Date: Thu, 28 Apr 2005 00:04:15 EST-10EDT,10,-1,0,7200,3,-1,0,7200,3600 Subject: [UnixOS2_Archive] No. 542 ************************************************** Wednesday 27 April 2005 Number 542 ************************************************** Subjects for today 1 Test version of Python 2.4.1 compiled with Innotek GCC : Paul Smedley 2 Re: Test version of Python 2.4.1 compiled with Innotek GCC : lsunley at mb.sympatico.ca 3 Building Freetype2 with Innotek_libc : Dave Yeo" 4 Re: Building Freetype2 with Innotek_libc : Andy Willis 5 Re: Test version of Python 2.4.1 compiled with Innotek GCC : Paul Smedley **= Email 1 ==========================** Date: Wed, 27 Apr 2005 07:15:04 +0930 From: Paul Smedley Subject: Test version of Python 2.4.1 compiled with Innotek GCC Hi All, Haven't had time to test this at all yet, but I finally got around to looking at Python again using Innotek GCC and was totally successful in getting a clean compile this time - including the dynamically loaded modules (damn underscores!! :P) If anyone has a chance to have a play - they can grab the zip file from http://smedley.info/python241-os2-innotekgcc.zip 5,889,534 bytes Cheers, Paul. **= Email 2 ==========================** Date: Tue, 26 Apr 2005 21:41:50 -0500 From: lsunley at mb.sympatico.ca Subject: Re: Test version of Python 2.4.1 compiled with Innotek GCC Thanks Downloading now.... Now I can get plpython built for postgreSQL Lorne In <426EB660.1090402 at smedley.info>, on 04/27/05 at 07:15 AM, Paul Smedley said: >Hi All, >Haven't had time to test this at all yet, but I finally got around to >looking at Python again using Innotek GCC and was totally successful in >getting a clean compile this time - including the dynamically loaded >modules (damn underscores!! :P) >If anyone has a chance to have a play - they can grab the zip file from >http://smedley.info/python241-os2-innotekgcc.zip 5,889,534 bytes >Cheers, >Paul. -- ----------------------------------------------------------- lsunley at mb.sympatico.ca ----------------------------------------------------------- **= Email 3 ==========================** Date: Tue, 26 Apr 2005 20:19:37 -0800 From: "Dave Yeo" Subject: Building Freetype2 with Innotek_libc While attempting to build freetype-2.1.9 with Innotek_libc 6b4 I see these errors (going the configure route here) gcc -ansi -I/usr/src/freetype-2.1.9/objs -I./builds/unix -I/usr/src/freetype-2.1.9/include -c -Wall -D__EMX__ -DOS2 -Zmtd -D__ST_MT_ERRNO__ -Zexe -O2 -fomit-frame-pointer -Zbin-files -Dstrncasecmp=strnicmp -Dstrcasecmp=stricmp -DFT_CONFIG_OPTION_SYSTEM_ZLIB -DDARWIN_NO_CARBON "-DFT_CONFIG_CONFIG_H=" /usr/src/freetype-2.1.9/src/base/ftsystem.c -DDLL_EXPORT -DPIC -o /usr/src/freetype-2.1.9/objs/.libs/ftsystem.o In file included from I:/usr/include/ctype.h:29, from /usr/src/freetype-2.1.9/include/freetype/config/ftstdlib.h:74, from builds/unix/ftconfig.h:43, from /usr/src/freetype-2.1.9/src/base/ftsystem.c:29: I:/usr/include/_ctype.h:162: error: syntax error before "unsigned" I:/usr/include/_ctype.h:169: error: syntax error before "int" I:/usr/include/_ctype.h:174: error: syntax error before "int" I:/usr/include/_ctype.h:181: error: syntax error before "__ct_rune_t" I:/usr/include/_ctype.h:188: error: syntax error before "__ct_rune_t" I:/usr/include/_ctype.h:196: error: syntax error before "unsigned" I:/usr/include/_ctype.h:203: error: syntax error before "int" I:/usr/include/_ctype.h:208: error: syntax error before "__wint_t" I:/usr/include/_ctype.h:215: error: syntax error before "__wint_t" I:/usr/include/_ctype.h:223: error: syntax error before "int" make: *** [/usr/src/freetype-2.1.9/objs/ftsystem.lo] Error 1 All apparently related to static inline statements. The relevent part of the makefiles is this # ANSIFLAGS: Put there the flags used to make your compiler ANSI-compliant. # ANSIFLAGS := -pedantic -ansi Removing the -ansi allows the compile to finish with a lot of warnings, removing both gets rid of the warnings. This is true with going the configure route or letting make discover we're on OS/2 (need to set MAKESHELL=cmd.exe). With PGCC 2.95.3 the build works fine out of the box (no warnings) gcc 2.8.1 doesn't like the -fno-strict-aliasing. Is this to be expected wiht Innotek_libc+gcc? And is there any drawbacks in removing the -ansi? Dave **= Email 4 ==========================** Date: Tue, 26 Apr 2005 23:02:16 -0600 From: Andy Willis Subject: Re: Building Freetype2 with Innotek_libc Dave Yeo wrote: >While attempting to build freetype-2.1.9 with Innotek_libc 6b4 I see >these errors (going the configure route here) > gcc -ansi -I/usr/src/freetype-2.1.9/objs -I./builds/unix >-I/usr/src/freetype-2.1.9/include -c -Wall -D__EMX__ -DOS2 -Zmtd >-D__ST_MT_ERRNO__ -Zexe -O2 -fomit-frame-pointer -Zbin-files >-Dstrncasecmp=strnicmp -Dstrcasecmp=stricmp >-DFT_CONFIG_OPTION_SYSTEM_ZLIB -DDARWIN_NO_CARBON >"-DFT_CONFIG_CONFIG_H=" >/usr/src/freetype-2.1.9/src/base/ftsystem.c -DDLL_EXPORT -DPIC -o >/usr/src/freetype-2.1.9/objs/.libs/ftsystem.o >In file included from I:/usr/include/ctype.h:29, > from >/usr/src/freetype-2.1.9/include/freetype/config/ftstdlib.h:74, > from builds/unix/ftconfig.h:43, > from /usr/src/freetype-2.1.9/src/base/ftsystem.c:29: >I:/usr/include/_ctype.h:162: error: syntax error before "unsigned" >I:/usr/include/_ctype.h:169: error: syntax error before "int" >I:/usr/include/_ctype.h:174: error: syntax error before "int" >I:/usr/include/_ctype.h:181: error: syntax error before "__ct_rune_t" >I:/usr/include/_ctype.h:188: error: syntax error before "__ct_rune_t" >I:/usr/include/_ctype.h:196: error: syntax error before "unsigned" >I:/usr/include/_ctype.h:203: error: syntax error before "int" >I:/usr/include/_ctype.h:208: error: syntax error before "__wint_t" >I:/usr/include/_ctype.h:215: error: syntax error before "__wint_t" >I:/usr/include/_ctype.h:223: error: syntax error before "int" >make: *** [/usr/src/freetype-2.1.9/objs/ftsystem.lo] Error 1 > >All apparently related to static inline statements. >The relevent part of the makefiles is this ># ANSIFLAGS: Put there the flags used to make your compiler >ANSI-compliant. ># >ANSIFLAGS := -pedantic -ansi > >Removing the -ansi allows the compile to finish with a lot of warnings, >removing both gets rid of the warnings. This is true with going the >configure route or letting make discover we're on OS/2 (need to set >MAKESHELL=cmd.exe). >With PGCC 2.95.3 the build works fine out of the box (no warnings) gcc >2.8.1 doesn't like the -fno-strict-aliasing. >Is this to be expected wiht Innotek_libc+gcc? And is there any >drawbacks in removing the -ansi? >Dave > > > > It built fine with ftjam. Andy **= Email 5 ==========================** Date: Wed, 27 Apr 2005 20:47:21 +0930 From: Paul Smedley Subject: Re: Test version of Python 2.4.1 compiled with Innotek GCC Hi All, I've now refreshed this build to fix a few problems that popped up when I tested tonight. The build now also includes support for zlib and crypt. I aim to include the other modules Andy was including too when I get more time. As I write, I'm using this build of Python to download using Bittorrent so it is at least partly functional :) Any comments appreciated! Cheers, Paul. Paul Smedley wrote: > Hi All, > Haven't had time to test this at all yet, but I finally got around to > looking at Python again using Innotek GCC and was totally successful in > getting a clean compile this time - including the dynamically loaded > modules (damn underscores!! :P) > > If anyone has a chance to have a play - they can grab the zip file from > http://smedley.info/python241-os2-innotekgcc.zip 5,889,534 bytes > > Cheers, > > Paul. >