On Wed, 2003-02-26 at 09:32, Bill Moseley wrote:
> > Does POSIX standard mention "_timeb", "_ftime"? I did not find them.
> I've sent mail to the person that added that code asking why that was
> used. I wonder if it was because of Windows (as it was originally wrapped
> in a #ifdef WIN32.
I suspect that was an oversight on my part when I replaced the #ifdef
WIN32 with a more general version. Windows prefixes Standard C
functions with _ unless STDC (or something like that) is defined. It
should be safe to remove the _ prefix completely.
> > 2. if I not specify "./configure --with-zlib=/usr/local [...some
> > stuff...]", the generated Makefile does NOT specify "-lz" and symbols
> > "compress/uncompress" can NOT be resolved; configure does not produce a
> > right Makefile if "zlib" is not explicitly mentioned
That is a problem with the autoconf's detection of zlib. Zlib doesn't
provide a zlib-config script to provide complete linker and cpp
dependency information. I have it automatically detect whether or not
zlib exists but it can fail to find the location. Any improvements
would be welcome.
> I remember once on Solaris that the zlib header file was found but the
> library was installed in a non-standard location so building failed.
I think maybe that's the same problem. In that case cpp found the
header but the linker did not find a library. The test program ends up
compiled but not linked to the correct libraries. So, incorrect
information is returned to configure. Sounds like a misconfigured
compiler to me. That is probably not uncommon on systems without
vendor-provided builds of GCC. The cpp: and link: sections of the GCC
specs file can be changed to fix those sorts of problems. That might
explain why it normally builds fine on Linux, *BSD, MacOSX, etc but
randomly fails on Solaris, HPUX, etc.
ICQ - 412039
Received on Wed Feb 26 15:53:07 2003