[Rd] R 3.1.0: 'R CMD Sweave' deletes non tex files created upon batch mode exit

Martin Maechler maechler at stat.math.ethz.ch
Thu Apr 17 11:22:04 CEST 2014

>>>>> Duncan Murdoch <murdoch.duncan at gmail.com>
>>>>>     on Tue, 15 Apr 2014 08:54:34 -0400 writes:

    > On 15/04/2014 8:12 AM, peter dalgaard wrote:
    >> On 15 Apr 2014, at 13:14 , Duncan Murdoch <murdoch.duncan at gmail.com> wrote:
    >> > On 14/04/2014, 10:23 AM, Marc Schwartz wrote:
    >> >> Hi Martin,
    >> >>
    >> >> Thanks for your confirmation on this.
    >> >>
    >> >> I normally do not use R CMD Sweave, as I too run under ESS in normal day to day operations. This finding was a quirk of having a particular Rnw document that I occasionally run using R CMD Sweave and I had done so over the weekend, realizing this behavior.
    >> >
    >> > This sounds like an argument for dropping R CMD Sweave, rather than fixing it.  The bug was introduced in July, 2013, and nobody noticed it because so few people use that feature, and apparently nobody who does use it bothers to test pre-release versions.
    >> >
    >> I'd say that that cat is out of the bag. There are probably umpteen documents around suggesting "R CMD Sweave". As people use Sweave only sporadically, it could take years before the old usage got stamped out. And anyways, the command format is the obvious way to generate documents in scripts and makefiles, isn't it?

    > I use Rscript, and I expect most others do too, but I was wrong about 
    > how long this went unreported.  Martin Morgan reported it in February in 
    > R-devel and I didn't notice.  He even gave a link to it in a message in 
    > March on another topic; I replied to the March message, but didn't 
    > follow the link.

    > We discourage people from using bugs.r-project.org for pre-release 
    > issues; perhaps we shouldn't do that.

    > Duncan Murdoch

I agree:  Rather we should encourage such volunteer testers reporting
such problems.

The 'R version' is already part of the bug report form;
I'm not sure we can easily *list* all the "R version == R-devel"
bugs easily via web interface, but even w/o that,
we should rather be glad for such reports.

PS: I'm currently testing a patch where 'R CMD Sweave' will
    revert to not deleting anything after running the R code by default.

Martin Maechler

More information about the R-devel mailing list