > Even though my packageDESCRIPTIONwas 1.13.11. Turned out it was picking up the book’sDESCRIPTIONjumping from 1.0.0 to 1.1.0, even though all the book contents were hidden away in a subdirectory of thevignettes/. Is this intentional?
Nitesh Turaga (13:40:42): > what package is this?
Aaron Lun (13:40:52): > simpleSingleCell.
Nitesh Turaga (13:40:52): > ooops…Simple Single Cell
Nitesh Turaga (13:40:55): > yep…
Aaron Lun (13:41:01): > I deleted the offending lines in the book’s description
Aaron Lun (13:41:12): > but it seems a bit odd that the hook would care about it.
Nitesh Turaga (13:41:13): > Does it work after that?
Aaron Lun (13:41:15): > yep.
Nitesh Turaga (13:41:34): > Let me take a look
Nitesh Turaga (13:43:07): > Ok, so, the hook looks at a file in thediffcalled DESCRIPTION…
Nitesh Turaga (13:43:19): > so when you changed your version number in the DESCRIPTION file in your book….
Nitesh Turaga (13:43:32): > it’s like…i’m going to check this just like I do any other package…
Nitesh Turaga (13:43:37): > and i’ll throw an error…
Aaron Lun (13:43:45): > shouldn’t it look for./DESCRIPTION?
Nitesh Turaga (13:43:50): > The hook is not smart enough to expect more than one DESCRIPTION file.
Nitesh Turaga (13:44:19): > Well, it didn’t need to before this …
Nitesh Turaga (13:44:21): > :stuck_out_tongue:
Nitesh Turaga (13:44:49): > considering all the other R packages didn’t think to put description files elsewhere haha
Aaron Lun (13:44:52): > I’ll note that I actually have at least two other examples with test packages inside a package.
Aaron Lun (13:45:03): > beachmatand I thinkbasilisk.
Nitesh Turaga (13:46:11): > I can keep it on my radar and improve the hook….
Aaron Lun (13:46:21): > :+1:
Nitesh Turaga (13:46:22): > Maybe in the next couple sprints…sorry:confused:
2020-09-06
Aaron Lun (15:01:06): > @Nitesh TuragaCan you give@Alan O’Cwrite access to the scater git repo?
Aaron Lun (15:01:24): > He should already have an account for thesnifterpackage, amongst other things.
2020-09-07
Michael Stadler (03:44:31): > @Michael Stadler has joined the channel
Michael Stadler (04:05:44): > Hi@Nitesh Turaga. (edit: this seems to be a duplicate of the message from Alina Selega from yesterday to the bioc-devel mailing list). I have just tried to push updates for two of the packages that I am maintaining (QuasRandeisaR) togit.bioconductor.org, and I received an error message: > > >git push upstream master > Counting objects: 58, done. > Delta compression using up to 96 threads. > Compressing objects: 100% (36/36), done. > Writing objects: 100% (48/48), 8.26 KiB | 0 bytes/s, done. > Total 48 (delta 21), reused 0 (delta 0) > remote: fatal: this operation must be run in a work tree > remote: ERROR:root:Exception: Command '['git', 'rev-parse', '--show-toplevel']' returned non-zero exit status 128 > remote: ERROR:root:local variable 'package_name' referenced before assignment > remote: Note: failed to update RSS feed;git repository updated successfully. > To git@git.bioconductor.org:packages/eisaR > 6358c47..e56d151 master -> master > > >git pull upstream master > From git.bioconductor.org:packages/eisaR > * branch master -> FETCH_HEAD > Already up-to-date. > > >git push upstream master > Everything up-to-date > > I think that my changes were pushed correctly, as for both packages, a subsequent pull/push did not show any further changes that needed to be synchronized. Is this something I should worry about? > > Here is my remote setup (here for eisaR, but it’s the same also forQuasR): > > >git remote -vv > origin git@github.com:fmicompbio/eisaR.git (fetch) > origin git@github.com:fmicompbio/eisaR.git (push) > upstream git@git.bioconductor.org:packages/eisaR (fetch) > upstream git@git.bioconductor.org:packages/eisaR (push) >
Alan O’C (06:25:33): > @Alan O’C has joined the channel
Nitesh Turaga (12:29:02) (in thread): > Send an email to Bioc-devel about this. For record keeping.
Nitesh Turaga (12:29:23): > I am aware of the RSS feed issue. I’ll take a look tomorrow.
Aaron Lun (13:40:44) (in thread): > done
2020-09-08
Nitesh Turaga (10:11:31): > <!here>The issue with RSS feed should be fixed now.
2020-09-14
Belinda Phipson (07:19:59): > @Belinda Phipson has joined the channel
2020-09-17
Aaron Lun (19:44:47): > @Nitesh Turagaany chance of getting TSCAN access?
Nitesh Turaga (19:45:06): > TSCAN access?
Nitesh Turaga (19:45:32): > Did you co-develop?
Aaron Lun (19:45:40): > yeah; there was an email on the BioC-devel mailing list about it.
Aaron Lun (19:46:10): > No, but I’m going to share some of the maintainership responsibilities with the current maintainers.
Nitesh Turaga (19:46:12): > Aaah..
Nitesh Turaga (19:46:13): > thanks!
Nitesh Turaga (19:46:17): > Now I see it.
Nitesh Turaga (19:46:24): > one minute
Nitesh Turaga (19:47:48): > Try now.
Aaron Lun (19:48:34): > thanks, looks like it works
Aaron Lun (19:48:50): > :+1:
2020-10-04
Aaron Lun (05:23:42): > I’m looking at my GitHub Actions again and there is this part where Issh-keyscanto get the host key fingerprint for the Bioc git servers.
Aaron Lun (05:24:49): > apparently, how I’mmeantto be doing it is to get the fingerprint from some published source, otherwise it’s not secure to man-in-the-middle attacks.
Aaron Lun (05:25:43): > Do we have some SSH host key fingerprints that I can just slap into my scripts?
2020-10-26
Jeffrey O’Brien (14:03:10): > @Jeffrey O’Brien has joined the channel
2020-10-27
Vince Carey (05:38:27): > @Vince Carey has joined the channel
Vince Carey (05:45:04): > What are the prospects of abstracting a few of the key bioc-git operations into functions in a package? There has been a lot of discussion on bioc-builds suggesting that more high-level concept packaging could help developers who, like me, are not expert in git, and do not use it much. Package functions could warn about pushes that don’t include a version bump. There could also be a protocol for checking for existence of a log demonstrating that CMD check was run before a push.
Federico Marini (05:48:08): > I have recently seengert
Federico Marini (05:48:09): > https://docs.ropensci.org/gert/ - Attachment (docs.ropensci.org): Simple Git Client for R > Simple git client for R based on libgit2 with support for SSH and HTTPS remotes. All functions in gert use basic R data types (such as vectors and data-frames) for their arguments and return values. User credentials are shared with command line git through the git-credential store and ssh keys stored on disk or ssh-agent. On Linux, a somewhat recent version of libgit2 is required; we provide a PPA for older Ubuntu LTS versions.
Federico Marini (05:48:35): > maybe some of these functions can build up a base for that
Vince Carey (05:58:46): > Interesting slide set at the gert site given above. It could be useful in bioc extensions to add some visualizations to the doc to illustrate branch structure, and to provide defaults that reduce cognitive load for common steps in bioc package repo maintenance. Another concept to work in is the simultaneous use of github for many package repos … if that makes sense.
Nitesh Turaga (07:08:14): > This is a great discussion. I’ll add my two cents, > 1. Unless we automate a lot of the finer functionality ofgitand hide it under some function saybioc_git_push(), I wonder if a package like ‘gert’ is actually that useful. It’ll be harder to debug (for the user) if something goes wrong in between when we hide most of the details under said functionbioc_git_push(). But I take the point that if we do decide to do this,gertcan provide a good base (i’ve not actually explored it though). > 2. Another thing is, since moving togitto make package contributions we have forced developers in a meaningful way to learn about version control. I don’t think this is a bad thing. It is a useful thing to require for developers in this day and age. > 3. There are ways we can leverage ‘RStudio’ which has a UI to interact withgit. We can make the process of committing easier, produce instructions via the UI (create a video, or take screenshots). > 4. If we need to generate a warning about a push not having a version bump, we can do this via agit hook. We can leave a messageon pushif a version bump is missing.
Federico Marini (09:14:30): > My 1 cent on your 2 cents,@Nitesh Turaga2- Yep! It is a very nice positive side effect! > 4- The git hook way is exactly how for exampleusethisdoes to check that, if you have aREADME.Rmdto generate themd, that would need to be re-rendered before pushing. > In a similar way we can have that check performed when the user wants tobioc_git_push()Maybe this can be an idea indeed for the Bioc pendant ofusethis, i.e. the new addition to the upcoming releaseBiocThisfrom@Leonardo Collado Torres
Nitesh Turaga (09:18:33): > These were the git-hooks I was talking abouthttps://github.com/Bioconductor/git-hooks. Bioconductor has it’s own set of pre-receive and post-recieve hooks. Just in case any one was curious.
Federico Marini (09:19:36): > Oh ok, then probablyusethisexploits a slightly different mechanism?
Nitesh Turaga (09:19:57): > yes
Federico Marini (09:21:09): > https://github.com/r-lib/usethis/blob/master/R/readme.R#L48-L53+https://github.com/r-lib/usethis/blob/master/inst/templates/readme-rmd-pre-commit.sh
Alan O’C (09:25:04): > Yeah I think abstracting git to an R wrapper will just make it harder to find help outside of bioc. Some pre-commit hooks would be useful for bioc packages though
Vince Carey (10:06:02): > Good discussion. I would not propose abstracting git, but abstracting the most commonly performed operations that seem to require frequent support from the dev team in the project. Thus I would tend to agree with Nitesh’s point 1. I will try to learn about the hooks, these might be a good start.
Nicholas Cooley (10:07:03): > @Nicholas Cooley has joined the channel
2020-11-05
FelixErnst (13:12:31): > @FelixErnst has joined the channel
Leonardo Collado Torres (13:22:39): > Hi
was brought to my attention today, which I didn’t know about. I’m a bit confused on why files such as./github/workflows/something.ymlwould not be allowed on the default branch of a GitHub repo that gets submitted to Bioc. > > This seems to disable using GitHub Actions on Bioc packages as well as having template GitHub files under.githubsuch ashttps://github.com/LieberInstitute/recount3/tree/master/.githubfor example. I mean, I could make a copy of the default branch and add the GHA workflow there, but it seems unnecessary. File size-wise, these are small files that can be ignored by.Rbuildignoresuch ashttps://github.com/LieberInstitute/recount3/blob/master/.Rbuildignore#L6. Otherwise, you have to keep 2 branches around and make sure they are in sync and all that (unless there’s a way to tell GHA to run a workflow from another branch). > > Similarly say adding adev/folder with some R scripts that is ignored by.Rbuildignore. I use those R scripts to include code that I need to re-run frequently. For examplehttps://github.com/LieberInstitute/recount3/blob/master/dev/04_update.R. > > I do get the point that nobigextra files should be on a repo. Like say, some rendered Rmarkdown HTML files. Is there an example that prompted this change? There might be use cases you’ve seen that I’m missing and that motivated this change.
Aaron Lun (13:23:35): > amen.
Nitesh Turaga (13:24:42): > @Lori Shepherdmight want to look at this.
Nitesh Turaga (13:26:01): > I think it was done to prevent cloning of those files into our git repo. But i’ll refrain, because i’m not aware of this change.
Spencer Nystrom (13:54:12): > @Spencer Nystrom has joined the channel
Lori Shepherd (14:55:55): > yes that was the original intention as we did not want to have unneeded additional files cloned into thegit.bioconductor.orgrepositories – but it can be up for discussion
Nitesh Turaga (14:58:37): > Since they are not data files, I don’t see any harm of having them there.
Nitesh Turaga (14:59:22): > But maybe others would like to pitch in here…
Nitesh Turaga (14:59:30): > I’ll tag@Martin Morganas well.
Martin Morgan (15:08:13): > One concern is that these files have side-effects that one would not expect without deep knowledge of the package, for instance when I clone fromgit.bioconductor.organd push to my own github repo, I get the workflows unexpectedly, and spend time understanding what is going on; a similar concern is with .Rproj files, imposing configuration choices on my RStudio session. I think more generally one wants to get a ‘package’, and not other material that may or may not be relevant. > > And I think that a ‘bioconductor’ branch is a good development practice for a package with extensive github development. > > From a project administration process, of course we ‘know’ about github workflows, but in general we do not want to be in the position of investigating each non-standard file submitted to arrive at a statement of marginal utility – 270 votes and we’ll let the file in!
Nitesh Turaga (15:11:14): > Hahaha!:smile:
Alan O’C (15:18:51): > It seems like forcing CI configuration files out of the default branch will discourage people from using CI and just result in more broken packages. Surely CI outside of the BBS should be encouraged not discouraged? But maybe I’m telling on my own laziness there.
Alan O’C (15:19:28): > Unless this is strictly for submission and after submission it’s fine, but then that seems counter to the point of review
Vince Carey (15:22:27): > I’d say this is a policy change that needs more discussion. At the moment I don’t see a significant downside to including the .github/workflows on the default branch.
Kasper D. Hansen (15:38:09): > Note that R CMD check will complain about such files and one can include an.Rbuildirgnorewhich - to me - acknowledges the existence of files that are not part of the package per se.
Kasper D. Hansen (15:38:39): > You could also argue that this could be solved by Bioconductor building from a “biocdevel” branch
Kasper D. Hansen (15:40:10): > And finally, is Bioc git just a place to put final files or do we intent final development to be done here. If its the later, I do think the developer should have a decent amount fo control over their workflow. Ie. I think that we should expect for sure that R CMD build gives a “clean” package but as long as that is satisfied …
Aaron Lun (16:02:45): > The proposed policy is quite obstructive. If I’ve marked everything out in the.Rbuildignore, there is no reason for the build system to complain. If a package reviewer is irritated by rubbish files in the repo, then a good approach might be to review the package tarball afterR CMD build .. If the rubbish files are explicitly listed in the.Rbuildignore, then they are probably important to the developer but won’t show up on the tarball, and it’s a win-win for everyone. If they’re not listed in the.Rbuildignore, then the review comment would be to either list them or delete them, and developers can make up their own minds as to whether something is important enough to have to manually list it in the.Rbuildignore.
2020-11-06
Federico Marini (03:41:50) (in thread): > and commit and commit and commit and commithttps://twitter.com/AllTransLivesM1/status/1324433218431557633?s=20 - Attachment (twitter): Attachment > I can’t :joy:
Vince Carey (06:37:40): > One issue is that Bioconductor core manages more than what is in the post build tarball. We manage using git and not github. We have write privileges over all the files in git. So the policy needs to be thought through with care.
2020-11-10
Geet (12:48:48): > @Geet has joined the channel
2020-11-18
eugenia.galeota (05:38:34): > @eugenia.galeota has joined the channel
2020-11-30
James MacDonald (16:58:16): > @James MacDonald has joined the channel
Gabriel J. Odom (17:15:36): > @Gabriel J. Odom has joined the channel
Gabriel J. Odom (17:24:23): > Hi all, > Would anyone be able to answer a (hopefully) quick question for a neophyte? My package,pathwayPCA, currently has a failing build. I believe that I have a solution, and I was able to commit and push said potential fix to devel / master on the Bioc side. However, I asked Lori, and she said that she hasn’t seen a change in RELEASE_3_12 (only master). So far, I believe that I have followed all the steps here (https://bioconductor.org/developers/how-to/git/sync-existing-repositories/), but I’ve clearly missed something. > Any help would be greatly appreciated
Marcel Ramos Pérez (17:28:09): > Hi Gabriel, perhaps you need to do Step 4+ from here?https://bioconductor.org/developers/how-to/git/bug-fix-in-release-and-devel/Feel free to file an issue / PR athttps://github.com/Bioconductor/bioconductor.org/if the documentation is not clear. Thanks!
Gabriel J. Odom (21:31:30) (in thread): > Thank you! I think that I was getting hung up on there being two different version numbers in the DESCRIPTION. So my master branch is the “development” branch, and has version 1.7.2, while RELEASE_3_12 is a “production” branch with version 1.6.1–even though the code is identical (I made a bug fix in both).
Marcel Ramos Pérez (22:34:57) (in thread): > I don’t really know the specifics but FWIW atomic commits are easier to cherry-pick. I usually separate bug fix and version bump commits.
2020-12-01
Lluís Revilla (11:17:17): > @Lluís Revilla has joined the channel
Kevin Stachelek (22:15:26): > @Kevin Stachelek has joined the channel
2020-12-02
Konstantinos Geles (Constantinos Yeles) (05:19:27): > @Konstantinos Geles (Constantinos Yeles) has joined the channel
rohitsatyam102 (06:04:35): > @rohitsatyam102 has joined the channel
2020-12-09
Federico Marini (11:59:33): > Not really abioc git thing but still: - File (PNG): image.png
Federico Marini (12:00:31): > Not so sure how useful this could be - I like that discussions start with a “minimal effort” to formulate an issue
Federico Marini (12:01:33): > Or is this their attempt to phagocitate gitter?
Alan O’C (12:02:37): > What does it look like?
Federico Marini (12:03:29): > no clue, did not try yet (#IShouldBeWriting)
Kevin Rue-Albrecht (12:03:58): > I’ve just enabled it forhttps://github.com/kevinrue/BiocChallenges/discussions/24
Federico Marini (12:04:51): > Looks like a simple running chat?
Kevin Rue-Albrecht (12:05:22): > might be the right kind of feature for that purpose, might not, worth a shot I guess
Federico Marini (12:05:37): > ah no, it is indeed with the possibility of more Discussions items
Kevin Rue-Albrecht (12:05:41): > lol - File (PNG): image.png
Federico Marini (12:05:50): - File (PNG): image.png
Kevin Rue-Albrecht (12:06:37): > What the hell is a ‘show and tell’ category?
Federico Marini (12:07:48): > Aha:https://en.wikipedia.org/wiki/Show_and_tell_(education) - Attachment: Show and tell (education) > Show and tell (sometimes called show and share) is the practice of showing something to an audience and describing it to them, usually a toy or other children’s-oriented item. In the United Kingdom, North America, New Zealand and Australia, it is a common classroom activity in early elementary school. It is used to teach young children about public speaking. For example, a child will bring an item from home and will explain to the class why they chose that particular item, where they got it, and other relevant information.
Kevin Rue-Albrecht (12:08:26): > so it’s a journal club?:stuck_out_tongue:
Federico Marini (12:08:30): > I want so that when my kids will attend such sessions they will say “My papa calls this a vignette”
Kevin Rue-Albrecht (12:08:51) (in thread): > you got the better analogy!
Alan O’C (13:35:42): > So it is basically gitter? I’ve barely used gitter but it seems like just a discussion forum
Alan O’C (13:37:25): > Not sure if I prefer that over people opening issues just so they can ask a question. At least with the latter I can just palm them off to StackOverflow and say that an issue isn’t the right place to ask:sweat_smile:
2020-12-13
Huipeng Li (20:36:16): > @Huipeng Li has joined the channel
2020-12-14
Thomas Naake (08:55:37): > @Thomas Naake has joined the channel
2020-12-17
Manojkumar Selvaraju (11:11:33): > @Manojkumar Selvaraju has joined the channel
2020-12-18
Leonardo Collado Torres (19:56:00): > Hi Nitesh, > > I encountered a new type of error (for me) which I’m guessing might be due to some permissions. I made a commit that I was able to push to bioconductor’s git on themasterbranch for thebrainflowprobespackage. However, once I cherry-picked that commit intoRELEASE_3_12and resolved the merge, I’m not able to push it to bioconductor (I did push it to GitHub though). Here’s the error: > > $ git push upstream > Enumerating objects: 21, done. > Counting objects: 100% (21/21), done. > Delta compression using up to 4 threads > Compressing objects: 100% (10/10), done. > Writing objects: 100% (11/11), 2.48 KiB | 2.48 MiB/s, done. > Total 11 (delta 8), reused 0 (delta 0), pack-reused 0 > remote: FATAL: W refs/heads/RELEASE_3_12 packages/brainflowprobes l.collado-torres DENIED by fallthru > remote: error: hook declined to update refs/heads/RELEASE_3_12 > To git.bioconductor.org:packages/brainflowprobes.git > ! [remote rejected] RELEASE_3_12 -> RELEASE_3_12 (hook declined) > error: failed to push some refs to 'git@git.bioconductor.org:packages/brainflowprobes.git' >
2020-12-19
Martin Morgan (11:06:28): > @Leonardo Collado Torresplease try again; the permissions file had been edited to provide you with access to the devel but not release branch.
Nitesh Turaga (23:23:19): > Thanks@Martin MorganFor taking a look at this.
Nitesh Turaga (23:24:28): > @Leonardo Collado TorresThe easiest way to deal with some of these issues is if you send an email tomaintainer@bioconductor.orgor bioc-devel. It’s easy to miss things on slack because of the number of channels.
Leonardo Collado Torres (23:57:03): > Thanks Martin for fixing the permissions and ok, I’ll use bioc-develop next time then Nitesh. Have a good weekend!
2020-12-23
Kozo Nishida (11:45:56): > @Kozo Nishida has joined the channel
Kozo Nishida (11:50:32): > Hi Nitesh, > I could not fetch upstream ofhttps://github.com/Bioconductor/Contributions/issues/1825(And for that reason, I can not fix the warning status.) > > kozo2@DESKTOP-9IP2NJT:~/Projects/AHPathbankDbs$ cat ~/.ssh/id_ed25519.pub > ssh-ed25519 AAAAC3NzaC1lZDI1NTE5AAAAIIyTixKuuRgkDs4Nm5lQYUfxCOhXERAEFBn6dB69xMto kozo2@DESKTOP-9IP2NJT > kozo2@DESKTOP-9IP2NJT:~/Projects/AHPathbankDbs$ curl[https://github.com/kozo2.keys](https://github.com/kozo2.keys)|grep AAAAC3NzaC1lZDI1NTE5AAAAIIyTixKuuRgkDs4Nm5lQYUfxCOhXERAEFBn6dB69xMto > % Total % Received % Xferd Average Speed Time Time Time Current > Dload Upload Total Spent Left Speed > 100 486 100 486 0 0 601 0 --:--:-- --:--:-- --:--:-- 600 > ssh-ed25519 AAAAC3NzaC1lZDI1NTE5AAAAIIyTixKuuRgkDs4Nm5lQYUfxCOhXERAEFBn6dB69xMto > kozo2@DESKTOP-9IP2NJT:~/Projects/AHPathbankDbs$ git remote -v > origin git@github.com:biopackathon/AHPathbankDbs.git (fetch) > origin git@github.com:biopackathon/AHPathbankDbs.git (push) > upstream git@git.bioconductor.org:packages/AHPathbankDbs.git (fetch) > upstream git@git.bioconductor.org:packages/AHPathbankDbs.git (push) > kozo2@DESKTOP-9IP2NJT:~/Projects/AHPathbankDbs$ git fetch --all > Fetching origin > Warning: Permanently added the RSA host key for IP address '52.69.186.44' to the list of known hosts. > Fetching upstream > git@git.bioconductor.org: Permission denied (publickey). > fatal: Could not read from remote repository. > > Please make sure you have the correct access rights > and the repository exists. > error: Could not fetch upstream > > What should I check?
Nitesh Turaga (11:53:30): > Usessh-rsa
Nitesh Turaga (11:53:40): > It’ll work.
Kozo Nishida (11:54:32): > Oh I see! Thanks!
Kozo Nishida (11:58:09): > I would like to send pull-request tohttps://bioconductor.org/developers/how-to/git/new-package-workflow/(about ssh type of the key). Is the document managed onhttps://github.com/Bioconductor/bioconductor.org?
Nitesh Turaga (11:58:35): > Please do. Thank you for you contribution.
Nitesh Turaga (11:59:03): > Let me know if using the RSA type key worked for you.
Kozo Nishida (12:04:30): > It doesn’t work yet. > Is it because I have to wait for a while? > > kozo2@DESKTOP-9IP2NJT:~/Projects/AHPathbankDbs$ curl[https://github.com/kozo2.keys](https://github.com/kozo2.keys)|grep ssh-rsa > % Total % Received % Xferd Average Speed Time Time Time Current > Dload Upload Total Spent Left Speed > 100 1039 100 1039 0 0 3073 0 --:--:-- --:--:-- --:--:-- 3073 > ssh-rsa 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 > kozo2@DESKTOP-9IP2NJT:~/Projects/AHPathbankDbs$ cat ~/.ssh/id_rsa.pub > ssh-rsa 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 kozo2@DESKTOP-9IP2NJT > kozo2@DESKTOP-9IP2NJT:~/Projects/AHPathbankDbs$ git remote -v > origin git@github.com:biopackathon/AHPathbankDbs.git (fetch) > origin git@github.com:biopackathon/AHPathbankDbs.git (push) > upstream git@git.bioconductor.org:packages/AHPathbankDbs.git (fetch) > upstream git@git.bioconductor.org:packages/AHPathbankDbs.git (push) > kozo2@DESKTOP-9IP2NJT:~/Projects/AHPathbankDbs$ git fetch --all > Fetching origin > Fetching upstream > git@git.bioconductor.org: Permission denied (publickey). > fatal: Could not read from remote repository. > > Please make sure you have the correct access rights > and the repository exists. > error: Could not fetch upstream >
Kozo Nishida (13:16:22): > I can’t git fetch --all yet. > Maybe it’s not due to the ssh key type.
Nitesh Turaga (13:17:06): > Do you have a BiocCredentials ID?
Kozo Nishida (13:19:43): > I’m sorry, I didn’t know about it. Let me check…
Nitesh Turaga (13:20:16): > Has the package been submitted to Bioconductor or accepted?
Kozo Nishida (13:22:30): > Yes submittedhttps://github.com/Bioconductor/Contributions/issues/1825But not accepted yet
Kozo Nishida (13:26:58): > I cangit fetch --allnow. > It seems that the reason was that the Bioconductor Git Credentials and the GitHub Credentials were out of sync.
Nitesh Turaga (13:27:22): > ok, great
Nitesh Turaga (13:27:44): > All your keys are still ed25519 type though.
Kozo Nishida (13:28:08): > I manually added my public SSH key (it’s ed25519 not rsa) in the Bioconductor Git Credentials web UI and now I can fetch it successfully.
Nitesh Turaga (13:28:37): > ok great.
Nitesh Turaga (13:29:25): > That’s fine. It seems ed25519 is supported as well.
Kozo Nishida (13:34:02): > Thanks Nitesh. I didn’t fully understand what Bioconductor Git Credentials does (now I understand it). I will try to spread this (How to use Bioconductor Git Credentials) to Japanese people.
Nitesh Turaga (13:38:49): > Thanks for your help Kozo. > > So the BiocCredentials app is supposed to help people know what packages they have access to, and directly access their keys and change them. It gives you access to the unique identifier (in your case “kozo2”) and SSH keys tied to it, along with the email address. What you are seeing is basically one entry from the database we keep to match Github credentials to packages.
Kozo Nishida (14:01:28): > Sorry there is still a problem. I can’t push theoriginupdates toupstream > > kozo2@DESKTOP-9IP2NJT:~/Projects/AHPathbankDbs$ git fetch --all > Fetching origin > Fetching upstream > From git.bioconductor.org:packages/AHPathbankDbs > * [new branch] master -> upstream/master > kozo2@DESKTOP-9IP2NJT:~/Projects/AHPathbankDbs$ git merge upstream/master > Already up to date. > kozo2@DESKTOP-9IP2NJT:~/Projects/AHPathbankDbs$ git merge origin/main > Already up to date. > kozo2@DESKTOP-9IP2NJT:~/Projects/AHPathbankDbs$ git push upstream master > error: src refspec master does not match any > error: failed to push some refs to 'git@git.bioconductor.org:packages/AHPathbankDbs.git' > kozo2@DESKTOP-9IP2NJT:~/Projects/AHPathbankDbs$ git fetch --all > Fetching origin > Fetching upstream > kozo2@DESKTOP-9IP2NJT:~/Projects/AHPathbankDbs$ git push origin main > Everything up-to-date > kozo2@DESKTOP-9IP2NJT:~/Projects/AHPathbankDbs$ git push upstream main > Enumerating objects: 8, done. > Counting objects: 100% (8/8), done. > Delta compression using up to 6 threads > Compressing objects: 100% (4/4), done. > Writing objects: 100% (6/6), 1.32 KiB | 1.32 MiB/s, done. > Total 6 (delta 4), reused 4 (delta 2) > remote: FATAL: W refs/heads/main packages/AHPathbankDbs kozo2 DENIED by fallthru > remote: error: hook declined to update refs/heads/main > To git.bioconductor.org:packages/AHPathbankDbs.git > ! [remote rejected] main -> main (hook declined) > error: failed to push some refs to 'git@git.bioconductor.org:packages/AHPathbankDbs.git' >
Kozo Nishida (14:02:41): > Is this because the default branch name on GitHub has changed frommastertomain…?
Nitesh Turaga (14:04:33): > Can you trygit push upstream master?
Kozo Nishida (14:05:29): > Yes. It failed like this > > kozo2@DESKTOP-9IP2NJT:~/Projects/AHPathbankDbs$ git push upstream master > error: src refspec master does not match any > error: failed to push some refs to 'git@git.bioconductor.org:packages/AHPathbankDbs.git' >
Nitesh Turaga (14:05:59): > There is no branch calledmasteron your local git repo.
Nitesh Turaga (14:06:11): > So, you’ll have to do something like…
Nitesh Turaga (14:06:39): > > git checkout main > git checkout -b master > git push upstream master >
Kozo Nishida (14:07:37): > Thanks it works well > > kozo2@DESKTOP-9IP2NJT:~/Projects/AHPathbankDbs$ git checkout main > Already on 'main' > Your branch is up to date with 'origin/main'. > kozo2@DESKTOP-9IP2NJT:~/Projects/AHPathbankDbs$ git checkout -b master > Switched to a new branch 'master' > kozo2@DESKTOP-9IP2NJT:~/Projects/AHPathbankDbs$ git push upstream master > Enumerating objects: 8, done. > Counting objects: 100% (8/8), done. > Delta compression using up to 6 threads > Compressing objects: 100% (4/4), done. > Writing objects: 100% (6/6), 1.32 KiB | 1.32 MiB/s, done. > Total 6 (delta 4), reused 4 (delta 2) > To git.bioconductor.org:packages/AHPathbankDbs.git > f5c2cb4..c81cb30 master -> master >
Kozo Nishida (14:10:31): > I didn’t know that I couldn’t push without making the branch names same.
Nitesh Turaga (14:12:02): > The branch names have to make sense…Either you have to connect your branchmainto themasterbranch on git.bioc or you have to make a master branch yourself.
Nitesh Turaga (14:14:19): > You can set your default branch tomasteron your Github too.
Kozo Nishida (14:15:23): > I see. I think the branch name manipulation is difficult, so it may be better to recommend using master instead of main as the default branch name when developing (or starting) a Bioconductor project… (for now)
Martin Morgan (14:34:49): > I think you cangit push upstream main:master, I would guess there is git magic to have to only do this once…
Kozo Nishida (14:42:12): > Thanks@Martin MorganI learned one more thing about Git… > I don’t have updates in my local repo yet, but I think it’s working fine. > > kozo2@DESKTOP-9IP2NJT:~/Projects/AHPathbankDbs$ git branch -a > * main > master > remotes/origin/HEAD -> origin/main > remotes/origin/main > remotes/upstream/master > kozo2@DESKTOP-9IP2NJT:~/Projects/AHPathbankDbs$ git push upstream main:master > Everything up-to-date >
Kozo Nishida (16:14:54): > I sent a pull-request aboutmain``masterdifferencehttps://github.com/Bioconductor/bioconductor.org/pull/79/files
2020-12-24
Martin Morgan (14:36:22) (in thread): > Thanks, I’ve merged the pull request. I think the topic needs a more general treatment, and I draftedhttps://github.com/Bioconductor/bioconductor.org/compare/master-main; please comment on that branch (kozo, or anyone else…)
2020-12-27
Aaron Lun (02:18:13): > Indeed, I have a little script that strips outmainand replaces it with good ol’master
Aaron Lun (02:19:31): > not that it really matters, I never let GitHub make the default branch choice for me anyway; I’m always pushing from pre-initialized Git repos that default tomaster
2021-01-22
Annajiat Alim Rasel (15:41:51): > @Annajiat Alim Rasel has joined the channel
2021-01-26
Dario Righelli (11:03:39): > Hi, I’m having a warning running theBiocCheck()on my local repo > > "The following files are over 5MB in size: '.git/objects/pack/pack-f0d1d177387fca4bc9767cde531008381a79178f.pack'" > > I tried to solve it by following this guide, but it doesn’t solve the warning.http://bioconductor.org/developers/how-to/git/remove-large-data/Any ideas? > thanks!
Marcel Ramos Pérez (11:06:16): > You may want to adjust the BFG--strip-blobs-bigger-thanflag.. also I would recommend that you have a backup clone of your repo just in case things go wrong
Dario Righelli (11:11:03): > yes I used 5M instead of 100M for that flag, it seems to do something, but the warning is still there…
Dario Righelli (11:12:40): > Also it seems to not working on different branches than master (I don’t know if it’s the right behaviour)
Marcel Ramos Pérez (11:12:57): > What package is it for?
Dario Righelli (11:13:07): > SpatialExperiment ^^’
Lukas Weber (11:22:25): > @Lukas Weber has joined the channel
Marcel Ramos Pérez (11:51:14): > I’m not sure BFG cleans.packfiles only large binary blobs.. you might have to usegit-filter-repohttps://github.com/newren/git-filter-repofor which you may get some savings with their Auto shrink feature
Nitesh Turaga (11:53:23): > @Dario RighelliAre these changes on your github repo? Mind sending me a link?
Dario Righelli (11:54:07) (in thread): > https://github.com/drighelli/SpatialExperiment
Dario Righelli (11:54:12) (in thread): > thanks!
Dario Righelli (11:54:21) (in thread): > thanks I’ll try…
Nitesh Turaga (12:40:13): > Hi@Dario Righelli, there seem to be a.packfile of 7.7MB after you use BFG to clean the repo. This is what is causing the second issue.
Dario Righelli (12:40:36): > oh ok!
Dario Righelli (12:40:54): > Indeed on my local machine I have two bigpackfiles…
Dario Righelli (12:43:16): > okok after pulling from github I just have one of 53M!
Nitesh Turaga (12:43:55): > Right, it’s because you didn’t push your local repo up to your github after cleaning with BFG.
Nitesh Turaga (12:44:01): > You have to do a force push
Dario Righelli (12:45:09): > ok, but I still have a bigpackfile, right?:slightly_smiling_face:
Dario Righelli (12:45:56) (in thread): > ok I try
Nitesh Turaga (12:47:15): > The 7.7MB one…
Nitesh Turaga (12:48:05): > yes you do.
Nitesh Turaga (12:55:57): > ok, so, i’m trying to figure out which of your files is causing the .pack file to be that large. You can go from.packor.idx–> commit id –> file to investigate. As an example > > ## FIGURE OUT 5 LARGEST COMMITS > $$ nturaga-instance:~/SpatialExperiment$ git verify-pack -v .git/objects/pack/pack-4c317a984fb5a102ce55be6c328865539744b30d.idx | sort -k 3 -n | tail -5 > b4204976e2eb7fd302111b4153fe80b78a457875 blob 902095 201170 1128407 > 60067d9575f6b4b9370202bbb1eebfab52f58bb0 blob 973309 170451 1335975 > 8830919963074121448f53b81216c12d8a530a0d blob 1301288 1284262 6568423 > cecae85479c47ee3f12c25782d962b06e4c42e54 blob 2231149 2226422 4342001 > 7302613041e96c25827d01e8794afb02ab568323 blob 2964387 590778 3132957 > > ## Search each commit to see which file is that big > $$ nturaga-instance:~/SpatialExperiment$ git rev-list --objects --all | grep \ b4204976e2eb7fd302111b4153fe80b78a457875 > b4204976e2eb7fd302111b4153fe80b78a457875 inst/extdata/seqFISH/seqfish_normalized_cortex_b2_testing.txt > > $$ nturaga-instance:~/SpatialExperiment$ git rev-list --objects --all | grep \ 60067d9575f6b4b9370202bbb1eebfab52f58bb0 > 60067d9575f6b4b9370202bbb1eebfab52f58bb0 inst/extdata/seqFISH/tasic_scRNAseq_training_b2.txt >
Nitesh Turaga (12:56:07): > You can go about this method to investigate more.
Dario Righelli (13:02:44): > ok thanks, and by removing the oldest commits “cleaning the tree” I should be able to shrink this pack, right?
Nitesh Turaga (13:03:53): > At this point it’s just investigating the tree, and figuring it out to shrink the size as far as I know. Maybe other git fans can step in here and fill in the blanks for us?
Dario Righelli (13:08:51): > thanks again!
2021-01-27
Dario Righelli (08:11:25) (in thread): > Hi Nitesh, I tried multiple times but it always says:>SpatialExperiment % git push -f origin``Everything up-to-date
2021-01-30
Aaron Lun (22:19:58): > @Martin MorganI would like to move all of the new OSCA.* subbooks into the BioC GH org.
2021-02-16
Kozo Nishida (07:50:54): > I transferred a GitHub repository after it was accepted. > (https://github.com/Bioconductor/Contributions/issues/1825is the package.) > Do I need to contact Bioconductor core team in such case?
Nitesh Turaga (08:11:53): > The set up of the Github repo is now up to you@Kozo Nishida. Bioconductor only deals with the package in the Bioconductor private git repo. Everything else outside of it, is your responsibility.
Kozo Nishida (08:43:43) (in thread): > Thank you@Nitesh Turaga. > I understand.
2021-02-17
abdullah hanta (16:05:29): > @abdullah hanta has joined the channel
2021-02-26
Krithika Bhuvanesh (13:19:59): > @Krithika Bhuvanesh has joined the channel
Nitesh Turaga (13:20:26): > @Krithika BhuvaneshFeel free to ask your questions on this channel:smile:Welcome.
Krithika Bhuvanesh (13:25:26): > Hello, We are the package maintainer for the CINdex Bioc package. 5 years ago when we our package first got accepted into BioC, it was before the Github integration. We have made changes in our package to address the build errors for the upcoming release. I have now completed Steps# 1 - 7 and Step # 9 from here:http://bioconductor.org/developers/how-to/git/sync-existing-repositories/(I think Step # 8 does not apply to me). > > Am I required to do Step # 10 ? I’m not clear on what a “release branch” is. Should i say this ? Pls advise. Thank you > > git checkout RELEASE_13_3 ? >
Nitesh Turaga (13:26:32): > There is no RELEASE_3_13 as of right now. It is just yourmasterbranch (this is the development branch).
Nitesh Turaga (13:28:14): > So, there are a couple of things here. It would help understand your sitaution better if you shared your github repo link.
Krithika Bhuvanesh (13:28:33): > Sure. The latest code is herehttps://github.com/ICBI/CINdex
Nitesh Turaga (13:30:48): > May I suggest a whole new approach for you? It would involve delete your current Github repo?
Nitesh Turaga (13:30:56): > A little radical, maybe.
Nitesh Turaga (13:31:27): > You first need to follow this,http://bioconductor.org/developers/how-to/git/maintain-github-bioc/.
Nitesh Turaga (13:32:34): > ^ this approach is a lot easier than the one you are taking.
Nitesh Turaga (13:34:10): > save the changes your made in your current repo so you don’t lose them.
Krithika Bhuvanesh (13:34:26): > So in this new approach. - I have to do the following. Is that correct ? > * -delete old repo > * clone existing CINdex package from BioC , and follow all steps fromhttp://bioconductor.org/developers/how-to/git/maintain-github-bioc/. > * Then do I commit/push as usual ?
Nitesh Turaga (13:35:08): > Yes, but save a copy of the old repo in your local machine.
Krithika Bhuvanesh (13:35:48) (in thread): > Okay, I can rename that.
Nitesh Turaga (13:36:02): > Perfect.
Krithika Bhuvanesh (13:36:59): > But just curios, is there a difference between the two ? I thought the steps I did worked okay ? Was there something I had missed ?
Nitesh Turaga (13:50:32): > So, your CINdex repo, had a few merge markers in them, so you wouldn’t actually be able to “push” that togit.bioconductor.org
Krithika Bhuvanesh (16:28:10): > Hi@Nitesh TuragaI followed all steps fromhttp://bioconductor.org/developers/how-to/git/maintain-github-bioc/andhttp://bioconductor.org/developers/how-to/git/push-to-github-bioc/. I updated 3 files in the github repo. When i didgit push upstream master, it gave me this error : > > remote: Error: You cannot push without resolving merge conflicts. > remote: > remote: Please check the files in the commit pushed to the git-server > remote: for merge conflict markers like '<<<<<<<', '========', '>>>>>>>'. > remote: > To git.bioconductor.org:packages/CINdex.git > ! [remote rejected] master -> master (pre-receive hook declined) > error: failed to push some refs to 'git@git.bioconductor.org:packages/CINdex.git' > > Should i do git merge upstream to resolve conflicts ? Pls advise
Nitesh Turaga (17:14:36): > You have to clear your merge conflicts.
Nitesh Turaga (17:14:43): > You cannot push with merge conflicts.
Krithika Bhuvanesh (17:20:38) (in thread): > But when I try git status, it says this. I’m confusedOn branch master``Your branch is up to date with 'origin/master'.``nothing to commit, working tree clean
Krithika Bhuvanesh (17:27:54): > No matter what merge command I try fromhttps://bioconductor.org/developers/how-to/git/resolve-conflicts/, it saysup to date, Seems its not able to find any conflicts. Not sure what to do
Martin Morgan (17:42:08): > The ‘merge commits’ are lines likehttps://github.com/ICBI/CINdex/blob/08f73b91cb8aa44b52c619522b7ac45c17b5178d/vignettes/PrepareInputData.Rmd#L43-L57where you’ve made commits to different branches that can’t be resolved automatically. You have to resolve them manually by removing the lines that you want to remove; maybe you’d resolve > > <<<<<<< HEAD > library(TxDb.Hsapiens.UCSC.hg18.knownGene) > library(Homo.sapiens) > > > {seg} > #Load example segment data into the workspace. > data("grl.data") > ======= > > {seg} > #Load example segment data into the workspace. > data("grl.data") > >>>>>>> upstream/master > > to > > library(TxDb.Hsapiens.UCSC.hg18.knownGene) > library(Homo.sapiens) > > > {seg} > #Load example segment data into the workspace. > data("grl.data") > > There are several of these conflicts that need resolving in the vignette, which I found by cloning your repository and grepping for them > > CINdex master$ grep -r "====" * > vignettes/PrepareInputData.Rmd:======= > vignettes/PrepareInputData.Rmd:======= > vignettes/PrepareInputData.Rmd:======= > vignettes/PrepareInputData.Rmd:======= > vignettes/PrepareInputData.Rmd:======= >
Krithika Bhuvanesh (21:46:41) (in thread): > Thanks for the explanation. Will try this out
Krithika Bhuvanesh (22:44:50): > Thank you so much for the help. I was able to push to master . Just to confirm - are these the next steps for me ? > * Since there is no release at this time, no need to push to RELEASE branch. > Question- Now that I have pushed to upstream master, will the push to release get done automatically for the release ? > * I will to check BioC devel report after a few days to see if the build errors are gone. > Is there anything else I need to do ? > Thank you very much
2021-02-27
Martin Morgan (09:33:07): > At the next release, ‘we’ will create a new branch RELEASE_3_13. > > You will then be responsible for updating that branch, if necessary, by pulling the branch into your local repository, creating a new local branch, committing changes to the local branch, and pushing those changes back to the RELEASE_3_13 branch on Bioconductor. You would only make ‘bug fixes’ to the RELEASE_3_13 branch; all new features etc would be introduced on ‘master’.
2021-02-28
Krithika Bhuvanesh (15:45:18): > Thank you
2021-03-20
watanabe_st (01:56:34): > @watanabe_st has joined the channel
2021-03-23
Lambda Moses (23:04:25): > @Lambda Moses has joined the channel
2021-03-31
Lisa Cao (12:49:54): > @Lisa Cao has joined the channel
2021-04-13
Nicholas Cooley (14:47:41): > I have an issue where the bioc build server seems to be building on a commit that is not the most recent: > > My most recent commit according to git log: > > commit 3ca4e207f3599453b748073b9772f05381286b02 (HEAD -> master, upstream/master, origin/master, origin/HEAD) > Author: npcooley <npcooley@gmail.com> > Date: Mon Apr 12 15:10:35 2021 -0400 > > whereas, the build server seems to be building from this commit: > > commit 099e01000afe56f30cfbd29276a2d3196e6ea74e > Author: Nicholas Cooley <nicholascooley@nicholass-mbp.myfiosgateway.com> > Date: Fri Apr 2 13:53:21 2021 -0400 > > at least according to the most recent build report:http://bioconductor.org/checkResults/devel/bioc-LATEST/SynExtend/I assume this has something to do with the author change, but the author line has changed across computers before?
Kasper D. Hansen (14:51:52): > Did you bump the version number?
Lori Shepherd (14:56:27): > also remember it can take up to 24 hours to appear on the build report. You just pushed yesterday and if it was after the builders did their git pull from the git repository it might not be on until tonights build
Nicholas Cooley (14:59:35): > The version number is bumped, and there was another push in between that should have registered a day or 2 ago? The version the build servers are working on is 1.3.7 where the repository should be 1.3.9? At least git diff tells me all there are no differences.
Lori Shepherd (15:02:34): > The push on Sunday April 11 would have been too late to register for Monday’s build and will probably be reflected in today’s build once its posted – The pushes you made yesterday were most likely too late to be pulled and reflect it today’s build and will be reflected in tomorrow’s build – see the top of this page:http://bioconductor.org/developers/how-to/troubleshoot-build-report/
Nicholas Cooley (15:03:10): > Wait, there are no weekend builds?
Nicholas Cooley (15:04:21): > ok, i’ll wait for the new build reports!
Lori Shepherd (15:05:20): > there are weekend builds for software - software are everyday - but the builders only build the packages once per day and only pull from git once per day
2021-04-23
Ashwini Shanmugam (04:14:30): > @Ashwini Shanmugam has joined the channel
Ashwini Shanmugam (04:15:26): > @Ashwini Shanmugam has left the channel
2021-04-28
Mateusz Staniak (18:16:47): > @Mateusz Staniak has joined the channel
2021-04-30
Gabriel J. Odom (20:03:33): > Hi all, is there a way for me to request that a package be added to my Bioconductor Git Credentials? My master’s student and I are currently submitting a package (coMethDMR), but I do not have access to push changes upstream. Is this because only one person can be a maintainer at a time? Thanks in advance.
2021-05-04
Leonardo Collado Torres (15:19:33) (in thread): > @Gabriel J. Odomif you send an email to bioc-devel asking for this, they’ll add you. Remember to cc the current maintainer + the package name and your BioC git credentials username
Gabriel J. Odom (15:20:21) (in thread): > Thank you!
2021-05-05
Gabriel J. Odom (18:50:19): > Hi all, I have another git question (package coMethDMR). We are using BFG to remove some data files from the git history, and all is well until we attempt to push these changes upstream to Bioconductor. git refuses the push because upstream has the data sets that we have deleted from the .git/ folder. The git error message recommends that we fetch first, but this adds the data sets back, and we are back to square 1. > I know that I am doing something stupid. Does anyone have any ideas? > Thanks
Spencer Nystrom (22:22:20) (in thread): > You’ll need to message the mailing list and have someone on the bioc end fix the upstream remote.
2021-05-07
Lukas Weber (08:52:23): > Hi all, I also have a question about BFG repo cleaner. We have some large files in git history that can be removed, but these are all in non-core branches (i.e. notmaster,RELEASE_3_12, or any other branches used by Bioconductor).BiocCheck::BiocCheck()still detects these large files in history and gives a warning (although only on local checks, not on the build reports). So my question is: if we use BFG repo cleaner to remove the large files from history and modify commit hashes in non-core branches on GitHub, does this affect the Bioconductor git repo, and do we need to contact the Bioc team? i.e. does Bioconductor git store thewholerepo (all branches) from GitHub, or only the master and release branches (in which case our change would not affect Bioconductor)?
Alan O’C (09:06:21) (in thread): > If they were pushed after the repo was accepted to Bioc then there’s no way Bioc would have picked the commits up. If they were there before, I guess it depends on how the repo is cloned from github to bioc but I imagine the same holds
Mike Smith (09:13:35) (in thread): > ~~~I think it also depends on whether you merged those other branches into one of the BioC branches. AFAIK no package on the BioC server has branches other than ~~~~~RELEASE_X_Ymastermaster
Lukas Weber (09:15:27) (in thread): > Ok thanks - it’s a completely separate branch, although I’m not sure if we created it before or after submission. BFG repo cleaner doesn’t change any of the commits inmasterorRELEASE_X_Y.
Mike Smith (09:17:22) (in thread): > I just read the bit about it not appearing on the build report. This package has been accepted and is part of Bioc, right? Because I don’t think the nightly builder runs BiocCheck, only the single package builder does as part of the submission review.
Lukas Weber (09:20:13) (in thread): > Ah ok - yes, it’s part of Bioc release and devel (package isSpatialExperiment). Ok maybe that explains why I get the warning on localBiocCheckbut not on the build report. I’m slightly concerned we may be wasting some space in the Bioc git repo. Alternatively if the large files in history are only on GitHub, then it is easy to fix and doesn’t need contacting the Bioc team.
Mike Smith (09:35:04) (in thread): > Maybe it’s sufficient to clone directly from the Bioc server to a temporary location. Then run the BFG cleaner on that copy to see if it picks anything up. If it doesn’t then you’re fine, if it does then you can explore removing them. > > FWIW there’s a couple of software packages with repos over 700MB!
Lukas Weber (09:35:33) (in thread): > ah yes good idea
Lukas Weber (09:39:13) (in thread): > Yes if I clone the whole repo (git clone --mirror) fromgit@git.bioconductor.orgit is only 1.7M, while if I get it from GitHub it is 177M. So I think this means the Bioc git repo does not have the other branches. Thanks!
Lukas Weber (09:42:01) (in thread): > Yeah the clone fromgit@git.bioconductor.orgcontains only branchesmasterandRELEASE_3_12. The one from GitHub contains the additional branches. So we should be able to modify it easily on GitHub without affecting Bioconductor. Thanks!
2021-05-11
Megha Lal (16:44:00): > @Megha Lal has joined the channel
2021-05-12
Devon Kohler (18:07:16): > @Devon Kohler has joined the channel
Devon Kohler (18:07:33): > Hi all, I have a question about maintainer changes for bioconductor packages. I am trying to push an update for the 3.13 release but I am not currently the package maintainer. The current maintainer does not have access to a computer the rest of the day, but can send an email asking to add me. What is the fastest way we can swap maintainers so that the update can go through for the version 3.13 deadline today?
Wancen Mu (18:27:31): > @Wancen Mu has joined the channel
Meena Choi (18:45:07): > @Meena Choi has joined the channel
Lori Shepherd (20:06:42): > Technically the deadline today is for adding new packages to the release. Maintainers can continue to push changes to their packages at any time. The deadline for passing R CMD build and check this Friday is mostly important for identifying packages that should be deprecated. If we know a maintainer is actively working on a package we will not deprecate.That said, please have the original maintainer respond on the email you sent to the mailing list requesting you have access. > We also recommend changes by this Friday so that changes can be accurately tested and reflected on they build report before release. If errors exist on tht devel branch into the release maintainers have the added burden of correcting on both branches.
Devon Kohler (20:23:42): > That sounds great, thanks for the info Lori. The maintainer actually just sent a new email referencing my email address.
Janani Ravi (22:21:20): > @Janani Ravi has joined the channel
2021-05-17
Nicholas Cooley (11:19:48): > I have a small question that I don’t seem to be able to find an answer to: when preparing for the new release, are the DESCRIPTION and NEWS files automatically updated by Bioconductor to reflect release, or do I need to edit those in one last commit?
Lori Shepherd (11:23:53): > The description version number is automatically adjusted at the release time by the core team so there is nothing you need to do for that. NEWS files are not automatically updated by Bioconductor. We scrap the NEWS files of packages to create the release announcement but otherwise do not alter those. Also any news articles that you want in the release announcement should be committed to git by tomorrow (for the freeze to create new git branches for RELEASE_3_13 on wed – also done ongit.bioconductor.orgautomatically for packages but maintainers would have to make their own github adjustments)
2021-05-19
Leonardo Collado Torres (19:19:54): > Hi channel! > > Lori created an issue forbiocthisin relation to version controlling thedev/*.Rscripts thatbiocthis::use_bioc_pkg_templates()creates. This new issue reminded me ofhttps://community-bioc.slack.com/archives/C6MVC96AZ/p1604600559030500which I think was well discussed, but where we didn’t reach a conclusion. Maybe this has since been discussed at the TAB or BioC-core or some other group. > > Anyway athttps://github.com/lcolladotor/biocthis/issues/21I linked back to this channel and that discussion we had. - Attachment: Attachment > Hi! > > https://github.com/Bioconductor/bioconductor.org/commit/9c7cabcb00170eb9d24a9597b6dd4c5ede588268 was brought to my attention today, which I didn’t know about. I’m a bit confused on why files such as ./github/workflows/something.yml would not be allowed on the default branch of a GitHub repo that gets submitted to Bioc. > > This seems to disable using GitHub Actions on Bioc packages as well as having template GitHub files under .github such as https://github.com/LieberInstitute/recount3/tree/master/.github for example. I mean, I could make a copy of the default branch and add the GHA workflow there, but it seems unnecessary. File size-wise, these are small files that can be ignored by .Rbuildignore such as https://github.com/LieberInstitute/recount3/blob/master/.Rbuildignore#L6. Otherwise, you have to keep 2 branches around and make sure they are in sync and all that (unless there’s a way to tell GHA to run a workflow from another branch). > > Similarly say adding a dev/ folder with some R scripts that is ignored by .Rbuildignore. I use those R scripts to include code that I need to re-run frequently. For example https://github.com/LieberInstitute/recount3/blob/master/dev/04_update.R. > > I do get the point that no big extra files should be on a repo. Like say, some rendered Rmarkdown HTML files. Is there an example that prompted this change? There might be use cases you’ve seen that I’m missing and that motivated this change.
Vince Carey (22:06:00): > Thanks for raising this up again@Leonardo Collado Torres. IMHO the core or TAB have not come to complete closure on this.@Marcel Ramos Pérezhas noted this approach using a distinct branch:https://github.com/lcolladotor/biocthis/issues/21#issuecomment-844614396… I hope we can come up with clear and satisfactory prescriptions soon.
2021-05-21
Leonardo Collado Torres (12:38:54): > Since Marcel and Lori replied on GitHub, I replied there too with: > * The suggestion to include a helper script / R function that goes through the.Rbuildignorecontents and removes those files. > * A list of non-standard R package files that I know of and/or use, plus a brief description of why I find them useful.
2021-05-24
Kasper D. Hansen (06:29:29): > I have to say that I disagree with this “Bioc git should only contain files relevant for the R package” perspective
Kasper D. Hansen (06:30:15): > From a historical perspective, back when we used SVN, we all used Bioc-SVN as the primary source code repository. This meant we have the R package but we also had TODO lists, notes and draft material in there.
Kasper D. Hansen (06:31:03): > So IMO we have a long history of hosting developer-related files which are not really part of the R package. I don’t see why we would want to go away from that.
Kasper D. Hansen (06:33:30): > I recognize that right now, most people use GitHub as their primary code/devel repos, but it is not clear to me how desirable that is long term. I understand that right now GH has tons of features we cannot provide, and perhaps that ship has sailed, but I do see a potential future benefit of returning to the old “bioc repos is main devel repos” eventually.
MarieV (13:42:35): > @MarieV has joined the channel
2021-05-26
Andres Wokaty (14:26:49): > @Andres Wokaty has joined the channel
2021-06-30
Ben Story (19:53:37): > @Ben Story has joined the channel
Ben Story (20:09:10): > For some reason my github public key isn’t working with the BioC git (see comment). Just trying to push some fixes to my package under review. Thanks.
Ben Story (20:10:05) (in thread): > git remote -v``origin git@github.com:benstory/mitoClone2 (fetch)``origin git@github.com:benstory/mitoClone2 (push)``upstream git@git.bioconductor.org:packages/mitoClone2.git (fetch)``upstream git@git.bioconductor.org:packages/mitoClone2.git (push)thengit fetch --all``Fetching origin``Fetching upstream``Permission denied (publickey).``fatal: Could not read from remote repository.``Please make sure you have the correct access rights``and the repository exists.``error: Could not fetch upstream
Spencer Nystrom (20:55:39) (in thread): > Did you submit your ssh key to:https://git.bioconductor.org/BiocCredentials/login/?next=/BiocCredentials/
Spencer Nystrom (20:56:04) (in thread): > (some additional detail here:https://www.bioconductor.org/developers/how-to/git/faq/)
2021-07-01
Ben Story (04:13:37) (in thread): > I tried to login/reset earlier but I think my e-mail is not currently associated with the package (i.e. my emailis not associated with a maintainer of a Bioconductor package. Although it is listed on the DESCRIPTIONhttps://github.com/Bioconductor/Contributions/issues/2158
Mike Smith (04:43:08) (in thread): > There seem to be quite a few people reporting git problems today on the devel mailing list:https://stat.ethz.ch/pipermail/bioc-devel/2021-June/018250.htmlNo idea if they’re all related, but perhaps there’s some underlying issue rather than your account specifically.
Nitesh Turaga (08:01:24) (in thread): > Please send us an email tobioc-devel@r-project.org
Nitesh Turaga (08:01:30) (in thread): > We like to keep a track of these things.
Ben Story (08:41:23) (in thread): > kk done
2021-07-19
Leo Lahti (17:00:12): > @Leo Lahti has joined the channel
2021-07-23
Batool Almarzouq (15:51:48): > @Batool Almarzouq has joined the channel
2021-08-04
TimNieuwenhuis (12:14:22): > @TimNieuwenhuis has joined the channel
2021-08-14
Sean Davis (11:13:01): > My GEOquery Bioconductor git repo is rejecting authentication with the my current ssh keys (likely due to my switching laptops and switching institutions….). Would it be possible to change .ssh key for GEOquery to be the same as, for example, my GenomicDataCommons package?
Nitesh Turaga (21:46:19): > Hi@Sean DavisI’ve fixed it and you should have access now.
2021-08-19
Henrik Bengtsson (08:12:41): > @Henrik Bengtsson has joined the channel
2021-08-31
Kevin Rue-Albrecht (07:39:48): > Hi all, just to point out that i’ve created#github-actionsto describe an issue that I’m having, without cross-posting it everywhere. > I also thought the channel could be useful as I haven’t found a channel here to specifically discuss issues about github actions for bioconductor packages.
2021-09-16
Henry Miller (18:35:14): > @Henry Miller has joined the channel
2021-09-25
Haichao Wang (07:20:25): > @Haichao Wang has joined the channel
2021-09-26
David Zhang (15:40:16): > @David Zhang has joined the channel
2021-10-15
Yagmur Simsek (12:19:42): > @Yagmur Simsek has joined the channel
2021-10-18
Nicholas Cooley (18:24:16): > Hi sorry if this has been answered, but when running BiocCheck and getting the error:Maintainer must add package name to Watched Tags on the support site; Edit your Support Site User Profile to add Watched Tags.Is the email / account that BiocCheck is looking for the one that is associated with the github account the package is homed under? Or the email used in theDESCRIPTIONfile? My github, and my bioconductor support site profile are under my personal gmail account, but my package uses my current academic email. Can I just change the email on my support site profile to match the package?
Alan O’C (18:27:28) (in thread): > It’ll be the email inDESCRIPTION. Github is just used for the package submission process; I don’t think Bioc ever really deals with your GH account beyond that. > > Not sure about changing emails on the support site but yeah I imagine that’s the way to go (or change the DESCRIPTION email, equivalently)
Nicholas Cooley (18:49:22) (in thread): > About how long should I expect BiocCheck to take to recognize the addition of a tag to a profile?
Alan O’C (18:52:13) (in thread): > No idea, but I imagine it does a http request so should be basically instant?
Marcel Ramos Pérez (19:14:34) (in thread): > You can try using something likeBiocCheck::checkWatchedTag("marcel.ramos@roswellpark.org", "TCGAutils")with the email in theDESCRIPTIONfile
2021-10-24
Sean Davis (17:10:05): > I’m trying to push a bug fix to GEOquery RELEASE_3_13. Here is what I get: > > $ git push bioc RELEASE_3_13Enumerating objects: 5, done. > Counting objects: 100% (5/5), done. > Delta compression using up to 8 threads > Compressing objects: 100% (3/3), done. > Writing objects: 100% (3/3), 307 bytes | 307.00 KiB/s, done. > Total 3 (delta 2), reused 0 (delta 0), pack-reused 0 > remote: FATAL: W refs/heads/RELEASE_3_13 packages/GEOquery sdavis2@mail.nih.gov DENIED by fallthru > remote: error: hook declined to update refs/heads/RELEASE_3_13 > To git.bioconductor.org:packages/GEOquery > ! [remote rejected] RELEASE_3_13 -> RELEASE_3_13 (hook declined) > error: failed to push some refs to 'git.bioconductor.org:packages/GEOquery' > > Any suggestions on what I need to do differently? I have been able to push tomainwithout problems.
Marcel Ramos Pérez (17:24:24): > Hi Sean, changes to 3.13 are no longer possible. The branch has been frozen.https://www.bioconductor.org/developers/release-schedule/
2021-10-26
Catherine Ross (11:48:32): > @Catherine Ross has joined the channel
Lori Shepherd (14:18:24): > we freeze the current release because of limited resources – the release 3.13 build machines become the next devel build machines and this involves some preparation time.
Marcel Ramos Pérez (14:20:37): > @Sean DavisThus the 1 - 2 week lag time between the freeze and the release. Thanks Lori!
Sean Davis (19:09:46) (in thread): > Yep. I knew there was a good reason–sorry if that came across as complaining. Thanks for the followup.
Sean Davis (19:11:07) (in thread): > And I really appreciate the clarity in communications around the release process. If I had been paying closer attention to those emails, I would have budgeted time differently–my bad there.
2021-11-08
Paula Nieto García (03:25:55): > @Paula Nieto García has joined the channel
2021-11-09
GuandongShang (21:35:02): > @GuandongShang has joined the channel
GuandongShang (21:35:49): > Hi. I developed a package in Bioc 3.13 and now its version is bumped by 1.0.0, but I am little confused about Git release branch and bump version of packages development. I am wondering whether someone can help me see if what I do is correct. > According to thehttps://bioconductor.org/developers/how-to/git/bug-fix-in-release-and-devel/, first, I sync my local repo with Github and Bioconductor > > git fetch --all > git checkout master > git merge upstream/master > git merge origin/master > > git checkout -b RELEASE_3_14 upstream/RELEASE_3_14 > git checkout RELEASE_3_14 > git merge upstream/RELEASE_3_14 > git merge origin/RELEASE_3_14 > > Then I modify some code in master branch > > git checkout master > # modify code > # bump version in DESCRIPTION into 1.1.1 > git add code file and DESCRIPTION > git commit -m "modify code" > > And then I switch to RELEASE_3_14 and cherry-pick the commits from master > > git checkout RELEASE_3_14 > git cherry-pick master > # Now the conflict appears in version section in DESCRIPTION file like this > > <<<<<<< HEAD > Version: 1.1.1 > ======= > Version: 1.0.0 > >>>>>>> upstream/master > > So I delete other line and only keep the Version 1.0.0 and bump it into 1.0.1. And I git add and commit > > git add DESCRIPTION > git commit -m "update RELEASE version number" > > Then I push my changes to Github and Bioconductor > > git checkout master > git push upstream master > git push origin master > > git checkout <RELEASE_X_Y> > git push upstream <RELEASE_X_Y> > git push origin <RELEASE_X_Y> >
2021-11-10
Arsenij Ustjanzew (03:55:33): > @Arsenij Ustjanzew has joined the channel
Marc Elosua (04:53:53): > @Marc Elosua has joined the channel
Federico Marini (05:22:06): > I think you are right@GuandongShang- the core concept is > * “new devels” go to the master devel branch (“only”) > * bug fixes should go to release as well > If you for example do a separate commit for the fix and one for the version bump, it can be that you also avoid the conflict, but that is an easy one to solve
GuandongShang (05:32:41) (in thread): > Thanks for your reply. I am wondering whether you can see I understand your “separate commit” correctly. > 1. switch to master -> modify the code -> git commit -m “fix bug” > 2. switch to RELEASE -> cherry-pick the commits from master -> bump version 1.0.1 -> git commit -m “update RELEASE version” > 3. switch to master -> bump version 1.1.1 -> git commit -m “update DEV version”
Kevin Rue-Albrecht (07:56:00) (in thread): > @Federico Marinii’ll make a note/crossref to this message in the mentorship materials repo. It also makes me think that it would be nice to have either/both a markdown page about it and/or a short video demonstrating this, as git can get quite confusing to explain, especially when different developers manage their git repos differently
Kevin Rue-Albrecht (08:00:17) (in thread): > https://github.com/kevinrue/bioc-mentorship-docs/issues/8:eyes:
Marcel Ramos Pérez (09:56:47) (in thread): > Thanks@GuandongShangwe will update the documentation to make it clear that commits should be atomic and version bumps should be separate commits. I can also record a loom video showing this.
Federico Marini (11:49:39) (in thread): > Sounds like an excellent idea@Marcel Ramos Pérez!
Marcel Ramos Pérez (14:24:01) (in thread): > http://contributions.bioconductor.org/git-version-control.html#bug-fix-in-release-and-devel - Attachment (contributions.bioconductor.org): Chapter 28 Git Version Control | Bioconductor Packages: Development, Maintenance, and Peer Review > The Bioconductor project is maintained in a Git source control system. Package maintainers update their packages by pushing changes to their git repositories. This chapter contains several…
2021-11-23
GuandongShang (08:11:03): > Hi. everyone. I fix a bug and my dev version now is 1.1.1(not bump) and my release version is 1.0.1(not bump). Now I am ready to bump my version. But I am not sure how I set the version in my NEWS.md, should I write 1.0.2 or 1.1.2? Or I write 1.0.2 in NEWS.md in master branch and write 1.1.2 in NEWS.md in RELEASE_3_14 branch.
Luke Zappia (09:14:38) (in thread): > I’m not sure there is any standard for this. I usually put the release version in the release branch and both versions in the dev branch but as far as I know this isn’t checked anywhere so whatever makes sense for your own record keeping.
Lori Shepherd (11:02:36) (in thread): > anything in release and devel news will be picked up for the release announcement. I agree we haven’t standardized this but probably should.
GuandongShang (19:42:27) (in thread): > Thanks ：）
2021-11-24
Laurent Gatto (11:32:31): > @Laurent Gatto has joined the channel
2021-11-26
Francesc Català (06:44:31): > @Francesc Català has joined the channel
2021-12-14
Megha Lal (08:23:26): > @Megha Lal has left the channel
2022-02-01
Stephanie Hicks (20:26:21): > @Stephanie Hicks has left the channel
2022-03-23
Henrik Bengtsson (18:22:54): > Hi. Is there a delay betweengit.bioconductor.orgreceiving a push and the commit(s) appearing onhttps://code.bioconductor.org/browse/? I pushed to affxparser on 22:09:31 UTC but it still hasn’t appeared onhttps://code.bioconductor.org/browse/affxparser/commits/masternow 13 minutes later (I’ve re-cloned the git repo to verify that the commit has been pushed.)
Alan O’C (18:26:01): > I vaguely remember reading that the cron job is every 30mins, but I might be making that up. So hopefully just ~5mins left:slightly_smiling_face:
Alan O’C (18:26:58) (in thread): > Ah, yes every 30minshttps://code.bioconductor.org/about.html - Attachment (code.bioconductor.org): Bioconductor Code: About > Details of code.bioconductor.org
Henrik Bengtsson (18:27:54) (in thread): > I failed the “… please wait at least 30 minutes before wondering why.” part.
Henrik Bengtsson (18:30:07) (in thread): > ~~~~~~I’m curious though, what’s the limitation that prevents commits from appear in a near-live fashion? Is it that ~code.bioconductor.org~~ is not “mounted” to ~~git.bioconductor.org, but pulls from there every 30 mins?~~~Time for me to actually read what’s written. … and have a:coffee:. … but I’d add to the wish list that it would be great if this sync could run much more frequently.
Henrik Bengtsson (19:03:24) (in thread): > FWIW, it took more like 45 mins or so; my commit appeared only now, at the top of the hour.
Martin Morgan (19:28:51) (in thread): > I bet this is a poll fromcode.bioconductor.org, something like updating each repository. It would be much more satisfactory for a commit hook on thegit.bioconductor.orgside to push to code.bioconductor.org; I think there’s a github repo athttps://github.com/grimbough/code.bioconductor.orgso not impossible to come up with a pull request (including suitable git hook…). Could be a nice community project…
Martin Morgan (19:30:01) (in thread): > …and then of course looking at the repo I see that updates are based on the RSS feed, which is based on a git hook, so we’re already there, basically, except for perhaps consulting the RSS feed (which shouldn’t be expensive?) more regularly…
2022-03-24
Mike Smith (11:27:20) (in thread): > I think this is mostly a legacy of the development cycle. My first prototype didn’t use the RSS feeds and polled every repository for changes. This took a surprisingly long time and I think I set the CRON job to be every 2 hours. Once I switched to using the RSS feeds I brought it down to 30 mins. There’s probably no reason is can’t be a higher frequency. If a new RSS checker detects there’s already a running container still checking out updates (like at release time) the newly spawned instance will just shut down again. > > Not sure why it would be 45 minutes in the end. Maybe there’s a lag between pushing and the RSS feed updating, or there were several other packages updated in that 30 minute window and it took a while to check them out first. > > The system prints a lot of log files, but because I don’t actually administer the kubernetes instance it’s quite hard to look at them.
2022-03-30
Sergio Oller (23:13:50): > @Sergio Oller has joined the channel
2022-03-31
Julia Wrobel (10:48:58): > @Julia Wrobel has joined the channel
Julia Wrobel (10:51:12): > I’m new to bioconductor and trying to submit my first package- I just bumped my version and am trying to push togit.bioconductor.orgfor the first time, but I get the following error when I run the commandgit push upstream main:
Julia Wrobel (10:51:45): > > To git.bioconductor.org:packages/VectraPolarisData.git > ! [remote rejected] main -> main (hook declined) > error: failed to push some refs to 'git@git.bioconductor.org:packages/VectraPolarisData.git' >
Julia Wrobel (10:51:57): > any tips on what might be going on and how I can resolve this?
Nitesh Turaga (10:53:05): > You cannot push tomain.
Nitesh Turaga (10:53:21): > The branch in thegit.bioconductor.orgserver is calledmaster
Nitesh Turaga (10:54:23): > So, try,git push upstream master.
Julia Wrobel (10:58:34): > thanks for the quick reply!
Julia Wrobel (10:58:48): > i’ve been following the instructions here:https://bioconductor.org/developers/how-to/git/new-package-workflow/
Julia Wrobel (10:59:42): > and they have the following advice: > > ## push to git@git.bioconductor.org. If your github default > ## branch name is main, replace master with main:master > git push upstream master > > ## push to your github repository > ## If your github default branch name is main, replace master with main > git push origin master > > which, is why I used main (my GitHub package is new so it uses main as the default branch).
Julia Wrobel (11:00:05): > how do I check which branch is used in thegit.bioconductor.orgserver?
Nitesh Turaga (11:00:26): > git branch -ashould show you all the branches available
Julia Wrobel (11:01:32): > weird I have both main and master branches at remotesdevelop_tg* main > remotes/origin/develop_tg > remotes/origin/main > remotes/origin/master > remotes/upstream/master
Nitesh Turaga (11:03:03): > Just usemasterto avoid confusion for now.
Julia Wrobel (11:04:53): > thanks - still not working but I’m going to try a couple things then come back here
Nitesh Turaga (11:05:31): > Hang on.
Nitesh Turaga (11:06:11): > Just make themasterbranch the default branch on your github
Nitesh Turaga (11:06:38): > And then you should be able to figure it out easier.
Julia Wrobel (11:06:43): > i don’t have a master branch on github- about a year ago github started making main the default i believe
Julia Wrobel (11:07:04): > and just creating that instead when you generate a new repo
Nitesh Turaga (11:07:21): > ok, try this.
Nitesh Turaga (11:08:29): > > git checkout main > git checkout -b master > git push -u origin master > > > > git branch -D main > > https://docs.github.com/en/repositories/configuring-branches-and-merges-in-your-repo[…]aging-branches-in-your-repository/changing-the-default-branch
Julia Wrobel (11:12:13): > great thanks! that worked to changed the default and i didn’t get any errors pushing upstream either
Nitesh Turaga (11:12:21): > ok great.
Nitesh Turaga (11:12:41): > Please usemasterfor now. We have some work to do on our end regarding the documentation.
Julia Wrobel (11:12:56): > sounds good!
Julia Wrobel (11:13:10): > thanks so much for your help!
Marcel Ramos Pérez (13:17:54): > As a side note,git push upstream main:mastershould work as well. The newer instructions are here (see #5):https://contributions.bioconductor.org/git-version-control.html?q=New%20package#new-package-workflow
2022-04-01
Lukas Weber (19:07:42): > I have been usinggit push upstream main:masterfor my repos too, so can second this suggestion too
2022-04-03
Sergio Oller (23:01:52): > I am used to add to my package’s git repositories the correspondingpackagename.RProjfile (from RStudio) as well as some yaml files to define GitHub CI workflows or even pkgdown yaml (to build a website with documentation). > > I see in theBioconductor package guidelinesthat the Rproj file should not be in the git repository (unacceptable is the word) but rather it should be kept locally. I find that file useful in git since, for instance, it allows any collaborator who uses RStudio to use my project tabs/spaces preferences (which should be aligned with Bioconductor). Other non-essential files are also convenient for CI purposes. > > I have also seen some other Bioconductor package(s) including those files in their git repositories. > > So I wonder why the rule exists and whether it is applied. I was thinking it existed because of a strict desire to save space (¿?), but I guess that’s not the case, since my .Rproj file will be anyway available as a deleted file in the git commit history. > > It’s not a big deal, but it would help me to understand why the rules/guidelines are the way they are so besides following them I can build some criteria as well
2022-04-04
Lori Shepherd (06:38:50) (in thread): > We don’t encourage non essential files, configurations, templates for a number of reasons. Firstly, a lot of extra files makes the package harder to review. Reviewers have to take time to figure out if non standard files are needed, not needed, dangerous, etc . Secondly if Bioconductor switches to use any of the templates/resources in tht future we then have to worry about conflicts and cleaning repos. These are the main reasons among many others… Which is why we encourage extra to be on a different branch and keeping a Bioconductor branch clean.
Sergio Oller (06:45:23) (in thread): > Thank you very much for the explanation. I had not thought about all those reasons and they make a lot of sense. I will do my best to have a branching workflow which is convenient to me but that it preserves the repository as clean as it can be. Thanks again
Leonardo Collado Torres (08:06:12) (in thread): > We’ve had some discussion about related things athttps://github.com/lcolladotor/biocthis/issues/21where I have a different point of view from the Bioc-core
2022-04-05
Vince Carey (06:34:31) (in thread): > I am sorry to admit that this is my first read-through of the comments at biocthis/issues/21 … I found this comment from Levi one that needs to be addressed in depth: > > IMO there should be a clear cost:benefit rationale (accounting for costs to contributors, including those without advanced git skills), along with a clearly defined recommended workflow, to justify asking contributors to move convenience tools such as @lcolladotor outlines from the main branch. Otherwise it's hard to understand why things not noted by R CMD CHECK (because they're in .Rbuildignore) should be of concern to Bioconductor, if they have no impact on users of the software. If the cost:benefit analysis does favor elimination of convenience configurations and decorations, perhaps it would make more sense for Bioconductor to use a dedicated devel branch. That would come with its own challenges and might reduce the frequency with which contributors merge to devel and push upstream as opposed to relying on continuous integration on GitHub, but I think it would feel less controlling and easier to understand. I'd have to see the proposed workflows in each scenario to give an opinion on which I think would be the least burden for the most contributors. >
Vince Carey (06:35:13) (in thread): > I would like to see a thorough discussion of this by a subgroup of contributors after 3.15 is released. I will try to arrange the call in May.
Vince Carey (07:02:19) (in thread): > Perhaps I can make a comment that could foster some strategic thinking in advance of a significant discussion. I have spent a lot of time trying to understand and exposit what I call “the deep stack” of Bioconductor. This is a software system that addresses a) the structure and management of thousands of git repositories with which hundreds of developers interact, b) the invitation and evaluation of hundreds of new repositories and data resources per year “from the wild”, c) testing and repackaging of 2000+ R packages on a background of another 1500 or so CRAN packages, all on three platforms (to become four) in two streams (release and devel). This deep stack is currently “production only”. To date we have not had capacity to branch the “deep stack” in ways that would foster experimental development. Many aspects of the deep stack predate several new technologies of interest in this domain. > > Our access to academic cloud resources may help us to find compute capacity to branch the deep stack and start experimenting with radically new approaches to managing the whole ecosystem in a maintainable way. But I have to confess that I do not see a way of keeping the current rate of growth and level of service to the global genomics community constant WHILE redesigning ecosystem management methods in a profound way that incorporates new and reliable technologies WITHOUT a significant inflow of personnel and intelligence equal to the task of understanding and evolving the deep stack. This latter confession may be mostly a reflection of my limitations and I hope that a broader discussion in May 2022 will show that steps towards adopting new technologies and methods, feasible with available resources, will get us to a 3.16 or 3.17 ecosystem that is more satisfactory to all parties.
Lori Shepherd (07:47:24) (in thread): > We are also discussing and planning the renaming of the “master” branch for bioconductor – instead of using ‘main’ the team has been leaning towards ‘devel’ 1. for better clarity and 2. then a default branch with all the extras can be the defualt while having clean Bioconductor only branches on github
2022-04-06
Sergio Oller (05:54:31) (in thread): > I was considering using a similar approach to the one you suggest,@Lori Shepherdfor my packages, but I feel I would have to keep cherry-picking my commits to avoid merging CI and extra files into the Bioconductor branches. It still feels a significant overhead / barrier of entry for many developers who may be just beginning to use git.@Vince CareyI’m barely a newcomer here, but I would be happy to be involved in the call if you feel it’s appropriate (and I would not feel bad at all if you feel it’s better otherwise).
Vince Carey (06:38:19) (in thread): > Thank you@Sergio OllerI will try to keep you posted, and will announce the call opportunity on slack in May. Just to give a little more motivation for a conservative approach: It is important to think about the “error surface” that we have to try to control in managing the ecosystem. A .Rproj file may seem intrinsically innocuous to those skilled in the use of RStudio. It is just a file in Debian Control Format. But if a mac user “clicks on it” in Finder, it starts RStudio in that folder and certain things happen. Those things must be significant – otherwise the .Rproj file concept would probably not exist. Now the questions arise: Are we sure that everyone who acquires the package from Bioconductor wants that file and wants those potential events? Do we know that the presence of the file and its contents could not induce some peculiar error in the production or use of products of the build system? Does the contributing developer verify that any changes to the .Rproj file format/content conventions are respected as RStudio evolves over time? Does R CMD check assess the file for problems? I don’t know the answers to these questions – I have never knowingly used such a file in 20-odd years of working with R. “Banning it” is a crude way of alleviating all these concerns, admittedly. I am all too aware of the complications of dealing with multiple code branches in git. It would be nice to have some software “on top of” git to help reduce complexity of these operations. Maybe further discussion of the details will help folks –and me – find a more satisfactory approach.
Leonardo Collado Torres (12:59:20) (in thread): > Thanks for the thoughts and updates Vince@Vince Careyand Lori@Lori Shepherd! I’m all in favor of discussing these extra files more and if it comes down to an expectation that bioconductor developers will need to know how to deal with multiplegitbranches, then to so be it, and I’ll updatebiocthisaccordingly (and my Bioc pkgs). > > I’m going to cc Marcel@Marcel Ramos Pérezand Levi@Levi Waldronwho I have thoughts on this topic. Anyways, this topic keeps evolving and small discussions can be found on different github issues / threads, so centralizing it will be great. > > See you around! Hopefully at BioC 2022 too ^^
Leonardo Collado Torres (13:01:26) (in thread): > And Sergio@Sergio Oller, input from beginners is very much welcomed ^^. It’s a very important perspective if we want to bring in new developers to the BioC world ^^ > > I’m going to cc some new developers from my team too: Louise@Louise Huukiand Josh@Josh Stolzas well as Nick@Nick Eagles. I’ve told them some part of the story of this topic to them in the past.
Louise Huuki (13:01:30): > @Louise Huuki has joined the channel
Nick Eagles (13:01:30): > @Nick Eagles has joined the channel
2022-05-03
Ray Su (06:55:53): > @Ray Su has joined the channel
2022-05-09
Levi Waldron (11:55:05): > @Levi Waldron has left the channel
2022-05-13
GuandongShang (22:09:32): > @GuandongShang has joined the channel
GuandongShang (22:15:30): > Hi, guys. > I have a question about version bump. I have a package named FindIT2:https://bioconductor.org/packages/release/bioc/html/FindIT2.htmlI fix my code according to thehttps://contributions.bioconductor.org/git-version-control.html#bug-fix-in-release-and-develBut I find my package in release branch, which is RELEASE_3_15 now do not bump the version.https://bioconductor.org/packages/release/bioc/html/FindIT2.htmlWhile the development version did bump the version.https://bioconductor.org/packages/devel/bioc/html/FindIT2.htmlI have checked the checkResulthttp://bioconductor.org/checkResults/release/bioc-LATEST/FindIT2/and code Browser, it did get my commit.https://code.bioconductor.org/browse/FindIT2/treegraph?branch=masterI am wondering whether you can help me:) - Attachment (Bioconductor): FindIT2 > This package implements functions to find influential TF and target based on different input type. It have five module: Multi-peak multi-gene annotaion(mmPeakAnno module), Calculate regulation potential(calcRP module), Find influential Target based on ChIP-Seq and RNA-Seq data(Find influential Target module), Find influential TF based on different input(Find influential TF module), Calculate peak-gene or peak-peak correlation(peakGeneCor module). And there are also some other useful function like integrate different source information, calculate jaccard similarity for your TF. - Attachment (contributions.bioconductor.org): Chapter 21 Git Version Control | Bioconductor Packages: Development, Maintenance, and Peer Review > The Bioconductor project is maintained in a Git source control system. Package maintainers update their packages by pushing changes to their git repositories. This chapter contains several… - Attachment (Bioconductor): FindIT2 (development version) > This package implements functions to find influential TF and target based on different input type. It have five module: Multi-peak multi-gene annotaion(mmPeakAnno module), Calculate regulation potential(calcRP module), Find influential Target based on ChIP-Seq and RNA-Seq data(Find influential Target module), Find influential TF based on different input(Find influential TF module), Calculate peak-gene or peak-peak correlation(peakGeneCor module). And there are also some other useful function like integrate different source information, calculate jaccard similarity for your TF. - Attachment (code.bioconductor.org): Bioconductor Code: Browse > Browse the content of Bioconductor software packages.
2022-05-14
Lori Shepherd (08:03:01): > I’m away from my computer to check but When did you commit? The release builds only rebuild Monday Wednesday Friday so if you missed Fridays build it won’t be reflected until after Mondays build
2022-05-15
GuandongShang (08:37:48) (in thread): > according to the code browser:https://code.bioconductor.org/browse/FindIT2/commit/f3d4d25a78c442e318d9d26692b3695a7a1ff83bGuandongShangauthored on 10/05/2022 08:27:44 - Attachment (code.bioconductor.org): Bioconductor Code: Browse > Browse the content of Bioconductor software packages.
Lori Shepherd (12:08:55) (in thread): > After further investigation there seemed to be a problem with the release propagation scripts. It should be remedied shortly@Andres Wokaty
2022-05-16
Andres Wokaty (12:54:53) (in thread): > Thank you for raising the issue@GuandongShang. It’s propagated today.
GuandongShang (21:43:53) (in thread): > Thanks for your help,@Lori Shepherd@Andres Wokaty
2022-06-12
Karat Sidhu (12:37:08): > @Karat Sidhu has joined the channel
2022-07-01
Nicholas Cooley (12:56:06): > Hi, not sure what the right way to ask this question is, but our package (SynExtend) now has another maintainer; what is the correct way to give him access to the devel branch? We’d like to set it up so that he’s not pushing changes to my github repository, and instead we’re both pushing to devel.
Alan O’C (14:42:23) (in thread): > > Only one person should be listed as the Maintainer to ensure a single point of contact. This person by default will have commit access to the git repository ongit.bioconductor.org. Commit access can be given to other developers by request on the bioc-devel mailing list. > https://contributions.bioconductor.org/description.html?q=maintainer#description-authors - Attachment (contributions.bioconductor.org): Chapter 6 The DESCRIPTION file | Bioconductor Packages: Development, Maintenance, and Peer Review > The DESCRIPTION file must be properly formatted. The following sections will review some important notes regarding fields of the DESCRIPTION file and associated files. 6.1 Package This is the…
Alan O’C (14:42:58) (in thread): > Although I don’t know if multiple people with commit access is common
2022-07-29
Ryan C. Thompson (15:42:21): > @Ryan C. Thompson has joined the channel
2022-09-27
Jennifer Holmes (16:15:29): > @Jennifer Holmes has joined the channel
2022-10-25
Ludwig Geistlinger (21:03:30): > @Ludwig Geistlinger has joined the channel
Ludwig Geistlinger (21:05:34): > Hi I have a branch-specific problem it appears. I can commit and push to devel. But I can’t push to RELEASE_3_15 for my bugsigdbr package: > > ludwig@Ludwigs-MacBook-Pro bugsigdbr % git push upstream RELEASE_3_15 > Enumerating objects: 16, done. > Counting objects: 100% (16/16), done. > Delta compression using up to 16 threads > Compressing objects: 100% (10/10), done. > Writing objects: 100% (10/10), 831 bytes | 831.00 KiB/s, done. > Total 10 (delta 7), reused 0 (delta 0), pack-reused 0 > remote: FATAL: W refs/heads/RELEASE_3_15 packages/bugsigdbr l.geistlinger DENIED by fallthru > remote: error: hook declined to update refs/heads/RELEASE_3_15 > To git.bioconductor.org:packages/bugsigdbr.git > ! [remote rejected] RELEASE_3_15 -> RELEASE_3_15 (hook declined) > error: failed to push some refs to 'git.bioconductor.org:packages/bugsigdbr.git' > > How to best proceed in such a case?
Lori Shepherd (21:30:54): > The release 3.15 Branch was frozen Oct 17 in accordance with the release schedulehttps://bioconductor.org/developers/release-schedule/it was announced on bioc-devel and the support site
Ludwig Geistlinger (21:38:59): > Oh I see. My bad. Thanks, Lori!
2022-11-29
Assa (09:09:09): > @Assa has joined the channel
2022-12-13
Xiangnan Xu (18:31:42): > @Xiangnan Xu has joined the channel
2023-01-10
Robert Shear (14:09:00): > @Robert Shear has joined the channel
2023-01-12
Martin Grigorov (13:05:31): > @Martin Grigorov has joined the channel
2023-01-18
José Basílio (13:09:11): > @José Basílio has joined the channel
2023-01-26
Chenyue Lu (16:48:35): > @Chenyue Lu has joined the channel
2023-01-31
Francesc Català-Moll (14:34:08): > @Francesc Català-Moll has joined the channel
2023-02-06
Ying Chen (21:35:57): > @Ying Chen has joined the channel
2023-02-17
Peter Hickey (03:37:52): > I’m having trouble pushing to theRELEASE_3_16branch ofOSCA.basic > > OSCA.basic % git log > commit 7f72c9d1b5fcc5ad36135c844b6cc53888cc0e65 (HEAD -> RELEASE_3_16, origin/RELEASE_3_16) > Author: Peter Hickey <peter.hickey@gmail.com> > Date: Wed Feb 15 19:43:05 2023 +1100 > > Version bump > > commit 02aaa25edad393e75ee3aac310afbe7d9176c433 > Author: Peter Hickey <peter.hickey@gmail.com> > Date: Wed Feb 15 19:41:24 2023 +1100 > > Fix link > > - See[https://github.com/OSCA-source/OSCA.basic/issues/10#issuecomment-1430829599](https://github.com/OSCA-source/OSCA.basic/issues/10#issuecomment-1430829599)- Closes #10 > > commit 634198fdcfb2e9b640d6d977ee2ee977dd6e7f0a (upstream/RELEASE_3_16) > Author: Peter Hickey <peter.hickey@gmail.com> > Date: Wed Feb 15 16:59:02 2023 +1100 > > Update README > > OSCA.basic % git push upstream RELEASE_3_16 > Enumerating objects: 13, done. > Counting objects: 100% (13/13), done. > Delta compression using up to 4 threads > Compressing objects: 100% (8/8), done. > Writing objects: 100% (8/8), 829 bytes | 165.00 KiB/s, done. > Total 8 (delta 5), reused 0 (delta 0) > remote: Traceback (most recent call last): > remote: File "hooks/pre-receive.h00-pre-receive-hook-dataexp-workflow", line 89, in <module> > remote: apply_hooks(hooks_dict) > remote: File "hooks/pre-receive.h00-pre-receive-hook-dataexp-workflow", line 71, in apply_hooks > remote: if hooks_dict["pre-receive-hook-merge-markers"]: # enable hook > remote: KeyError: 'pre-receive-hook-merge-markers' > To git.bioconductor.org:packages/OSCA.basic > ! [remote rejected] RELEASE_3_16 -> RELEASE_3_16 (pre-receive hook declined) > error: failed to push some refs to 'git@git.bioconductor.org:packages/OSCA.basic' >
Peter Hickey (03:40:21): > > OSCA.basic % git remote --verbose > origin git@github.com:OSCA-source/OSCA.basic.git (fetch) > origin git@github.com:OSCA-source/OSCA.basic.git (push) > upstream git@git.bioconductor.org:packages/OSCA.basic (fetch) > upstream git@git.bioconductor.org:packages/OSCA.basic (push) >
Peter Hickey (03:41:05): > Why is this push being rejected? I can’t make sense of the pre-receive hook error message
Lori Shepherd (07:21:13): > sorry thats on me most likely – I updated the hooks yesterday . I’ll investigate
Lori Shepherd (07:55:21): > I found the issue and am actively working on the remedy. This should not affect software packages but data experiment, workflows, and books are affected. I hope to have it remedied shortly.
2023-03-06
Yagmur Simsek (12:01:24): > > $ git push -f upstream main > Enter passphrase for key '/c/Users/Yagmur Simsek/.ssh/id_ed25519': > Enumerating objects: 969, done. > Counting objects: 100% (880/880), done. > Delta compression using up to 8 threads > Compressing objects: 100% (306/306), done. > Writing objects: 100% (812/812), 285.90 KiB | 1.59 MiB/s, done. > Total 812 (delta 539), reused 731 (delta 469), pack-reused 0 > remote: Resolving deltas: 100% (539/539), completed with 39 local objects. > remote: FATAL: W refs/heads/main packages/miaSim YagmurSimsekk DENIED by fallthru > remote: error: hook declined to update refs/heads/main > To git.bioconductor.org:packages/miaSim.git > ! [remote rejected] main -> main (hook declined) > error: failed to push some refs to 'git.bioconductor.org:packages/miaSim.git' >
Yagmur Simsek (12:01:45): > Hi! I am the maintainer of the package. I get this error even though I follow thehttps://contributions.bioconductor.org/git-version-control.html#sync-existing-repositories. - Attachment (contributions.bioconductor.org): Chapter 21 Git Version Control | Bioconductor Packages: Development, Maintenance, and Peer Review > The Bioconductor project is maintained in a Git source control system. Package maintainers update their packages by pushing changes to their git repositories. This chapter contains several…
Yagmur Simsek (12:02:39): > This is how my remote URLs looks like > > $ git remote -v > origin[https://github.com/microbiome/miaSim.git](https://github.com/microbiome/miaSim.git)(fetch) > origin[https://github.com/microbiome/miaSim.git](https://github.com/microbiome/miaSim.git)(push) > upstream git@git.bioconductor.org:packages/miaSim.git (fetch) > upstream git@git.bioconductor.org:packages/miaSim.git (push) >
Marcel Ramos Pérez (12:10:44): > Hi Yagmur, did you meangit push upstream main:master? There is nomainbranch atgit.bioconductor.org. See the ‘Note’ in point #5https://contributions.bioconductor.org/git-version-control.html#new-package-workflow
Yagmur Simsek (13:09:48) (in thread): > It worked! Thanks a lot
2023-03-07
Martin Grigorov (02:46:03) (in thread): > also be aware that the branch will be renamed todeveltomorrow -https://bioconductor.github.io/biocblog/posts/2023-03-01-transition-to-devel/ - Attachment (bioconductor.github.io): Bioconductor community blog - Renaming the Default Branch to Devel > During the 3.17 devel cycle, Bioconductor will rename the default branch devel.
Patricia Brianna (03:03:27): > @Patricia Brianna has joined the channel
Patricia Brianna (03:03:58): > hello good people, nice to here as an outreachy applicant
Shravani Nag (11:50:38): > @Shravani Nag has joined the channel
Monalika Patnaik (15:08:14): > @Monalika Patnaik has joined the channel
2023-03-08
Ibizugbe Merit (00:17:35): > @Ibizugbe Merit has joined the channel
Unyime William (01:15:36): > @Unyime William has joined the channel
Atrayee Samanta (02:25:38): > @Atrayee Samanta has joined the channel
Namusisi Sharon (09:57:01): > @Namusisi Sharon has joined the channel
Lori Shepherd (14:30:00): > The Bioconductor Core Team has moved forward with the branch renaming from master to devel ongit.bioconductor.org. > > Please see > BiocBlog:https://bioconductor.github.io/biocblog/posts/2023-03-01-transition-to-devel/Branch Renaming FAQ:https://contributions.bioconductor.org/branch-rename-faqs.htmlAnyone having trouble can post here - Attachment (bioconductor.github.io): Bioconductor community blog - Renaming the Default Branch to Devel > During the 3.17 devel cycle, Bioconductor will rename the default branch devel. - Attachment (contributions.bioconductor.org): Chapter 27 Branch Rename FAQs | Bioconductor Packages: Development, Maintenance, and Peer Review > The Bioconductor project has chosen ‘devel’ to be the default branch for GitHub and git repositories. To prepare for this transition, we have compiled a list of Frequently Asked Questions (FAQs)…
2023-03-09
Sweta Kumari (03:51:37): > @Sweta Kumari has joined the channel
Helen Lindsay (07:37:58): > @Helen Lindsay has joined the channel
K Nodia (08:07:27): > @K Nodia has joined the channel
2023-03-10
Leonardo Collado Torres (12:47:51): > > % git branch -m master devel > > % git fetch origin > From github.com:LieberInstitute/spatialLIBD > * branch devel -> FETCH_HEAD > > % git branch -u origin/devel devel > fatal: the requested upstream branch 'origin/devel' does not exist > hint: > hint: If you are planning on basing your work on an upstream > hint: branch that already exists at the remote, you may need to > hint: run "git fetch" to retrieve it. > hint: > hint: If you are planning to push out a new local branch that > hint: will track its remote counterpart, you may want to use > hint: "git push -u" to set the upstream config as you push. > hint: Disable this message with "git config advice.setUpstreamFailure false" >
Leonardo Collado Torres (12:48:31): > > % git version > git version 2.39.2 >
Leonardo Collado Torres (12:48:50): > I haven’t had time to edit and upload my LIBD rstats club video from today, but well, we ran into this error a few times
Leonardo Collado Torres (12:49:00): > > % git push -u origin devel > branch 'devel' set up to track 'origin/devel'. > Everything up-to-date >
Leonardo Collado Torres (12:49:06): > that seems to things
Leonardo Collado Torres (12:49:27): > though you still get: > > % git remote set-head origin -a > error: Not a valid ref: refs/remotes/origin/devel >
Leonardo Collado Torres (12:49:47): > > % git remote -v > origin git@github.com:LieberInstitute/spatialLIBD.git (fetch) > origin git@github.com:LieberInstitute/spatialLIBD.git (push) > upstream git@git.bioconductor.org:packages/spatialLIBD.git (fetch) > upstream git@git.bioconductor.org:packages/spatialLIBD.git (push) >
Leonardo Collado Torres (12:52:42): > https://jh.zoom.us/rec/share/8F_PzD8sAm-IDj4zYQlvGSSqBQ9PwvqjB83a5gC1OMLOG8eAC3O-OGC7_5hMgRQf.UYIlB3K1m7PoCXdmPasscode: PXn+%#2N
Leonardo Collado Torres (12:52:56): > (I’ll edit that video some time next week)
Marcel Ramos Pérez (12:53:00): > You can only set the upstream branch toorigin/develafter you’ve renamed on GitHub, seegit branch -a
Leonardo Collado Torres (12:53:13): > I had already renamed it on GitHub
Leonardo Collado Torres (12:53:31): > on the video, the instructions worked well for Louise, but for me, they haven’t on 2 repos
Leonardo Collado Torres (12:53:56): > I mean,git push -u origin develmakes it all ok, but others might be confused by the errors I ran into
Leonardo Collado Torres (12:55:33): > > % git branch -a > RELEASE_3_12 > RELEASE_3_14 > RELEASE_3_15 > RELEASE_3_16 > * devel > gh-pages > remotes/origin/gh-pages > remotes/origin/master > remotes/upstream/RELEASE_3_11 > remotes/upstream/RELEASE_3_12 > remotes/upstream/RELEASE_3_13 > remotes/upstream/RELEASE_3_14 > remotes/upstream/RELEASE_3_15 > remotes/upstream/RELEASE_3_16 > remotes/upstream/master >
Leonardo Collado Torres (12:59:16) (in thread): > @Louise Huuki
Marcel Ramos Pérez (13:01:05): > Strange, it doesn’t look likeremotes/origin/develis there even though you’ve fetched. I see it on a fresh clone from github
Marcel Ramos Pérez (13:34:55) (in thread): > I’m able to reproduce if I don’t do agit fetch -aorgit fetch originand try to dogit branch -u origin/devel devel. Thegit fetch originshould say: > > > git fetch origin > From github.com:<ORG>/<REPOSITORY> > * [new branch] devel -> origin/devel > > I can add a note in our docs about usinggit push -u origin develfor when thefetch originisn’t downloading the new branch
Marcel Ramos Pérez (14:09:02) (in thread): > BTW can you show the output of > > git remote show origin > > ?
Edel Aron (15:22:20): > @Edel Aron has joined the channel
Leonardo Collado Torres (15:29:10) (in thread): > sure, here it is: > > % git remote show origin > * remote origin > Fetch URL: git@github.com:LieberInstitute/spatialLIBD.git > Push URL: git@github.com:LieberInstitute/spatialLIBD.git > HEAD branch: devel > Remote branch: > gh-pages tracked > Local branches configured for 'git pull': > devel merges with remote devel > gh-pages merges with remote gh-pages > Local refs configured for 'git push': > RELEASE_3_12 pushes to RELEASE_3_12 (up to date) > RELEASE_3_14 pushes to RELEASE_3_14 (up to date) > RELEASE_3_15 pushes to RELEASE_3_15 (up to date) > RELEASE_3_16 pushes to RELEASE_3_16 (up to date) > devel pushes to devel (fast-forwardable) > gh-pages pushes to gh-pages (local out of date) >
Leonardo Collado Torres (15:29:36) (in thread): > cool, thanks Marcel!
Marcel Ramos Pérez (15:46:14) (in thread): > I guess you’ve sync’ed since. It would be good to see this output whengit fetch originis not working as it should.
Leonardo Collado Torres (16:23:39) (in thread): > well, I still get the error: > > % git branch -u origin/devel devel > fatal: the requested upstream branch 'origin/devel' does not exist > hint: > hint: If you are planning on basing your work on an upstream > hint: branch that already exists at the remote, you may need to > hint: run "git fetch" to retrieve it. > hint: > hint: If you are planning to push out a new local branch that > hint: will track its remote counterpart, you may want to use > hint: "git push -u" to set the upstream config as you push. > hint: Disable this message with "git config advice.setUpstreamFailure false" >
Leonardo Collado Torres (16:24:00) (in thread): > you can see athttps://community-bioc.slack.com/archives/C6MVC96AZ/p1678470471996299that I did rungit fetch origin - Attachment: Attachment > % git branch -m master devel > > % git fetch origin > From [github.com:LieberInstitute/spatialLIBD](http://github.com:LieberInstitute/spatialLIBD) > * branch devel -> FETCH_HEAD > > % git branch -u origin/devel devel > fatal: the requested upstream branch 'origin/devel' does not exist > hint: > hint: If you are planning on basing your work on an upstream > hint: branch that already exists at the remote, you may need to > hint: run "git fetch" to retrieve it. > hint: > hint: If you are planning to push out a new local branch that > hint: will track its remote counterpart, you may want to use > hint: "git push -u" to set the upstream config as you push. > hint: Disable this message with "git config advice.setUpstreamFailure false"
Leonardo Collado Torres (16:25:14) (in thread): > in other words,git fetch originseems to be working, but then for some reason, theorigin/develbranch is still not locally available when I rungit branch -u origin/devel devel
Leonardo Collado Torres (16:44:32): > also athttps://contributions.bioconductor.org/branch-rename-faqs.html#what-if-i-am-an-rstudio-usercan we maybe change the wording a bit for: > > 27.5 What if I am an RStudio user? > > Maintainers that use RStudio should be familiar with runninggitcommands via theTerminaltab in RStudio. > It kind of reads like: well, you are screwed if you don’t know how to rungitcommands in the terminal. > > Maybe we edit it a bit and add a little extra sentence like: > > Maintainers that use RStudio are expected to be familiar with runninggitcommands via theTerminaltab in RStudio. However, if you are unfamiliar with usinggitcommands in theTerminal, you might find the material athttps://happygitwithr.com/git-commands.htmluseful to understand how somegitbuttons on the RStudio interface translate togitcommands on theTerminal.
Marcel Ramos Pérez (16:46:06) (in thread): > It hasFETCH_HEADinstead oforigin/develI’m not sure what is causing that
Marcel Ramos Pérez (16:46:47) (in thread): > The normal output should look like this:https://community-bioc.slack.com/archives/C6MVC96AZ/p1678473295306909?thread_ts=1678471265.892219&cid=C6MVC96AZ - Attachment: Attachment > I’m able to reproduce if I don’t do a git fetch -a or git fetch origin and try to do git branch -u origin/devel devel . The git fetch origin should say: > > git fetch origin > From [http://github.com:<ORG](http://github.com:<ORG)/[github.com:<ORG](REPOSITORY)/<REPOSITORY>> > * [new branch] devel -> origin/devel > I can add a note in our docs about using git push -u origin devel for when the fetch origin isn’t downloading the new branch
Leonardo Collado Torres (16:53:23) (in thread): > hmm
Leonardo Collado Torres (16:53:46) (in thread): > at some I guess I did open both of those withstreehttps://www.sourcetreeapp.com/ - Attachment (SourceTree): Sourcetree | Free Git GUI for Mac and Windows > A Git GUI that offers a visual representation of your repositories. Sourcetree is a free Git client for Windows and Mac.
Leonardo Collado Torres (16:54:04) (in thread): > because my local clone ofTREGby@Louise Huukiworked without issues too
Leonardo Collado Torres (16:54:24) (in thread): > so it’s not an issue with my localgitversion
Leonardo Collado Torres (16:54:30) (in thread): > that’s the only idea that comes to mind
Marcel Ramos Pérez (17:03:42) (in thread): > Perhaps the app is changing your~/.gitconfigsettings
2023-03-13
Alan O’C (08:02:30) (in thread): > I don’t know that that git link is actually that helpful for somebody who’s not familiar with git on the commandline? It’s quite terse
2023-03-15
Alex Mahmoud (13:49:43): > @Alex Mahmoud has joined the channel
2023-03-29
Matteo Tiberti (10:11:45): > @Matteo Tiberti has joined the channel
Himel Mallick (16:56:31): > @Himel Mallick has joined the channel
2023-04-03
Mengbo Li (20:45:12): > @Mengbo Li has joined the channel
2023-04-05
Josué Curto Navarro (20:59:26): > @Josué Curto Navarro has joined the channel
2023-05-04
Leopoldo Valiente (16:46:00): > @Leopoldo Valiente has joined the channel
2023-05-18
Sarvesh Nikumbh (03:37:35): > @Sarvesh Nikumbh has joined the channel
Oluwafemi Oyedele (05:53:31): > @Oluwafemi Oyedele has joined the channel
2023-05-24
Rishi (04:44:36): > @Rishi has joined the channel
Rishi (04:44:59): > Hi, > I have git related question. Recently mypackagehas moved from devel to master branch ingit.bioconductor.org. I am trying to push my code with old command which was working till now but now I am getting following error.git push upstream master:develThe errors are > > To git.bioconductor.org:packages/DELocal.git > ! [rejected] master -> devel (non-fast-forward) > error: failed to push some refs to 'git.bioconductor.org:packages/DELocal.git' > hint: Updates were rejected because a pushed branch tip is behind its remote > hint: counterpart. Check out this branch and integrate the remote changes > hint: (e.g. 'git pull ...') before pushing again. > hint: See the 'Note about fast-forwards' in 'git push --help' for details. > > The output of “git remote -v” > > DELocal git@github.com:dasroy/DELocal.git (fetch) > DELocal git@github.com:dasroy/DELocal.git (push) > origin git@github.com:dasroy/DELocal.git (fetch) > origin git@github.com:dasroy/DELocal.git (push) > upstream git@git.bioconductor.org:packages/DELocal.git (fetch) > upstream git@git.bioconductor.org:packages/DELocal.git (push) > > Please let me know how can I fix this - Attachment (Bioconductor): DELocal > The goal of DELocal is to identify DE genes compared to their neighboring genes from the same chromosomal location. It has been shown that genes of related functions are generally very far from each other in the chromosome. DELocal utilzes this information to identify DE genes comparing with their neighbouring genes.
Alan O’C (04:52:52) (in thread): > The upstream branch has been updated, as the (admittedly cryptic) git message tells you the pushed branch is behind its remote counterpart. > > I think you need to rungit pull upstream develor equivalentlygit fetch upstream && git merge upstream/develfrom the master branch before pushing. Or you could switch to using adevelbranch locally and on github, seehttps://bioconductor.github.io/biocblog/posts/2023-03-01-transition-to-devel/ - Attachment (bioconductor.github.io): Bioconductor community blog - Renaming the Default Branch to Devel > During the 3.17 devel cycle, Bioconductor will rename the default branch devel.
Rishi (04:55:08) (in thread): > Thanks for your reply. Now I am getting thisgit pull upstream devel > > From git.bioconductor.org:packages/DELocal > * branch devel -> FETCH_HEAD > hint: You have divergent branches and need to specify how to reconcile them. > hint: You can do so by running one of the following commands sometime before > hint: your next pull: > hint: > hint: git config pull.rebase false # merge (the default strategy) > hint: git config pull.rebase true # rebase > hint: git config pull.ff only # fast-forward only > hint: > hint: You can replace "git config" with "git config --global" to set a default > hint: preference for all repositories. You can also pass --rebase, --no-rebase, > hint: or --ff-only on the command line to override the configured default per > hint: invocation. > fatal: Need to specify how to reconcile divergent branches. >
Alan O’C (04:58:52) (in thread): > This is a git setting, see:https://stackoverflow.com/questions/62653114/how-can-i-deal-with-this-git-warning-pulling-without-specifying-how-to-reconci - Attachment (Stack Overflow): How can I deal with this Git warning? “Pulling without specifying how to reconcile divergent branches is discouraged” > After a git pull origin master, I get the following message: > warning: Pulling without specifying how to reconcile divergent branches is > discouraged. You can squelch this message by running one of the
Alan O’C (05:00:27) (in thread): > I don’t know that I can recommend it, but I use the default strategy listed in the hint there, set as the default preference for all repositories
Rishi (05:20:50) (in thread): > Thanks I hope I have been able to push now correctly with a new version bump. Should I receive an email notification stating that a new build will initiate?
Alan O’C (05:22:11) (in thread): > The next build will be the nightly build, you won’t get emails unless your package is failing I think, but you can also check the build pageshttp://bioconductor.org/checkResults/release/bioc-LATEST/DELocal/http://bioconductor.org/checkResults/devel/bioc-LATEST/DELocal/
Marcel Ramos Pérez (12:15:52) (in thread): > Note. You can check to see if your commit went through athttps://code.bioconductor.org/browse/DELocal/commits/devel
Rishi (12:17:44) (in thread): > Thanks Marcel and Alan, now I can see my commit went through.
2023-07-13
Mike Smith (06:31:08): > I have a fun little git puzzle. Does anyone have any idea what might have happened to the git histories ofSTATegRaandflowDensity? They’ve each got ~40,000 commits and ~700 contributing authors, but it mostly looks like nonsense. When you take a look at the commits they don’t actually refer to any of the files in the repositories and seem to have come from unrelated packages. It’s almost like commits from the entire BioC version control across all packages were somehow merge. I don’t think it’s really causing any problems, but I’ve never seen anything like it and was curious whether anyone could shed any light on it. Perhaps@Nitesh Turagawas most familiar with the git repo around the time it seems to have happened.
Nitesh Turaga (12:19:40): > It’s funny cause there are also no duplicate commits.
Nitesh Turaga (12:21:55): > The names of the authors are also messed up. It’s like they just took any genericFirstname Lastnameand made a git username asf.lastnameappear.
Nitesh Turaga (12:22:57): > I’ve no memory of how this could have happened.
Martin Morgan (13:58:32) (in thread): > @Mike Smithare these the only packages looking weird? How did you find them?
Lluís Revilla (13:59:48): > Maybe tracking the history to try to find when all those users contributed might help finding the issue. I just know that STATegRa moved between github repositories (I also know one of the maintainers, in case you want me to reach out to them)
Lori Shepherd (14:15:27): > It seems like all the git history from the transition from svn to git is accurate and appropriate — it is pre conversion, svn commits that are wonky – so my gut feeling is it happened on the conversion from svn to git in Bioconductor, also not sure how that exactly works with the transition if these svn commits are/would be visible
Mike Smith (14:31:18) (in thread): > > are these the only packages looking weird? How did you find them? > I thought it might be interesting to look at the total number of commits in the project, maybe the most frequent days commits are made, how much happens a week before release etc. I think you used to present similar data for download counts. > > Since I have all the software repos cloned locally I made a data frame with all the commits from the devel branches. Those two packages stood out as crazy outliers for total number of commits. The next two are IRanges and Biostrings with ~3200 and ~2200 respectively, which seems more in keeping with what one might expect.
Nitesh Turaga (16:45:24): > I think those commits are “empty” and you could just do something along the lines of > > git filter-branch --prune-empty --tag-name-filter cat -- --all > > There might be a little bit more work after this though
2023-07-17
Leonardo Collado Torres (15:15:01): > https://twitter.com/lcolladotor/status/1681019609150005249 - Attachment (Twitter): :flag-mx: Leonardo Collado-Torres on Twitter > From 2023-03-10 we have “Updating your @Bioconductor packages: devel #Git branch and #biocthis updates” > > @LIBDrstats @LieberInstitute #rstats @github > > :spiral_note_pad: https://t.co/WBOBMoH2yH > :video_camera: https://t.co/9iEFiQHOSl
Leonardo Collado Torres (15:21:55): > This one is maybe more for#biocbooks, but well, I’ll share it herehttps://twitter.com/lcolladotor/status/1681020937494466560 - Attachment (Twitter): :flag-mx: Leonardo Collado-Torres on Twitter > Last @LIBDrstats video from me so far in 2023 is 06-23 “Making a documentation website book with #bookdown (@xieyihui) & @GitHub actions” > > More by the team soon! > > #biocthis @Bioconductor #rstats @LIBDrstats @LieberInstitute > > :spiral_note_pad: https://t.co/wKayvU2Sd5 > :video_camera: https://t.co/v0BqsHOKu3
2023-07-28
Konstantinos Daniilidis (13:47:19): > @Konstantinos Daniilidis has joined the channel
2023-08-01
Nikhil Mane (09:14:00): > @Nikhil Mane has joined the channel
2023-08-04
Trisha Timpug (09:34:36): > @Trisha Timpug has joined the channel
2023-08-16
mahamtariq58 (09:47:28): > @mahamtariq58 has joined the channel
2023-08-20
Jacques SERIZAY (10:38:37): > @Jacques SERIZAY has joined the channel
2023-08-24
Lachlan Baer (01:20:55): > @Lachlan Baer has joined the channel
2023-09-15
Leo Lahti (04:53:28): > @Leo Lahti has joined the channel
2023-09-21
Philippine Louail (16:39:58): > @Philippine Louail has joined the channel
2023-10-17
Anna Quaglieri (she/her) (02:01:06): > @Anna Quaglieri (she/her) has joined the channel
2023-11-03
Robert Castelo (15:00:33): > @Robert Castelo has joined the channel
Robert Castelo (15:07:09): > I’m experiencing for the first time 404 errors when trying to access my public GitHub repos, through the browser and in command line when doinggit push, I’m also seeing it with repos such ashttps://github.com/Bioconductor/Contributionsandhttps://www.githubstatus.comis reportingdegraded availability, anyone experiencing connecting problems as well?? - Attachment (githubstatus.com): GitHub Status > Welcome to GitHub’s home for real-time and historical data on system performance.
Robert Castelo (15:19:51): > GitHub seems to be back to operating normally.
2023-11-09
LAR (15:58:31): > @LAR has joined the channel
LAR (16:03:18): > Hello everyone: > > I recently made resolved some errors in my Bioconductor package (here). I’ve been following instructions for doing a version bump to push the new changes, following instructions under the “New package workflow” section (here), and made it up to the point 5 (“Push changes to the Bioconductor and GitHub repositories”). > > However, upon attempting the linegit push upstream devel, I receive the error:" ! [remote rejected] devel -> devel (pre-receive hook declined)``error: failed to push some refs to 'git.bioconductor.org:packages/bigPint.git'"I see in Bioc-devel mail archives (here) that this error can be something on the Bioconductor developer side. My question is how can I determine if it is a problem on my end or Bioconductor end? > > Thank you for any advice you may be able to provide!
Marcel Ramos Pérez (16:31:04): > Hi Lindsay,bigPintwasdeprecatedin3.18for failing to build in3.17. Perhaps that’s why you can’t push but I am not sure. Lori@Lori Shepherdshould have more information. For now, can you show the entire error message? What are yourgit remotes -v?
Marcel Ramos Pérez (16:33:46): > > Please do not undeprecate a package yourself. You must request undeprecation > from the Bioconductor core team at bioc-devel@r-project.org. >
LAR (18:16:41): > Thank you, Marcel. I will wait a bit to hear back from Lori and, if not, I will send an email to the team as you suggested.
LAR (18:16:55): > The full error is below: > > Enumerating objects: 114, done. > Counting objects: 100% (96/96), done. > Delta compression using up to 4 threads > Compressing objects: 100% (50/50), done. > Writing objects: 100% (72/72), 12.06 KiB | 3.01 MiB/s, done. > Total 72 (delta 53), reused 27 (delta 21), pack-reused 0 > remote: fatal: Not a valid object name f7c29bb3d981eb1080813b334bdc3731e6cd9b4d:"\033\033" > remote: Traceback (most recent call last): > remote: File "hooks/pre-receive.h00-pre-receive-hook-software", line 93, in <module> > remote: apply_hooks(hooks_dict) > remote: File "hooks/pre-receive.h00-pre-receive-hook-software", line 76, in apply_hooks > remote: prevent_large_files(oldrev, newrev, refname) > remote: File "/home/git/.gitolite/local/hooks/repo-specific/prevent_large_files.py", line 42, in prevent_large_files > remote: newrev + ":" + fl.decode()]) > remote: File "/usr/lib/python3.5/subprocess.py", line 626, in check_output > remote: **kwargs).stdout > remote: File "/usr/lib/python3.5/subprocess.py", line 708, in run > remote: output=stdout, stderr=stderr) > remote: subprocess.CalledProcessError: Command '['git', 'cat-file', '-s', 'f7c29bb3d981eb1080813b334bdc3731e6cd9b4d:"\\033\\033"']' returned non-zero exit status 128 > To git.bioconductor.org:packages/bigPint.git > ! [remote rejected] devel -> devel (pre-receive hook declined) > error: failed to push some refs to 'git.bioconductor.org:packages/bigPint.git' >
Marcel Ramos Pérez (18:31:59) (in thread): > You should send the email to request undeprecation. It seems like commits tobigPintare blocked. > > Note. I deleted yourgit remote -vbecause it seemed to contain sensitive information.
LAR (18:36:53) (in thread): > Thanks, Marcel. Very helpful. I will send an email now to request undepreciation.
2023-11-10
Lori Shepherd (13:03:21) (in thread): > I can check credentials when I’m back from vacation next week but some tips in the meantime. Make sure you’ve uploaded a ssh key to your credentials account.Also if the package was deprecated, you should be pushing changes to the RELEASE_3_18 branch ongit.Bioconductor.org; once the package builds cleanly in release we would consider undeprecating and adding back to the devel/3.19 branch
Lori Shepherd (13:04:27) (in thread): > The package won’t show in devel until we undeprecate. And we won’t undeprecate until the package is fixed
LAR (13:08:10) (in thread): > Thanks, Lori! I will prepare myself by reviewing the information you provided here, and be in touch with you when you return, to move this process forward. Thank you again:slightly_smiling_face:
2023-11-13
Lori Shepherd (08:55:29): > this looks like there is an issue with the hook trying to check large package files. I’ll look into it.
2023-11-14
Lori Shepherd (12:43:33): > still investigating but also FWIW I would guess that the hook is actually being hit and failing but not completing right or giving the right more helpful message. Does your package or could your package contain large files (files over 5MB)?
2023-11-15
LAR (08:27:20): > Thank you, Lori. I just rancheck()on my package, and it gives the following: > > ❯ checking installed package size ... NOTE > installed size is 7.8Mb > sub-directories of 1Mb or more: > data 2.1Mb > doc 2.4Mb > shiny-examples 3.0Mb >
LAR (08:27:45): > So, it seems individual files are not larger then 5mb, but the total size is 7.8mb.
LAR (08:30:36): > However, it does seem that one of my data files is >5MB. It isdata/soybean__cn_metrics.rda, which appears to be 5.5MB.
Kasper D. Hansen (08:52:25): > That seems like a weird discrepancy. Note thatcheck()doesn’t report the file size in the git repo, but the installed file size. And what matters for git, is the size of the git repos.
LAR (15:40:41): > Thanks, Kasper. I agree it is a strange discrepancy. If notcheck(), what should I use to determine the file size of the git repo? I’m still a bit unsure whether my package contains large files (over 5MB). It had been on Bioconductor for a while, and, as far as I remember, I had not increased any file sizes.
Kasper D. Hansen (15:43:21): > Well, the terminal
Kasper D. Hansen (15:43:51): > You’re saying you have a `data/soybean__cn_metrics.rdawhich is 5.5MB. Isn’t that the answer
2023-11-16
LAR (15:33:43): > I’m not entirely sure. The package has been on Bioconductor for years, and I do not believe I have altered this dataset. I’m not sure how else its size might have changed.
2023-12-28
David Rach (08:58:58): > @David Rach has joined the channel
2024-01-31
Clemens Kohl (07:21:07): > @Clemens Kohl has joined the channel
2024-03-06
Caitlin Page (00:00:51): > @Caitlin Page has joined the channel
2024-03-22
Hervé Pagès (17:20:19): > @Hervé Pagès has left the channel
2024-03-25
Pedro Sanchez (06:27:29): > @Pedro Sanchez has joined the channel
2024-04-04
Tung Trinh (23:39:15): > @Tung Trinh has joined the channel
2024-04-25
Mercedes Guerrero (05:02:47): > @Mercedes Guerrero has joined the channel
2024-05-15
Sunil Nahata (08:31:20): > @Sunil Nahata has joined the channel
2024-05-30
Kevin Rue-Albrecht (04:15:35): > I’d like to clean up myvelociraptorrepo. I can seedevelandmasterbranches both locally and on the bioc remote (upstream). > On github,develis the default branch. > Is it automatically the case on the bioc git? > Can I safely delete all themasterbranches, locally and remotely? > > $ git branch -a > * devel > kra-conda > master > testload > remotes/multimeric/devel > remotes/multimeric/testload > remotes/origin/devel > remotes/origin/gh-pages > remotes/origin/kra-conda > remotes/origin/velocyto > remotes/upstream/RELEASE_3_12 > remotes/upstream/RELEASE_3_13 > remotes/upstream/RELEASE_3_14 > remotes/upstream/RELEASE_3_15 > remotes/upstream/RELEASE_3_16 > remotes/upstream/RELEASE_3_17 > remotes/upstream/RELEASE_3_18 > remotes/upstream/RELEASE_3_19 > remotes/upstream/devel > remotes/upstream/master >
Martin Grigorov (04:43:54) (in thread): > I think it is safe to delete all branches butdevelandRELEASE_**.https://code.bioconductor.org/browse/velociraptor/has only those
Kevin Rue-Albrecht (04:45:38) (in thread): > ah! good point, didn’t think of checking there
Kevin Rue-Albrecht (04:45:43) (in thread): > thanks
Kevin Rue-Albrecht (04:54:20) (in thread): > strange, when I deletemaster, the next time Ifetch --allit find it again on the bioc git: > > $ git branch -d -r upstream/master > Deleted remote-tracking branch upstream/master (was ad13b4a). > $ git fetch --all --prune > Fetching origin > Fetching upstream > From git.bioconductor.org:packages/velociraptor > * [new branch] master -> upstream/master > Fetching multimeric >
Kevin Rue-Albrecht (04:55:00) (in thread): > Doeshttps://code.bioconductor.org/browse/velociraptor/hidemaster?:face_with_raised_eyebrow: - Attachment (code.bioconductor.org): Bioconductor Code: velociraptor > Browse the content of Bioconductor software packages.
Martin Grigorov (05:10:57) (in thread): > You deleted it locally. You need to dogit push :masterto delete it remotely. Or you can use GitHub UI to do it via button clicks
Alan O’C (05:20:02) (in thread): > It’s claiming to be from the bioc remote, not github, so I’d suspect if yougit branch -d master && git branch -d upstream/masterthen there’s nothing, unless bioc does have a hidden master
Martin Grigorov (05:21:51) (in thread): > right! it is not Github! I’m a bit distracted here …
Kevin Rue-Albrecht (05:23:44) (in thread): > It didn’t like it:cry: > > $ git branch -a > * devel > gh-pages > kra-conda > testload > remotes/multimeric/devel > remotes/multimeric/testload > remotes/origin/devel > remotes/origin/gh-pages > remotes/origin/kra-conda > remotes/origin/velocyto > remotes/upstream/RELEASE_3_12 > remotes/upstream/RELEASE_3_13 > remotes/upstream/RELEASE_3_14 > remotes/upstream/RELEASE_3_15 > remotes/upstream/RELEASE_3_16 > remotes/upstream/RELEASE_3_17 > remotes/upstream/RELEASE_3_18 > remotes/upstream/RELEASE_3_19 > remotes/upstream/devel > remotes/upstream/master > $ git push upstream --delete master > remote: FATAL: + refs/heads/master packages/velociraptor k.rue-albrecht DENIED by fallthru > remote: error: hook declined to update refs/heads/master > To git.bioconductor.org:packages/velociraptor > ! [remote rejected] master (hook declined) > error: failed to push some refs to 'git.bioconductor.org:packages/velociraptor' >
Alan O’C (05:25:46) (in thread): > fwiw I also see a (remotes/)bioc/master on densvis + scater, which also don’t show up on the code.bioc browser
Kevin Rue-Albrecht (05:26:42) (in thread): > > unless bioc does have a hidden master > so many jokes possible on that, but none worthy of writing out loud yet:stuck_out_tongue_winking_eye:
Alan O’C (05:27:53) (in thread): > I thought the (not so) hidden master is still Aaron:fast_parrot:
Lori Shepherd (08:01:04) (in thread): > We have not officially removed (hidden) master from bioc git. We made links to allow people more time to adjust. We plan to remove over the next few weeks now that we were through the last release and sent the notification out
2024-06-05
Mike Smith (08:21:46) (in thread): > I think I choose not to clone themasterbranch oncode.bioconductor.orgsince it’s always a copy ofdevel. Relevant code would be:https://github.com/Bioconductor/code.bioconductor.org/blob/646116b1826363321f6ca0f201b0a83bb5b73d45/mirror-updater/git_update_functions.R#L95-L98
Lori Shepherd (08:25:52) (in thread): > yes and since its planned to be removed that is appropriate
2024-06-11
Ziru Chen (04:36:57): > @Ziru Chen has joined the channel
2024-07-04
Sounkou Mahamane Toure (15:28:10): > @Sounkou Mahamane Toure has joined the channel
2024-07-11
Sathish Kumar (06:03:12): > @Sathish Kumar has joined the channel
2024-07-16
Hiru (06:56:52): > @Hiru has joined the channel
2024-07-26
Qiwen Octavia Huang (19:51:53): > @Qiwen Octavia Huang has joined the channel
2024-08-09
Kevin Stachelek (13:51:43): > hello, I am trying to submit a bioconductor package and am trapped in a git cycle. I am trying to integrate the help from these two pages:http://contributions.bioconductor.org/git-version-control.html#new-package-workflowhttps://contributions.bioconductor.org/git-version-control.html#remove-large-data-files-and-clean-git-treeI have an upstream git@git.bioconductor.org:packages with large files. I run git filter-repo to remove them from my github repository but when i try to push upstream I am rejected because I haven’t pulled the large files from the upstreamgit.bioconductor.orgrepo.git push --forceto the upstream is rejected with this error message: > > remote: Resolving deltas: 100% (4802/4802), done. > remote: FATAL: + refs/heads/devel packages/chevreul whtns DENIED by fallthru > remote: error: hook declined to update refs/heads/devel > To git.bioconductor.org:packages/chevreul.git > ! [remote rejected] devel -> devel (hook declined) > error: failed to push some refs to 'git@git.bioconductor.org:packages/chevreul.git' > > Can someone help me resolve this? - Attachment (contributions.bioconductor.org): Chapter 21 Git Version Control | Bioconductor Packages: Development, Maintenance, and Peer Review > The Bioconductor project is maintained in a Git source control system. Package maintainers update their packages by pushing changes to their git repositories. This chapter contains several… - Attachment (contributions.bioconductor.org): Chapter 21 Git Version Control | Bioconductor Packages: Development, Maintenance, and Peer Review > The Bioconductor project is maintained in a Git source control system. Package maintainers update their packages by pushing changes to their git repositories. This chapter contains several…
Marcel Ramos Pérez (14:56:55) (in thread): > I think you would need admin to force push your updates togit.bioconductor.org.@Lori ShepherdIs this something we do?
Marcel Ramos Pérez (14:57:40) (in thread): > @Kevin StachelekPlease link the updated GitHub repository here
Kevin Stachelek (18:50:47) (in thread): > Thanks for the reply. I am still able to move forward for the moment because github is only issuing warnings regarding these large files and not limiting my ability entirely to push. Here is the repo:https://github.com/whtns/chevreul
2024-08-12
Lori Shepherd (07:23:18) (in thread): > yes normally we would either reset or force push for you as we do not allow maintainers to force push only admin.
Lori Shepherd (07:24:41) (in thread): > Is the github repo above ready to be reset on Bioconductor systems? I can look at this later this afternoon
2024-08-14
Kevin Stachelek (15:30:41) (in thread): > Thanks for this info, I’m following up on the contribution issue of my package. I’ll avoid large files whenever possible before submission in the future.
2024-08-19
Rema Gesaka (09:40:59): > @Rema Gesaka has joined the channel
2024-09-07
Tyler Sagendorf (14:35:09): > @Tyler Sagendorf has joined the channel
2024-09-13
Gobi Dasu (18:18:54): > @Gobi Dasu has joined the channel
2024-09-25
Martin Emons (07:49:03): > @Martin Emons has joined the channel
2024-10-02
Eva Hamrud (19:03:57): > @Eva Hamrud has joined the channel
2024-10-23
Hong Qin (17:46:33): > @Hong Qin has joined the channel
2025-02-13
JP Flores (13:54:44): > @JP Flores has joined the channel
2025-02-19
Anushka Paharia (14:06:34): > @Anushka Paharia has joined the channel
2025-03-21
W Sun (17:19:04): > @W Sun has joined the channel
W Sun (17:21:39): > Hello, a month ago, I attempted to use Rust in my package and push it to the Bioconductor development branch. However, Andreas was unable to install the Rust compiler on Windows. As a result, I need to revert to the previous version. I tried to rebase and push the Git repository to the Bioconductor development branch but encountered an error. Do you have any advice?
Lori Shepherd (18:17:54): > Bioconductor doesn’t allow force pushes required for reverts and rebase. We can under extreme circumstances assist with this but it is highly not recommended and will require core team assistance. a reset that keeps the history and then a new version bump is better for public repository so that anyone that has checked out your code can achieve a synced repoand should be possible with out a force push
Marcel Ramos Pérez (18:27:38) (in thread): > As Lori said, you can’t force push to the git repo (only to GH). You should revert to the commit where the package was working and then bump the version to something higher than when the unwanted changes were implemented. Also please link to the git repository either on GH orcode.bioconductor.orgfor more detailed guidance.
W Sun (18:33:46): > Thanks @Marcel and @Lori . I have technical question about git. I tried to reset the code into earlier version. with: > > ## after pull the repo from bioc > git reset --soft d1d354f46f281bbe3903dd94ff6e2a78afb9d677 > > And also bump the version. > But will I tried to push it to the git, it will report the error that: > > To git.bioconductor.org:packages/CellBarcode.git > ! [rejected] devel -> devel (non-fast-forward) > error: failed to push some refs to 'git.bioconductor.org:packages/CellBarcode.git' > hint: Updates were rejected because the tip of your current branch is behind > hint: its remote counterpart. If you want to integrate the remote changes, > hint: use 'git pull' before pushing again. > hint: See the 'Note about fast-forwards' in 'git push --help' for details. > > Any details about how to solve the issue.
Marcel Ramos Pérez (18:36:40) (in thread): > resetrequires a force push so it’s better torevert
W Sun (19:09:47): > Thank you both. finally got it to work byrevert.
2025-03-28
Khadija Juma (07:26:33): > @Khadija Juma has joined the channel