#r7

2021-03-18

Michael Lawrence (13:12:28): > @Michael Lawrence has joined the channel

Michael Lawrence (13:12:28): > set the channel description: Potential adoption of R7 within Bioconductor

Mike Smith (13:13:56): > @Mike Smith has joined the channel

Charlotte Soneson (13:16:14): > @Charlotte Soneson has joined the channel

James MacDonald (13:17:04): > @James MacDonald has joined the channel

Martin Morgan (13:22:11): > @Martin Morgan has joined the channel

Dirk Eddelbuettel (13:22:37): > @Dirk Eddelbuettel has joined the channel

Nathan Eastwood (13:39:40): > @Nathan Eastwood has joined the channel

Henrik Bengtsson (15:16:57): > @Henrik Bengtsson has joined the channel

Sanchit Saini (15:20:50): > @Sanchit Saini has joined the channel

Peter Hickey (15:29:20): > @Peter Hickey has joined the channel

Laurent Gatto (17:33:43): > @Laurent Gatto has joined the channel

Stuart Lee (17:57:29): > @Stuart Lee has joined the channel

2021-03-19

Vince Carey (09:34:11): > @Vince Carey has joined the channel

2021-03-20

watanabe_st (01:58:11): > @watanabe_st has joined the channel

2021-04-07

Mike Smith (08:31:22): > Do any of the OOP paradigms in R have the concept of a destructor (maybe that’s a finalizer with garbage collection)? I was thinking about the S4 objects we use to represent HDF5 object handles, and how if you dorm(h5_file_id)without first doingH5FClose(h5_file_id)it leaves an open handle. The same is true if you open something inside a function and forget to close. Then it’s lost to the user and we can be stuck with “this file is already open” style warnings. I haven’t really thought about the logistics, but it’d be cool if once there was no R object referencing an HDF5 handle it got closed automatically.

Michael Lawrence (13:53:09): > Are you looking forreg.finalizer()?

Henrik Bengtsson (18:07:41): > R.oocalls S3 genericfinalize()on any R.oo objectObjectwhen garbage collected, i.e. you can create custom finalizers by defining S3 methods. It relies onbase::reg.finalizer()internally. It goes an extra mile to try to make it “reentrant”, e.g. work also whenR.oohas been unloaded when the finalizer is called, and other obscure things that can happen with finalizers. - Attachment (cran.r-project.org): R.oo: R Object-Oriented Programming with or without References > Methods and classes for object-oriented programming in R with or without references. Large effort has been made on making definition of methods as simple as possible with a minimum of maintenance for package developers. The package has been developed since 2001 and is now considered very stable. This is a cross-platform package implemented in pure R that defines standard S3 classes without any tricks.

Michael Lawrence (18:50:27): > The methods package’s reference classes have a similar feature. We are currently deferring reference classes though.

Dirk Eddelbuettel (18:57:54): > And packages using compiled code often stick an external pointer object with an explicit finalizer call into their R6 or S4 structures.

2021-04-08

Martin Morgan (03:58:12): > I thinkreg.finalizeris not satisfactory for the purposes Mike mentions. The finalizer runs at a poorly specified time, so one can’t make any assumptions about available functions, other than base R (e.g., the package or even other state in a class definition may already have been garbage collected). The help page ?reg.finalizer also says > > R's interpreter is not re-entrant and the finalizer could be run > in the middle of a computation. So there are many functions which > it is potentially unsafe to call from 'f': one example which > caused trouble is 'options'. Finalizers are scheduled at garbage > collection but only run at a relatively safe time thereafter. > > which makes it very difficult to know what exactly one is allowed to do. A robust finalizer mechanism would be a valuable feature in r7.

Michael Lawrence (17:59:56): > Thanks for the suggestion. The ideal fix would be forreg.finalizer()to defer the finalizer call.

2021-05-03

Malte Thodberg (09:23:37): > @Malte Thodberg has joined the channel

2021-05-11

Megha Lal (16:45:28): > @Megha Lal has joined the channel

2021-10-18

John Kerl (09:16:15): > @John Kerl has joined the channel

2022-01-28

Megha Lal (11:14:14): > @Megha Lal has left the channel

2022-05-05

Aedin Culhane (12:22:55): > @Aedin Culhane has joined the channel

2022-05-18

Vince Carey (06:24:14): > @Vince Carey has left the channel

2023-03-17

Michael Milton (00:48:42): > @Michael Milton has joined the channel

2023-08-04

Lucio Queiroz (14:22:38): > @Lucio Queiroz has joined the channel

2023-08-07

Jiaji George Chen (11:22:13): > @Jiaji George Chen has joined the channel

2023-08-25

Malte Thodberg (10:00:15): > Is the plan for Bioconductor to adopt this in the long term?:https://cran.r-project.org/web/packages/S7/index.html - Attachment (cran.r-project.org): S7: An Object Oriented System Meant to Become a Successor to S3 and S4 > A new object oriented programming system designed to be a successor to S3 and S4. It includes formal class, generic, and method specification, and a limited form of multiple dispatch. It has been designed and implemented collaboratively by the R Consortium Object-Oriented Programming Working Group, which includes representatives from R-Core, ‘Bioconductor’, ‘Posit’/‘tidyverse’, and the wider R community.

2024-08-14

Kylie Bemis (22:34:54): > @Kylie Bemis has joined the channel