[M3commit] FilePosixC.c

[Jay]

Tony, Well, it is a matter of degree, but I really dislike the duplication
and I believe the portability is plenty attainable. It also didn't compile
for me on Cygwin, I did a bit more research and fixed that.
It is ok now for SOLgnu? I do have a SOLgnu/sun machine but didn't rebuild there yet.
L_SET is not portable.
 
 - Jay> From: hosking at cs.purdue.edu> To: jkrell at elego.de> Date: Wed, 17 Dec 2008 06:33:27 +1100> CC: m3devel at elegosoft.com; m3commit at elegosoft.com> Subject: Re: [M3commit] CVS Update: cm3> > This new file FilePosixC.c does not build on my SOLgnu Tinderbox > setup, so the regressions failed last night.> > Jay, this is precisely the reason that we don't want to have C-coded > files in the Modula-3 code-base -- C might proclaim its portability > but that portability is a lie. Why do we need this when the old pure > Modula-3 setup worked just fine! I know it meant some amount of > duplication, but that duplication was needed because of differences in > API specs on different platforms. I am not averse to having > *portable* C code, but in this case I think it is too low-level to be > properly portable. Perhaps better to keep it in Modula-3.> > On 15 Dec 2008, at 18:33, Jay Krell wrote:> > > CVSROOT: /usr/cvs> > Changes by: jkrell at birch. 08/12/15 18:33:17> >> > Modified files:> > cm3/m3-libs/libm3/src/os/POSIX/: FilePosix.i3 FilePosix.m3> > m3makefile> > Added files:> > cm3/m3-libs/libm3/src/os/POSIX/: FilePosixC.c> >> > Log message:> > Remove the need to untangle the #ifdefs around struct flock.> 
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://m3lists.elegosoft.com/pipermail/m3commit/attachments/20081217/526448ee/attachment-0002.html>