[Top][All Lists]
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: [Groff] Re: MSVC Port
From: |
Keith Marshall |
Subject: |
Re: [Groff] Re: MSVC Port |
Date: |
Sun, 30 Nov 2003 12:01:23 +0000 |
On Friday 28 November 2003 10:24 pm, Werner LEMBERG wrote:
> > Just a thought. While the combination of MSYS with MinGW perhaps
> > seems more natural, (both are free software from the same stable),
> > it could be interesting to try MSYS as an alternative to MKS, in
> > conjunction with the MSVC compiler. You probably wouldn't have as
> > many *nix style tools, but it is sufficient to run configure and
> > make, (the shell is bash, make is GNU make), and to create diffs,
> > (with GNU diff).
>
> Indeed, an interesting experiment -- but people are actually using the
> MKS framework, so groff should compile cleanly (with none or only
> minor alterations).
Yes, I agree entirely. Any implied inference to the contrary was
unintentional; I simply wanted to raise the possibility of an alternative
option. Unfortunately, since I do not have MSVC, I cannot attest to its
viability; had I been able, I would have checked it out, before posting the
suggestion.
> Ideally, I would like to support MSVC directly, but this requires a
> lot of work. My idea is to have a script file which creates the
> necessary files for compilation (Makefile, config.h, etc.) instead of
> executing the configure script. This script should be run under Unix
> (at `make dist' time), and output files should be placed into
> arch/msvc to avoid cluttering the source tree. The only prerequisite
> for compilation with MSVC would be to have GNU make (and the necessary
> binaries like gs or netpbm at run time). Since I don't use Windows
> this would be a volunteer's job.
Seems to me that this would likely require a great deal of effort. I don't
intend any disrespect here; I am just interested in your thought process.
If going to this length, should you not also aim to create a makefile which
is compatible with Microsoft's own 'nmake' tool? If you retain the
dependency on GNU make, why not also retain a requirement for a framework
capable of running the configure script? After all, the stated aim of the
MSYS project is to provide a minimal framework for running configure scripts
and GNU make, (which it also provides), to produce native binaries for
Windows-32.
BTW, the other binaries you mention are already freely available for
Windows-32, in the form of native binary packages; (see README.MinGW).
Best regards,
Keith.
Message not available