Mail Archives: djgpp/2010/04/06/17:00:18
Hi again,
On Apr 2, 12:21=A0pm, Rugxulo <rugx DOT DOT DOT AT gmail DOT com> wrote:
>
> On Apr 2, 1:10=A0am, Eli Zaretskii <e DOT DOT DOT AT gnu DOT org> wrote:
>
> > > From: Rugxulo <rugx DOT DOT DOT AT gmail DOT com>
> > > Date: Thu, 1 Apr 2010 15:23:22 -0700 (PDT)
> > > Bytes: 4549
>
> > > The Autoconf people explicitly target only POSIX, if at all possible.
>
> > That's not accurate: they explicitly cater to environments such as
> > Cygwin and MSYS which run on Windows. =A0And at least in the past, when
> > DJGPP was actively developed, our voice was also heard.
>
> > Besides, DJGPP is sufficiently Posix for that matter, anyway.
One (non-GNU) third-party app in particular won't port to DJGPP
because it needs mmap(), which I think I was told is default in a
newer POSIX version. Checking the newsgroup archives here shows that
DPMI 1.0 could maybe support such emulation, but nobody has bothered
(for the big obvious reason, it's not common).
Anyways, I think even the original Bourne shell itself isn't POSIX, so
Autoconf dudes don't support it. (Sounds odd to hear, but it's true!)
They even avoid using "test -z" because it's not always available.
BTW, Cygwin is allegedly 30% faster since Autoconf 2.64, but (for me)
DJGPP is still dog slow when configuring latest ZILE. It's just crazy
to check for so many things, it's just too slow. It takes longer to
configure than compile! Perhaps the default config.site should have
more pre-defined variables for some common things? Or maybe we need a
faster shell (Dash or Busybox Ash)? Or maybe we need a sh2pl or sh2py
translator?
I don't want DJGPP (support) to die, it's too dang useful. But even
when all the tools work (G++, GNU Make, Bison, Flex, Perl, PCRE,
SQLite, Lua), some developers still abandon it (e.g. latest Dungeon
Crawl: Stone Soup). Heck, I even (barely) built them a working binary,
and they still don't seem to care. FYI, Dungeon Crawl originated on
DJGPP! Oh well, another one bites the dust. ;-(
- Raw text -