RFC: New pexecute interface

[E. Weddington]

Ian Lance Taylor wrote:

>As noted in PR 14316, collect2 doesn't build on Windows due to the use
>of vfork.  There have been at least two patches to address this, one
>of them from me, one from Zack.
>
>My patch is here:
>    http://gcc.gnu.org/ml/gcc-patches/2004-03/msg01445.html
>Zack had some comments:
>    http://gcc.gnu.org/ml/gcc-patches/2004-04/msg00227.html
>    http://gcc.gnu.org/ml/gcc-patches/2004-04/msg00637.html
>    http://gcc.gnu.org/ml/gcc-patches/2004-04/msg01044.html
>And at that point I kind of ran out of steam.
>
>Anyhow, I've returned to the issue, and I've tried to put together an
>interface based on Zack's suggestion which can handle everything we
>might want to handle.  I'm not too crazy about it--there are several
>different requirements, and so the interface has again gotten kind of
>complicated.  It implements what collect2 needs, and I think can also
>do most of what gcc.c needs to implement -time, -pipe, and
>--save-temps.
>
>At the moment I've implemented a Unix version.  Versions for other
>operating systems should not be difficult.
>
>  
>

Thanks for taking a look at this again. A solution is badly needed.
Would this be targeted for 4.0, perhaps?

Eric