[Bioc-devel] Vignettes and package size

Hi I am the developper of the packages DAPAR and Prostar. Each of these packages contains two vignettes : a user manual and an tutorial. As these documents contain screen shots, I have to join the images (png files) in the packages in addition of the Rnw and pdf files. Unfortunately, the size

[Rd] R-exts update for new *check* argument in R 3.3.0

Hello dear dev team, In the recent devel R-exts manual I cannot find any information about new option `--ignore-vignettes` to `R CMD check`. I would expect it to be mentioned somewhere in "1.4.2 Non-Sweave vignettes". Checked on https://cran.r-project.org/doc/manuals/r-devel/R-exts.html and

Re: [Rd] S3 dispatch for S4 subclasses only works if variable "extends" is accessible from global environment

Agreed about Rscript being consistent R. For now, I'll modify S4_extends() so that it leads to S4 dispatch when dispatch is turned on (not just when methods is attached). On Tue, Apr 19, 2016 at 8:37 AM, Gabriel Becker wrote: > Does it make sense to be able to load an S4

Re: [Rd] S3 dispatch for S4 subclasses only works if variable "extends" is accessible from global environment

On 19 April 2016 at 11:21, Hadley Wickham wrote: | This might be too big a change - but is it worth reconsidering the | behaviour of Rscript? Maybe the simplest fix would be simply to always | load the methods package. (I think historically it didn't because | loading methods took a long time,

Re: [Rd] S3 dispatch for S4 subclasses only works if variable "extends" is accessible from global environment

On Tue, Apr 19, 2016 at 9:21 AM, Hadley Wickham wrote: > > This might be too big a change - but is it worth reconsidering the > behaviour of Rscript? Maybe the simplest fix would be simply to always > load the methods package. (I think historically it didn't because >

Re: [Rd] S3 dispatch for S4 subclasses only works if variable "extends" is accessible from global environment

This might be too big a change - but is it worth reconsidering the behaviour of Rscript? Maybe the simplest fix would be simply to always load the methods package. (I think historically it didn't because loading methods took a long time, but that is no longer true) Hadley On Tue, Apr 19, 2016

Re: [Rd] S3 dispatch for S4 subclasses only works if variable "extends" is accessible from global environment

Does it make sense to be able to load an S4 object without the methods package being attached? I'm not sure implementation-wise how easy this would be, but it seems like any time there is an S4 object around, the methods package should be available to deal with it. ~G On Tue, Apr 19, 2016 at

Re: [Rd] S3 dispatch for S4 subclasses only works if variable "extends" is accessible from global environment

Right, R_has_methods_attached() uses that. Probably not the right check, since it refers to S4 dispatch, while S4_extends() is used by S3 dispatch. Perhaps S4_extends() should force load the methods package? The above example works after fixing the check to ensure that R_MethodsNamespace is not

Re: [Rd] S3 dispatch for S4 subclasses only works if variable "extends" is accessible from global environment

See also .isMethodsDispatchOn, which is what trace uses to decide if the methods package needs to be loaded. ~G On Tue, Apr 19, 2016 at 5:34 AM, Michael Lawrence wrote: > Not sure why R_has_methods_attached() exists. Maybe Martin could shed > some light on that. > >

Re: [Rd] S3 dispatch for S4 subclasses only works if variable "extends" is accessible from global environment

Not sure why R_has_methods_attached() exists. Maybe Martin could shed some light on that. On Mon, Apr 18, 2016 at 11:50 PM, Kirill Müller wrote: > Thanks for looking into it, your approach sounds good to me. See also > R_has_methods_attached() >

Re: [Rd] S3 dispatch for S4 subclasses only works if variable "extends" is accessible from global environment

Thanks for looking into it, your approach sounds good to me. See also R_has_methods_attached() (https://github.com/wch/r-source/blob/42ecf5f492a005f5398cbb4c9becd4aa5af9d05c/src/main/objects.c#L258-L265). I'm fine with Rscript not loading "methods", as long as everything works properly with