Why does the source use `int64_t'. Is this suppose to be built with GCC? I have Mumit Kahns gcc, version 2.8.1. Do I need a 2.9.5 version to get the `int64_t' type? I have substituted a typedef of `long long' in os_types.h. This seems to make sense? The rint() macro also seems to be having some problems. I guess there are two meanings to __GNUC__ and _WIN32. One is the OS and the other is the compiler type. I tried the .dsw file in the distribution with VC6.00.2???, but nothing happened when I attempted to load the workspace. I would rather build with gcc anyways. Thanks, Bill -- Tune your stereo. Do something nice for a lout tomorrow. This month food would be a good thing to try. You'll sometimes be happy. --- >8 ---- List archives: xiph.org/archives Ogg project homepage: xiph.org/ogg
>>>>> "Bill" == Bill Pringlemeir <bpringle@teklogix.com> writes:[snip] Bill> The rint() macro also seems to be having some problems. I Bill> guess there are two meanings to __GNUC__ and _WIN32. One is Bill> the OS and the other is the compiler type. That is weird. The file `vorbis/lib/psytune.c' has a variable declared floor. This is hiding the function named floor(). I am unsure which one is right for rint(), but it would seem to be the `C' library function. I guess that I have produced both libraries. (libvorbis.a and vorbisfile.a) Is psytune to be run afterwards to tune some header files? Thanks again, Bill -- Have you ever had large lymph nodes tell you that your slave is Mars? Or been transported to an alternate reality where you can generate psychokinetic energy without following boxes of Borax? You will. And the company that will bring it to you: AT&T. --- >8 ---- List archives: xiph.org/archives Ogg project homepage: xiph.org/ogg
> Why does the source use `int64_t'.It needs a 64 bit type an int64_t is specified by POSIX and was supported by SysV and BSD anyway.> Is this suppose to be built > with GCC?It's meant to be compiled by a POSIX compiler, of which GCC is one. In most cases, though, this is a header issue, not a compiler issue.> I have Mumit Kahns gcc, version 2.8.1. Do I need a > 2.9.5 version to get the `int64_t' type?No, you need appropriate libc headers.> I have substituted a > typedef of `long long' in os_types.h. This seems to make sense?Sure, but this should be checked for in autoconf, not os_types.h, because there are several million possible OS/Compiler/Header variation/combinations, and we're not going to stuff them all in os_types.h. That's for systems (like win32) that nominally do not use autoconf.> The rint() macro also seems to be having some problems. I guess > there are two meanings to __GNUC__ and _WIN32. One is the OS and > the other is the compiler type.Ah, this is Windows, OK. Actually there has already been a proposed fix to the win32 entries in os_types.h to check for GCC and !UNIX... Monty --- >8 ---- List archives: xiph.org/archives Ogg project homepage: xiph.org/ogg
>Why does the source use `int64_t'. Is this suppose to be built >with GCC? I have Mumit Kahns gcc, version 2.8.1. Do I need a >2.9.5 version to get the `int64_t' type? I have substituted a >typedef of `long long' in os_types.h. This seems to make sense?I have simpler question. How did anyone building on windows generate the makefiles? I tried running the configure using bash or sh from the cygwin stuff but the script bails out with configure: error: can not run ./config.sub when attempting to run it. Dale --- >8 ---- List archives: xiph.org/archives Ogg project homepage: xiph.org/ogg
Hi Dale, I managed to get it to build using gcc-2.95.2-2 (mingw32), but had to resort to a little trickery. I ran the configure script on a linux box, and then transferred the makefiles to my win32 box. If you'd like, I can send you the makefiles, etc. that I use for mingw32 -- they should work with a minimum amount of modification for cygwin. Andy> >Why does the source use `int64_t'. Is this suppose to be built > >with GCC? I have Mumit Kahns gcc, version 2.8.1. Do I need a > >2.9.5 version to get the `int64_t' type? I have substituted a > >typedef of `long long' in os_types.h. This seems to make sense? > >I have simpler question. How did anyone building on windows generate the >makefiles? I tried running the configure using bash or sh from the cygwin >stuff but the script bails out with > >configure: error: can not run ./config.sub > >when attempting to run it. > >Dale > >--- >8 ---- >List archives: xiph.org/archives >Ogg project homepage: xiph.org/ogg________________________________________________________________________ Get Your Private, Free E-mail from MSN Hotmail at hotmail.com --- >8 ---- List archives: xiph.org/archives Ogg project homepage: xiph.org/ogg