[R-pkg-devel] Rd files: using \link[pkg]{foo} when file names differ between OSs

[Martin Morgan]

On 04/16/2018 12:31 PM, Duncan Murdoch wrote:
> On 16/04/2018 12:06 PM, Martin Maechler wrote:
>>>>>>> Duncan Murdoch <murdoch.duncan at gmail.com>
>>>>>>>      on Mon, 16 Apr 2018 11:52:10 -0400 writes:
>>
>>      > On 16/04/2018 11:35 AM, Ramon Diaz-Uriarte wrote:
>>      >> Dear All,
>>      >>
>>      >> Two recent threads in the bioconductor devel mailing list
>>      >> (https://stat.ethz.ch/pipermail/bioc-devel/2018-April/013156.html
>>      >> and
>>      >> https://stat.ethz.ch/pipermail/bioc-devel/2018-April/013259.html)
>>      >> are related to packages that have different names of html
>>      >> files in different operating systems.
>>      >>
>>      >> For example, parallel has a file called mclapply in
>>      >> Linux. So using, from the Rd file of another package,
>>      >> \link[parallel]{mclapply} works fine under Linux, but
>>      >> does not under Windows, because there is no mclapply.html
>>      >> file in Windows (there is a mcdummies file).
>>      >>
>>      >>
>>      >> Is there any recommended way to proceed in these cases?
>>      >>
>>      >>
>>      >> Yes, section 2.5 of Writing R Extensions indicates that
>>      >> \link[pkg]{foo} and \link[pkg:bar]{foo} are rarely
>>      >> needed; so the simplest way to proceed would be to avoid
>>      >> \link[pkg]{foo} and \link[pkg:bar]{foo}. I am asking for
>>      >> the cases where, as noted in 2.5, "more than one package
>>      >> offers help on a topic".
>>
>>      > You could make the links conditional on the OS.  For example,
>>
>>      > #ifdef windows
>>      > See \link[parallel]{mcdummies}.
>>      > #endif
>>      > #ifdef unix
>>      > See \link[parallel]{mclapply}.
>>      > #endif
>>
>>      > The other possibility (useful if there are major differences 
>> between the
>>      > platforms) is to have two copies of the help file, one in 
>> man/unix, one
>>      > in man/windows, but that doesn't seem appropriate from your 
>> description.
>>
>>      > Duncan Murdoch
>>
>> and mid-term, I really think R and (CRAN, Bioc, ...) packages
>> should not do what we (R core) did here.
>> Rather,  \alias{mclapply}  should exist both for windows and
>> non-windows, and hence \link{mclapply}  would just work.

\alias{mclapply} does exist...

> 
> I forget whether that would work here:  parallel being a base package 
> (used by the package in question?) might mean it would be found without 
> the [parallel] in the link.  But in general, links to other packages 
> using [pkg] go to the *filename*, not to the alias.  This oddity happens 
> because we want the links to work even if the referenced package is 
> installed later than the Rd file is processed.

People are quite concerned about fixing the 'WARNING' this generates. 
However from the text

   file link 'mclapply' in package 'parallel' does not exist and so has 
been  treated as a topic

the help pages are actually constructed correctly, finding the page on 
which the topic (aka alias) 'mclapply' is defined. This contrasts with a 
completely incorrect link (e.g., \link[stats]{mclapply}) generating a 
warning

   missing file link

and unable to link to the mclapply topic.

Perhaps the WARNING from \link[parallel]{mclapply} should be a NOTE?

There are other oddities in the threads that Ramon indicates

   - WARNINGs from \link[foo]{bar} that should really be 
\link[foo:baz]{bar} often only appear on Windows (parallel's use of 
mcdummies is a special case here). Shouldn't they be 
platform-independent? An example is

 
http://bioconductor.org/checkResults/3.7/bioc-LATEST/ADaCGH2/tokay2-checksrc.html

where as.MAList is defined in a file called asmalist.Rd so 
\link[limma]{as.MAList} is incorrect. It generates a warning only on our 
Windows machine (tokay2) not Linux or Mac. I had some recollection of 
Windows-specific help system behavior, but I think this dates back to 
the .chm days...

   - There were a spate of independent posts about this, suggesting that 
this is relatively new phenomenon (though it could also be that 
maintainers have been busy preparing their packages for the next 
release, so are now noticing the problem...)

   - It doesn't seem like good practice to link to the file name, which 
seems an internal aspect of a package, rather than to the alias, which 
seems at least for symbols to be a user-facing public aspect of the package.

   - And finally I did make an svn commit related to this, so am 
sensitive to blundering here

R-devel/src/library/tools/R$ svn log -c74129
------------------------------------------------------------------------
r74129 | morgan | 2018-01-17 16:52:53 -0500 (Wed, 17 Jan 2018) | 10 lines

correct warning when \link[base]{foo} is incorrect

- now "missing file link" rather than "file link 'foo' in package 'base'
   does not exist and will be treated as a topic" (foo is _not_ a topic
   in base; the link to foo.html is missing)
- vice-versa for \link[base]{rbind} (rbind _is_ a topic in base, 
documented in
   cbind.html, and will be resolved by the help system)
- see https://stat.ethz.ch/pipermail/r-devel/2009-October/055287.html and
   c50198

------------------------------------------------------------------------

   - There is a possibility that some of the oddities are Bioc 
build-system specific, and I really have not had a chance to dig to the 
bottom of this to my satisfaction.

Martin


> 
> The reason the file was named "mclapply" in man/unix but "mcdummies" in 
> man/windows was likely to save some time:  there are several aliases 
> documented in that one file on Windows, but they are in separate files 
> on Unix.
> 
> Duncan Murdoch
> 
>>
>> Martin Maechler
>>
>>
>>      >> Thanks,
>>      >> R.
>>
>>      >> --
>>      >> Ramon Diaz-Uriarte
>>      >> Department of Biochemistry, Lab B-25
>>      >> Facultad de Medicina
>>      >> Universidad Autónoma de Madrid
>>      >> Arzobispo Morcillo, 4
>>      >> 28029 Madrid
>>      >> Spain
>>
> 
> ______________________________________________
> R-package-devel at r-project.org mailing list
> https://stat.ethz.ch/mailman/listinfo/r-package-devel


This email message may contain legally privileged and/or...{{dropped:2}}