[Rd] system()/system2() using short paths of commands on Windows?

[Yihui Xie]

Okay, thanks everyone for the clear explanations! Now I understand it
much better.

Before I wrote to r-devel, I guessed it was probably a problem on TeX
Live's side rather than R, therefore I also reported it to them and
I'm still waiting for a response. If they cannot fix it, and R's
system() has to use short paths, I will resort to workarounds. Many
thanks again!

Regards,
Yihui



On Tue, Oct 31, 2023 at 4:22 AM Tomas Kalibera <tomas.kalibera using gmail.com> wrote:
>
> On 10/30/23 21:36, Yihui Xie wrote:
>
> I have read about "system() not using a shell on Windows" on the help
> page many times before but never understood what it means technically.
> Please forgive my ignorance. I still do not understand it, but thanks
> a lot for the explanation anyway!
>
> As the documentation says, system() on Windows runs the command directly, without a shell (without cmd.exe). As it says,
>
> 'command must be an executable (extensions ‘.exe’, ‘.com’) or a batch file (extensions ‘.cmd’ and ‘.bat’): these extensions are tried in turn if none is supplied. This means that redirection, pipes, DOS internal commands, ... cannot be used: see shell if you want to pass a shell command-line. '
>
> Things like redirection, pipes, etc are implemented by a shell, you can use shell() to run a command via "cmd.exe /c ...", so these would work.
>
> I'm just curious if the full path
> would work in system() today. If it still would not work because
> today's Windows is still like Windows 95 in this aspect, please ignore
> my question and I will ask Microsoft for a refund.
>
> I am not sure if you are asking a general question or specifically still for this use. In principle, short names are still useful to get rid of spaces (sometimes they are not quoted correctly, sometimes they cannot be quoted correctly such as in make). Also short names reduce the risk of running over the path-length limits. So R uses short names when they are available, but supports also long names and tries itself to quote properly.
>
> You have run into a case when an external wrapper has a bug and isn't able to deal with short names. There could easily be other wrappers around, which have a different bug and cannot deal with long names, e.g. because of spaces, when passing them to other programs. Very much like in luatex. And they may have been written in times when they actually were correct.
>
> Best Tomas
>
> Regards,
> Yihui
>
>
>
> On Mon, Oct 30, 2023 at 3:03 PM Prof Brian Ripley <ripley using stats.ox.ac.uk> wrote:
>
> On 30/10/2023 16:18, Yihui Xie wrote:
>
> Hi,
>
> It may have been so for 20+ years but I just discovered today that system()
> would always try to use the short path of a command on Windows:
> https://github.com/wch/r-source/blob/635a67/src/gnuwin32/run.c#L141 If
> that's true, I wonder if it could provide an option to disable this
> behavior, because we recently ran into a case in which short paths wouldn't
> work. I wonder what the original motivation of using short paths was. If it
> was to avoid spaces in paths, wouldn't shQuote() work? Thanks!
>
> No: system on Windows does not use a shell.
>
> The 'original motivation' was to work reliably!  Back in the days of
> Windows 95 when many parts of Windows only supported 8+3 names.
>
> Regards,
> Yihui
>
>       [[alternative HTML version deleted]]
>
> Please do re-read the posting guide.  It has ' been so for 20+ years '.
>
> My apologies! Sometimes I forget to switch to the plain-text mode when
> writing to R mailing lists.
>
> --
> Brian D. Ripley,                  ripley using stats.ox.ac.uk
> Emeritus Professor of Applied Statistics, University of Oxford
>
> ______________________________________________
> R-devel using r-project.org mailing list
> https://stat.ethz.ch/mailman/listinfo/r-devel