Dear GPCers:
I recently was forced to upgrade to Yosemite and lost my GPC compiler.
I tried to compile according to
* GNU Pascal 3.4.6u4 compiler sources and double-clickable building script (for Mac OS X 10.3 - 10.6) (22.2 MB)
and got: % build-gpc.command setenv GPC_BUILD_TOOLS 1 setenv GPC_BUILD_GP 1 setenv GPC_BUILD_GPC_SOURCES 1 setenv GPC_BUILD_GPC_CLEAN 1 setenv GPC_BUILD_GPC_CONFIGURE 1 setenv GPC_BUILD_GPC 1 setenv GPC_BUILD_GPC_DYLIB 1 setenv GPC_BUILD_GPC_CLEAN_CROSS 1 setenv GPC_BUILD_GPC_CONFIGURE_CROSS 1 setenv GPC_BUILD_GPC_CROSS 1 setenv GPC_BUILD_GPC_DYLIB_CROSS 1 setenv GPC_BUILD_GPC_DOCS 1 setenv GPC_BUILD_TESTSUITE 0 setenv GPC_BUILD_PACKAGE 0 setenv GPC_BISON bison-2.3 setenv GPC_FLEX flex-2.5.35 setenv GPC_HELP2MAN help2man-1.36.4 setenv GPC_SED sed-4.1.4 setenv GPC_GCC 3.4.6 setenv GPC_GCC_DIFF gcc-3.4.4.diff setenv GPC_SNAPSHOT gpc-20070904 setenv GPC_PDF gpc-20070904.pdf setenv GPC_RELEASE 346u4 setenv GPC_PACKAGE 3.4.6u4 setenv GPC_FPSAVE fpsave-20051224 setenv GPC_GP gp-0.61 setenv GPC_DIR /Developer/Pascal setenv GPC_GPC gpc346u4 setenv GPC_DYLIB 346u4 setenv GPC_CONTACT -- Please contact Adriaan van Os at gpc@microbizz.nl setenv GPC_OSX_VERSION `uname -r` uname -r switch ( 14.5.0 ) unset echo
Unforeseen version of Mac OS X. -- Please contact Adriaan van Os at gpc@microbizz.nl
-----
Adriaan has not responded for several days.
I don't see how to get a binary.
Has anyone migrated to 10.10.5? How do I do this?
Thanks,
Tom Schneider
Thomas D. Schneider, Ph.D. Senior Investigator National Institutes of Health National Cancer Institute Center for Cancer Research Gene Regulation and Chromosome Biology Laboratory Molecular Information Theory Group Frederick, Maryland 21702-1201 schneidt@mail.nih.gov http://schneider.ncifcrf.gov (current link) http://alum.mit.edu/www/toms (permanent link)
Hello all GPC gurus:
I recently sent out a posting about GPC not working on Mac OS X 10.10.5. I got no response. Is anyone listening?
Tom
Thomas D. Schneider, Ph.D. Senior Investigator National Institutes of Health National Cancer Institute Center for Cancer Research Gene Regulation and Chromosome Biology Laboratory Molecular Information Theory Group Frederick, Maryland 21702-1201 schneidt@mail.nih.gov https://schneider.ncifcrf.gov (current link) https://alum.mit.edu/www/toms (permanent link)
On 6 Oct 2015 at 14:11, Schneider wrote:
Hello all GPC gurus:
I recently sent out a posting about GPC not working on Mac OS X 10.10.5. I got no response. Is anyone listening?
Yes. But "anyone" does not have a Mac, and so cannot comment ;-)
Best regards, The Chief -------- Prof. Abimbola A. Olowofoyeku (The African Chief) web: http://www.greatchief.plus.com/
I'm here but I don't code the compiler; just used to be a tester and discussed different things with the GPC group more than 10 years ago. Just re-joined the group a few weeks ago.
And, sorry. Don't know MAC at all. Just wanted to let you know someone was indeed out here.
-Ken
On Tue, Oct 6, 2015 at 1:11 PM, Schneider schneidt@mail.nih.gov wrote:
Hello all GPC gurus:
I recently sent out a posting about GPC not working on Mac OS X 10.10.5. I got no response. Is anyone listening?
Tom
Thomas D. Schneider, Ph.D. Senior Investigator National Institutes of Health National Cancer Institute Center for Cancer Research Gene Regulation and Chromosome Biology Laboratory Molecular Information Theory Group Frederick, Maryland 21702-1201 schneidt@mail.nih.gov https://schneider.ncifcrf.gov (current link) https://alum.mit.edu/www/toms (permanent link)
Gpc mailing list Gpc@gnu.de https://www.g-n-u.de/mailman/listinfo/gpc
On 28/09/15 21:58, Schneider, Thomas (NIH/NCI) [E] wrote:
Dear GPCers:
I recently was forced to upgrade to Yosemite and lost my GPC compiler.
I tried to compile according to
- GNU Pascal 3.4.6u4 compiler sources and double-clickable building script (for Mac OS X 10.3 - 10.6) (22.2 MB)
and got: % build-gpc.command setenv GPC_BUILD_TOOLS 1 setenv GPC_BUILD_GP 1 setenv GPC_BUILD_GPC_SOURCES 1 setenv GPC_BUILD_GPC_CLEAN 1 setenv GPC_BUILD_GPC_CONFIGURE 1 setenv GPC_BUILD_GPC 1 setenv GPC_BUILD_GPC_DYLIB 1 setenv GPC_BUILD_GPC_CLEAN_CROSS 1 setenv GPC_BUILD_GPC_CONFIGURE_CROSS 1 setenv GPC_BUILD_GPC_CROSS 1 setenv GPC_BUILD_GPC_DYLIB_CROSS 1 setenv GPC_BUILD_GPC_DOCS 1 setenv GPC_BUILD_TESTSUITE 0 setenv GPC_BUILD_PACKAGE 0 setenv GPC_BISON bison-2.3 setenv GPC_FLEX flex-2.5.35 setenv GPC_HELP2MAN help2man-1.36.4 setenv GPC_SED sed-4.1.4 setenv GPC_GCC 3.4.6 setenv GPC_GCC_DIFF gcc-3.4.4.diff setenv GPC_SNAPSHOT gpc-20070904 setenv GPC_PDF gpc-20070904.pdf setenv GPC_RELEASE 346u4 setenv GPC_PACKAGE 3.4.6u4 setenv GPC_FPSAVE fpsave-20051224 setenv GPC_GP gp-0.61 setenv GPC_DIR /Developer/Pascal setenv GPC_GPC gpc346u4 setenv GPC_DYLIB 346u4 setenv GPC_CONTACT -- Please contact Adriaan van Os at gpc@microbizz.nl setenv GPC_OSX_VERSION `uname -r` uname -r switch ( 14.5.0 ) unset echo
Unforeseen version of Mac OS X. -- Please contact Adriaan van Os at gpc@microbizz.nl
Adriaan has not responded for several days.
I don't see how to get a binary.
Has anyone migrated to 10.10.5? How do I do this?
Thanks,
Tom Schneider
Thomas D. Schneider, Ph.D. Senior Investigator National Institutes of Health National Cancer Institute Center for Cancer Research Gene Regulation and Chromosome Biology Laboratory Molecular Information Theory Group Frederick, Maryland 21702-1201 schneidt@mail.nih.gov http://schneider.ncifcrf.gov (current link) http://alum.mit.edu/www/toms (permanent link)
Gpc mailing list Gpc@gnu.de https://www.g-n-u.de/mailman/listinfo/gpc
Hi Tom,
I don't own, or have access to a Mac, so cannot help directly.
The build error you reported;
.... setenv GPC_OSX_VERSION `uname -r` uname -r switch ( 14.5.0 ) unset echo
Unforeseen version of Mac OS X
suggests to me that the build is aborting because the version of OS X in use is 'Unforeseen'. That does not in itself imply that it could not work. Have you tried just hacking out this check? Maybe force “setenv GPC_OSX_VERSION” to be a version the script is happy with instead of using “uname -r”.
It might be worth asking for assistance on a Mac forum. The number of Mac users that can handle scripts, is probably much higher than the number of GPC users that are familiar with OS X.
Regards, Peter
More specifically, the immediate problem is the switch statement in build-gpc.command
As the OS version appears to be 14*, the 'default' [fail] option is used.
setenv GPC_OSX_VERSION `uname -r` switch ($GPC_OSX_VERSION) case "10*": setenv GPC_DARWIN darwin10 setenv GPC_OSX_VERSION_NAME 10.6 setenv GPC_SDK /Developer/SDKs/MacOSX10.5.sdk setenv GPC_ALWAYS_USE_SDK 1 setenv GPC_PRE_TIGER 0 breaksw case "9*": setenv GPC_DARWIN darwin9 setenv GPC_OSX_VERSION_NAME 10.5 setenv GPC_SDK /Developer/SDKs/MacOSX10.5.sdk setenv GPC_ALWAYS_USE_SDK 0 setenv GPC_PRE_TIGER 0 breaksw case "8*": setenv GPC_DARWIN darwin8 setenv GPC_OSX_VERSION_NAME 10.4 setenv GPC_SDK /Developer/SDKs/MacOSX10.4u.sdk setenv GPC_ALWAYS_USE_SDK 0 setenv GPC_PRE_TIGER 0 breaksw case "7*": setenv GPC_DARWIN darwin7 setenv GPC_OSX_VERSION_NAME 10.3 setenv GPC_SDK setenv GPC_ALWAYS_USE_SDK 0 setenv GPC_PRE_TIGER 1 breaksw default: unset echo echo echo "Unforeseen version of Mac OS X. $GPC_CONTACT" echo exit 1 endsw
Try adding an extra item to the switch statement such as
case "14*": setenv GPC_DARWIN darwin14 setenv GPC_OSX_VERSION_NAME 10.10 setenv GPC_SDK /Developer/SDKs/MacOSX10.5.sdk setenv GPC_ALWAYS_USE_SDK 1 setenv GPC_PRE_TIGER 0 breaksw
This might get you a bit farther. I'm guessing the filename of the sdk might change, or you could set GPC_ALWAYS_USE_SDK to 0.
Peter
P.S. Assuming you are using this; www.microbizz.nl/gpc346u4-sources.dmg
Peter:
Thanks for the great suggstion!
P.S. Assuming you are using this; www.microbizz.nl/gpc346u4-sources.dmg
Yes.
More specifically, the immediate problem is the switch statement in build-gpc.command
As the OS version appears to be 14*, the 'default' [fail] option is used.
setenv GPC_OSX_VERSION `uname -r` switch ($GPC_OSX_VERSION) case "10*":
...
Try adding an extra item to the switch statement such as
Turns out that Apple moved the SDK. After much futzing around I located it. The additional case is now:
switch ($GPC_OSX_VERSION) case "14*": # Case added by Thomas Schneider 2015 Nov 16 TDS setenv GPC_DARWIN darwin10 setenv GPC_OSX_VERSION_NAME 10.10.5 setenv GPC_SDK /Volumes/SSD2/Macintosh_HD/Applications/Xcode.app/Contents/Developer/Platforms/MacOSX.platform/Developer/SDKs/MacOSX10.11.sdk setenv GPC_ALWAYS_USE_SDK 1 setenv GPC_PRE_TIGER 0 breaksw
That worked! But then it crashed at:
# work around bug in gcc script setenv SHELL /bin/bash # go the the directory the script is in cd $0:h
That fails with: cd build-gpc.command build-gpc.command: Not a directory.
So I simply blocked the 'cd' with '#". It's not clear why it was there.
Anyway, then the compile ran for quite a while, so we are making progress.
That ran for a while until:
make[3]: *** [libgcc/./_muldi3.o] Error 1 make[2]: *** [libgcc.a] Error 2 make[1]: *** [stage1_build] Error 2 make: *** [bootstrap] Error 2
Now I'm stuck again ...
Tom
Thomas D. Schneider, Ph.D. Senior Investigator National Institutes of Health National Cancer Institute Center for Cancer Research Gene Regulation and Chromosome Biology Laboratory Molecular Information Theory Group Frederick, Maryland 21702-1201 schneidt@mail.nih.gov https://schneider.ncifcrf.gov (current link) https://alum.mit.edu/www/toms (permanent link)
Peter:
setenv GPC_SDK /Volumes/SSD2/Macintosh_HD/Applications/Xcode.app/Contents/Developer/Platforms/MacOSX.platform/Developer/SDKs/MacOSX10.11.sdk
Note! This is the location on my system only; presumably nobody else has SSD2 ... Find the Xcode.app and then follow the path.
You can find it using: xcode-select -p find . -name "*SDK*" -print | grep MacOSX.platform
which can be done on one line:
cd `xcode-select -p`;find . -name "*SDK*" -print | grep MacOSX.platform
Then one has to select the directory that ends in '.sdk'.
Tom
Thomas D. Schneider, Ph.D. Senior Investigator National Institutes of Health National Cancer Institute Center for Cancer Research Gene Regulation and Chromosome Biology Laboratory Molecular Information Theory Group Frederick, Maryland 21702-1201 https://schneider.ncifcrf.gov (current link) https://alum.mit.edu/www/toms (permanent link)
On 16/11/15 18:43, Schneider wrote:
Peter:
Thanks for the great suggstion!
P.S. Assuming you are using this; www.microbizz.nl/gpc346u4-sources.dmg
Yes.
More specifically, the immediate problem is the switch statement in build-gpc.command
As the OS version appears to be 14*, the 'default' [fail] option is used.
setenv GPC_OSX_VERSION `uname -r` switch ($GPC_OSX_VERSION) case "10*":
...
Try adding an extra item to the switch statement such as
Turns out that Apple moved the SDK. After much futzing around I located it. The additional case is now:
switch ($GPC_OSX_VERSION) case "14*": # Case added by Thomas Schneider 2015 Nov 16 TDS setenv GPC_DARWIN darwin10 setenv GPC_OSX_VERSION_NAME 10.10.5 setenv GPC_SDK /Volumes/SSD2/Macintosh_HD/Applications/Xcode.app/Contents/Developer/Platforms/MacOSX.platform/Developer/SDKs/MacOSX10.11.sdk setenv GPC_ALWAYS_USE_SDK 1 setenv GPC_PRE_TIGER 0 breaksw
That worked! But then it crashed at:
# work around bug in gcc script setenv SHELL /bin/bash # go the the directory the script is in cd $0:h
That fails with: cd build-gpc.command build-gpc.command: Not a directory.
So I simply blocked the 'cd' with '#". It's not clear why it was there.
Anyway, then the compile ran for quite a while, so we are making progress.
That ran for a while until:
make[3]: *** [libgcc/./_muldi3.o] Error 1 make[2]: *** [libgcc.a] Error 2 make[1]: *** [stage1_build] Error 2 make: *** [bootstrap] Error 2
Now I'm stuck again ...
Tom
Thomas D. Schneider, Ph.D. Senior Investigator National Institutes of Health National Cancer Institute Center for Cancer Research Gene Regulation and Chromosome Biology Laboratory Molecular Information Theory Group Frederick, Maryland 21702-1201 schneidt@mail.nih.gov https://schneider.ncifcrf.gov (current link) https://alum.mit.edu/www/toms (permanent link)
I would (for now anyway) try building without the sdk, try setting setenv GPC_ALWAYS_USE_SDK 0
Peter
Peter:
I would (for now anyway) try building without the sdk, try setting setenv GPC_ALWAYS_USE_SDK 0
Unfortunately it still ends with:
/var/folders/vg/ts3vj22j2y3d5n3cmmhw6gt9dzb4gf/T//ccQXON5q.s:522:2: error: unknown directive .stabn 224,0,0,Lscope0 ^ /var/folders/vg/ts3vj22j2y3d5n3cmmhw6gt9dzb4gf/T//ccQXON5q.s:524:9: error: unsupported directive '.stabs' .stabs "",100,0,0,Letext ^ make[3]: *** [libgcc/./_muldi3.o] Error 1 make[2]: *** [libgcc.a] Error 2 make[1]: *** [stage1_build] Error 2 make: *** [bootstrap] Error 2
Tom
Thomas D. Schneider, Ph.D. Senior Investigator National Institutes of Health National Cancer Institute Center for Cancer Research Gene Regulation and Chromosome Biology Laboratory Molecular Information Theory Group Frederick, Maryland 21702-1201 https://schneider.ncifcrf.gov (current link) https://alum.mit.edu/www/toms (permanent link)
Schneider wrote:
Unfortunately it still ends with:
/var/folders/vg/ts3vj22j2y3d5n3cmmhw6gt9dzb4gf/T//ccQXON5q.s:522:2: error: unknown directive .stabn 224,0,0,Lscope0 ^ /var/folders/vg/ts3vj22j2y3d5n3cmmhw6gt9dzb4gf/T//ccQXON5q.s:524:9: error: unsupported directive '.stabs' .stabs "",100,0,0,Letext ^ make[3]: *** [libgcc/./_muldi3.o] Error 1 make[2]: *** [libgcc.a] Error 2 make[1]: *** [stage1_build] Error 2 make: *** [bootstrap] Error 2
Xcode 7 and later have replaced the system assembler with Clang, which doesn't support stabs debug information. If there are any places where you can specify options to be used while compiling libgcc.a, try adding the -gdwarf-2 flag.
You could also try downgrading to Xcode 6.4, which still shipped with the old assembler. You can get it from https://developer.apple.com/downloads/
Jonas
Jonas:
Xcode 7 and later have replaced the system assembler with Clang, which doesn't support stabs debug information. If there are any places where you can specify options to be used while compiling libgcc.a, try adding the -gdwarf-2 flag.
Thanks for the suggestion.
libgcc.a is not mentioned in build-gpc.command. I guessed:
# setenv GPC_MAKE_FLAGS 'CC="gcc -arch i386"' # TDS: setenv GPC_MAKE_FLAGS 'CC="gcc -arch i386 -gdwarf-2"'
but it failed as above. However the code does say: # build and install a dynamic libgpc
I was not clear where to inject your suggestion.
Maybe one could force the use of an open source gcc instead of Clang?
You could also try downgrading to Xcode 6.4, which still shipped with the old assembler. You can get it from https://developer.apple.com/downloads/
I wouldn't want to do that because it would put us in a bad position later when Xcode 6.4 is no longer around. Better to move forward if possible.
Tom
Thomas D. Schneider, Ph.D. Senior Investigator National Institutes of Health National Cancer Institute Center for Cancer Research Gene Regulation and Chromosome Biology Laboratory Molecular Information Theory Group Frederick, Maryland 21702-1201 https://schneider.ncifcrf.gov (current link) https://alum.mit.edu/www/toms (permanent link)
libgcc.a is not mentioned in build-gpc.command. I guessed:
# setenv GPC_MAKE_FLAGS 'CC="gcc -arch i386"' # TDS: setenv GPC_MAKE_FLAGS 'CC="gcc -arch i386 -gdwarf-2"'
but it failed as above.
Also just guessing;
try instead adding --with-dwarf2 to
setenv GPC_CONFIG_EXTRAS "--with-arch=pentium-m --with-tune=prescott"
ie
setenv GPC_CONFIG_EXTRAS "--with-arch=pentium-m --with-tune=prescott --with-dwarf2"
Peter:
Also just guessing; setenv GPC_CONFIG_EXTRAS "--with-arch=pentium-m --with-tune=prescott --with-dwarf2"
mkdir -p -- libgcc if [ -f stmp-dirs ]; then true; else touch stmp-dirs; fi ./xgcc -B./ -B/Developer/Pascal/gpc346u4/i386-apple-darwin10/bin/ -isystem /Developer/Pascal/gpc346u4/i386-apple-darwin10/include -isystem /Developer/Pascal/gpc346u4/i386-apple-darwin10/sys-include -L/Users/schneidt/now/gpc/current/build/gcc/../ld -O2 -DIN_GCC -W -Wall -Wwrite-strings -Wstrict-prototypes -Wmissing-prototypes -Wold-style-definition -isystem ./include -fPIC -g -DHAVE_GTHR_DEFAULT -DIN_LIBGCC2 -D__GCC_FLOAT_NOT_NEEDED -I. -I. -I../../gcc-3.4.6/gcc -I../../gcc-3.4.6/gcc/. -I../../gcc-3.4.6/gcc/../include -DL_muldi3 -c ../../gcc-3.4.6/gcc/libgcc2.c -o libgcc/./_muldi3.o ../../gcc-3.4.6/gcc/libgcc2.c:0: internal compiler error: in default_no_named_section, at varasm.c:4715 Please submit a full bug report, with preprocessed source if appropriate. See URL:http://gcc.gnu.org/bugs.html for instructions. make[3]: *** [libgcc/./_muldi3.o] Error 1 make[2]: *** [libgcc.a] Error 2 make[1]: *** [stage1_build] Error 2 make: *** [bootstrap] Error 2
Wow, that blew the compiler!
Tom
Thomas D. Schneider, Ph.D. Senior Investigator National Institutes of Health National Cancer Institute Center for Cancer Research Gene Regulation and Chromosome Biology Laboratory Molecular Information Theory Group Frederick, Maryland 21702-1201 https://schneider.ncifcrf.gov (current link) https://alum.mit.edu/www/toms (permanent link)
On 17/11/15 17:16, Schneider wrote:
Peter:
Also just guessing; setenv GPC_CONFIG_EXTRAS "--with-arch=pentium-m --with-tune=prescott --with-dwarf2"
mkdir -p -- libgcc if [ -f stmp-dirs ]; then true; else touch stmp-dirs; fi ./xgcc -B./ -B/Developer/Pascal/gpc346u4/i386-apple-darwin10/bin/ -isystem /Developer/Pascal/gpc346u4/i386-apple-darwin10/include -isystem /Developer/Pascal/gpc346u4/i386-apple-darwin10/sys-include -L/Users/schneidt/now/gpc/current/build/gcc/../ld -O2 -DIN_GCC -W -Wall -Wwrite-strings -Wstrict-prototypes -Wmissing-prototypes -Wold-style-definition -isystem ./include -fPIC -g -DHAVE_GTHR_DEFAULT -DIN_LIBGCC2 -D__GCC_FLOAT_NOT_NEEDED -I. -I. -I../../gcc-3.4.6/gcc -I../../gcc-3.4.6/gcc/. -I../../gcc-3.4.6/gcc/../include -DL_muldi3 -c ../../gcc-3.4.6/gcc/libgcc2.c -o libgcc/./_muldi3.o ../../gcc-3.4.6/gcc/libgcc2.c:0: internal compiler error: in default_no_named_section, at varasm.c:4715 Please submit a full bug report, with preprocessed source if appropriate. See URL:http://gcc.gnu.org/bugs.html for instructions. make[3]: *** [libgcc/./_muldi3.o] Error 1 make[2]: *** [libgcc.a] Error 2 make[1]: *** [stage1_build] Error 2 make: *** [bootstrap] Error 2
Wow, that blew the compiler!
Tom
Thomas D. Schneider, Ph.D. Senior Investigator National Institutes of Health National Cancer Institute Center for Cancer Research Gene Regulation and Chromosome Biology Laboratory Molecular Information Theory Group Frederick, Maryland 21702-1201 https://schneider.ncifcrf.gov (current link) https://alum.mit.edu/www/toms (permanent link)
Maybe use --without-stabs instead of --with-dwarf2 ?
And/Or set one of the following to -gdwarf-2
LIBGCC2_DEBUG_CFLAGS TARGET_LIBGCC2_CFLAGS
I recommend following Jonas' suggestion to try Xcode 6.4, at least temporarily, to see if the debug format is the only issue, or whether there are a raft of other problems waiting to reveal themselves.
Peter:
And/Or set one of the following to -gdwarf-2
LIBGCC2_DEBUG_CFLAGS TARGET_LIBGCC2_CFLAGS
I did:
setenv LIBGCC2_DEBUG_CFLAGS -gdwarf-2 setenv TARGET_LIBGCC2_CFLAGS -gdwarf-2
/var/folders/vg/ts3vj22j2y3d5n3cmmhw6gt9dzb4gf/T//ccnnRCza.s:524:9: error: unsupported directive '.stabs' .stabs "",100,0,0,Letext ^ make[3]: *** [libgcc/./_muldi3.o] Error 1 make[2]: *** [libgcc.a] Error 2 make[1]: *** [stage1_build] Error 2 make: *** [bootstrap] Error 2
Maybe use --without-stabs instead of --with-dwarf2 ?
setenv GPC_CONFIG_EXTRAS "--with-arch=pentium-m --with-tune=prescott --without-stabs"
/var/folders/vg/ts3vj22j2y3d5n3cmmhw6gt9dzb4gf/T//cc5bKMiz.s:524:9: error: unsupported directive '.stabs' .stabs "",100,0,0,Letext ^ make[3]: *** [libgcc/./_muldi3.o] Error 1 make[2]: *** [libgcc.a] Error 2 make[1]: *** [stage1_build] Error 2 make: *** [bootstrap] Error 2
I recommend following Jonas' suggestion to try Xcode 6.4, at least temporarily, to see if the debug format is the only issue, or whether there are a raft of other problems waiting to reveal themselves.
Ok, I found: http://developer.apple.com/devcenter/download.action?path=/Developer_Tools/X...
This went into /Applications and so I set this inside the build script:
setenv GPC_SDK /Applications/Xcode.app/Contents/Developer/Platforms/MacOSX.platform/Developer/SDKs setenv GPC_ALWAYS_USE_SDK 1
It crashed:
The directory that should contain system headers does not exist: /Applications/Xcode.app/Contents/Developer/Platforms/MacOSX.platform/Developer/SDKs//usr/include make[2]: *** [stmp-fixinc] Error 1 make[1]: *** [stage1_build] Error 2 make: *** [bootstrap] Error 2
This is a total mess.
Tom
Thomas D. Schneider, Ph.D. Senior Investigator National Institutes of Health National Cancer Institute Center for Cancer Research Gene Regulation and Chromosome Biology Laboratory Molecular Information Theory Group Frederick, Maryland 21702-1201 schneidt@mail.nih.gov https://schneider.ncifcrf.gov (current link) https://alum.mit.edu/www/toms (permanent link)
Schneider wrote:
Ok, I found: http://developer.apple.com/devcenter/download.action?path=/Developer_Tools/X...
This went into /Applications and so I set this inside the build script:
setenv GPC_SDK /Applications/Xcode.app/Contents/Developer/Platforms/MacOSX.platform/Developer/SDKs
The actual SDK's are the MacOSX10.9.sdk and MacOSX10.10.sdk subdirectories of that directory.
setenv GPC_ALWAYS_USE_SDK 1
It crashed:
The directory that should contain system headers does not exist: /Applications/Xcode.app/Contents/Developer/Platforms/MacOSX.platform/Developer/SDKs//usr/include make[2]: *** [stmp-fixinc] Error 1 make[1]: *** [stage1_build] Error 2 make: *** [bootstrap] Error 2
It didn't crash. It just says that it cannot find any headers in the "SDK" directory you specified, which is correct.
Jonas
Jonas:
The actual SDK's are the MacOSX10.9.sdk and MacOSX10.10.sdk subdirectories of that directory. It didn't crash. It just says that it cannot find any headers in the "SDK" directory you specified, which is correct.
DUH. I knew that! I just cut/pasted the wrong thing! Thanks!
However:
/var/folders/vg/ts3vj22j2y3d5n3cmmhw6gt9dzb4gf/T//ccidJkLD.s:524:9: error: unsupported directive '.stabs' .stabs "",100,0,0,Letext ^ make[3]: *** [libgcc/./_muldi3.o] Error 1 make[2]: *** [libgcc.a] Error 2 make[1]: *** [stage1_build] Error 2 make: *** [bootstrap] Error 2
By the way, I'm just rerunning build-gpc.command each time. So I effectively assume it cleans out previous results. Could that be wrong?
Tom
Thomas D. Schneider, Ph.D. Senior Investigator National Institutes of Health National Cancer Institute Center for Cancer Research Gene Regulation and Chromosome Biology Laboratory Molecular Information Theory Group Frederick, Maryland 21702-1201 schneidt@mail.nih.gov https://schneider.ncifcrf.gov (current link) https://alum.mit.edu/www/toms (permanent link)
Schneider wrote:
However:
/var/folders/vg/ts3vj22j2y3d5n3cmmhw6gt9dzb4gf/T//ccidJkLD.s:524:9: error: unsupported directive '.stabs' .stabs "",100,0,0,Letext ^ make[3]: *** [libgcc/./_muldi3.o] Error 1 make[2]: *** [libgcc.a] Error 2 make[1]: *** [stage1_build] Error 2 make: *** [bootstrap] Error 2
What does as -version
show? If it doesn't show Apple Inc version cctools-870, GNU assembler version 1.38
(you have to press ctrl-c to quit it) try running sudo xcode-select -s /Applications/Xcode.app
Afterwards, try the above command again to verify it shows the expected result. Maybe the link between the dummy /usr/bin/as and clang still somehow exists from when you had Xcode 7.
By the way, I'm just rerunning build-gpc.command each time. So I effectively assume it cleans out previous results. Could that be wrong?
I have no idea.
Jonas
Jonas:
What does as -version
show? If it doesn't show Apple Inc version cctools-870, GNU assembler version 1.38> (you have to press ctrl-c to quit it) try running sudo xcode-select -s /Applications/Xcode.app
Afterwards, try the above command again to verify it shows the expected result. Maybe the link between the dummy /usr/bin/as and clang still somehow exists from when you had Xcode 7.
1% as -version clang: error: unknown argument: '-version'2% as --version Apple LLVM version 7.0.0 (clang-700.1.76) Target: x86_64-apple-darwin14.5.0 Thread model: posix
Ok, wrong version, thanks.
3% sudo xcode-select -s /Applications/Xcode.app Password: 4% as --version FATAL:/Applications/Xcode.app/Contents/Developer/Toolchains/XcodeDefault.xctoolchai n/usr/bin/../libexec/as/x86_64/as: I don't understand '-' flag! 5% as -version Apple Inc version cctools-870, GNU assembler version 1.38 ^CInterrupted by signal 2
Ok!
6% build-gpc.command ... configure: error: installation or configuration problem: C compiler cannot create e xecutables. make[2]: *** [p/rts/Makefile] Error 1 make[1]: *** [stage1_build] Error 2 make: *** [bootstrap] Error 2
Hmm.
Tom
Thomas D. Schneider, Ph.D. Senior Investigator National Institutes of Health National Cancer Institute Center for Cancer Research Gene Regulation and Chromosome Biology Laboratory Molecular Information Theory Group Frederick, Maryland 21702-1201 schneidt@mail.nih.gov https://schneider.ncifcrf.gov (current link) https://alum.mit.edu/www/toms (permanent link)
Schneider wrote on Wed, 18 Nov 2015:
6% build-gpc.command ... configure: error: installation or configuration problem: C compiler cannot create e xecutables. make[2]: *** [p/rts/Makefile] Error 1 make[1]: *** [stage1_build] Error 2 make: *** [bootstrap] Error 2
Hmm.
Check what the associated config.log says.
Jonas
Jonas:
Check what the associated config.log says.
I couldn't look at that at first because files were owned by root and other things ... I fixed those (to be owned by me) but that didn't change the results.
% more config.log This file contains any messages produced by compilers while running configure, to aid debugging if configure makes a mistake.
configure:581: checking host system type configure:602: checking target system type configure:620: checking build system type configure:675: checking for a BSD compatible install configure:2866: checking for i386-apple-darwin10-ar configure:2899: checking for ar configure:2938: checking for i386-apple-darwin10-as configure:2971: checking for as configure:3010: checking for i386-apple-darwin10-dlltool configure:3043: checking for dlltool configure:3082: checking for i386-apple-darwin10-ld configure:3154: checking for i386-apple-darwin10-nm configure:3187: checking for nm configure:3226: checking for i386-apple-darwin10-ranlib configure:3259: checking for ranlib configure:3298: checking for i386-apple-darwin10-windres configure:3331: checking for windres configure:3370: checking for i386-apple-darwin10-objcopy configure:3403: checking for objcopy configure:3442: checking for i386-apple-darwin10-objdump configure:3475: checking for objdump configure:3524: checking for i386-apple-darwin10-ar configure:3557: checking for ar configure:3596: checking for i386-apple-darwin10-as configure:3629: checking for as configure:3668: checking for i386-apple-darwin10-dlltool configure:3701: checking for dlltool configure:3740: checking for i386-apple-darwin10-ld configure:3773: checking for ld configure:3812: checking for i386-apple-darwin10-nm configure:3845: checking for nm configure:3884: checking for i386-apple-darwin10-ranlib configure:3917: checking for ranlib configure:3956: checking for i386-apple-darwin10-windres configure:3989: checking for windres configure:4056: checking whether to enable maintainer-specific portions of Makefiles
configure: error: installation or configuration problem: C compiler cannot create executables. make[2]: *** [p/rts/Makefile] Error 1 make[1]: *** [stage1_build] Error 2 make: *** [bootstrap] Error 2 egg 55%
The error message comes from
gpc346u4-sources/gcc-3.4.6/intl/configure: { echo "configure: error: installation or configuration problem: C compiler cannot create executables." 1>&2; exit 1; }
So gcc can't write for some reason ...?
Tom
Thomas D. Schneider, Ph.D. Senior Investigator National Institutes of Health National Cancer Institute Center for Cancer Research Gene Regulation and Chromosome Biology Laboratory Molecular Information Theory Group Frederick, Maryland 21702-1201 https://schneider.ncifcrf.gov (current link) https://alum.mit.edu/www/toms (permanent link)
Schneider wrote on Wed, 18 Nov 2015:
configure:4056: checking whether to enable maintainer-specific portions of Makefiles
configure: error: installation or configuration problem: C compiler cannot create executables. make[2]: *** [p/rts/Makefile] Error 1 make[1]: *** [stage1_build] Error 2 make: *** [bootstrap] Error 2 egg 55%
There doesn't seem to be any output for the failing test or command here.
The error message comes from
gpc346u4-sources/gcc-3.4.6/intl/configure: { echo "configure: error: installation or configuration problem: C compiler cannot create executables." 1>&2; exit 1; }
So gcc can't write for some reason ...?
That message can have many causes, e.g. the compiler segfaulting or the linker not finding a system library. Without the actual output, there's not really anything to go on.
As an aside, "gcc" is an alias for Clang on OS X nowadays (I think since Xcode 5), but shouldn't be relevant here (since you apparently got past the above stage with Xcode 7).
Jonas
Jonas:
Schneider wrote on Wed, 18 Nov 2015:
configure:4056: checking whether to enable maintainer-specific portions of Makefiles
configure: error: installation or configuration problem: C compiler cannot create executables. make[2]: *** [p/rts/Makefile] Error 1 make[1]: *** [stage1_build] Error 2 make: *** [bootstrap] Error 2 egg 55%
There doesn't seem to be any output for the failing test or command here.
The error message comes from
gpc346u4-sources/gcc-3.4.6/intl/configure: { echo "configure: error: installation or configuration problem: C compiler cannot create executables." 1>&2; exit 1; }
So gcc can't write for some reason ...?
That message can have many causes, e.g. the compiler segfaulting or the linker not finding a system library. Without the actual output, there's not really anything to go on.
As an aside, "gcc" is an alias for Clang on OS X nowadays (I think since Xcode 5), but shouldn't be relevant here (since you apparently got past the above stage with Xcode 7).
I've switched back to Xcode 7 (because fink objected when I tried to install a program).
% build-gpc.command ... chmod 755 ../fixinc.sh The directory that should contain system headers does not exist: /Applications/Xcode.app/Contents/Developer/Platforms/MacOSX.platform/Developer/SDKs/MacOSX10.10.sdk///usr/include make[2]: *** [stmp-fixinc] Error 1 make[1]: *** [stage1_build] Error 2 make: *** [bootstrap] Error 2
Hmm.
Tom
Thomas D. Schneider, Ph.D. Senior Investigator National Institutes of Health National Cancer Institute Center for Cancer Research Gene Regulation and Chromosome Biology Laboratory Molecular Information Theory Group Frederick, Maryland 21702-1201 https://schneider.ncifcrf.gov (current link) https://alum.mit.edu/www/toms (permanent link)
On 18 Nov 2015, at 14:21, Schneider schneidt@mail.nih.gov wrote:
The error message comes from
gpc346u4-sources/gcc-3.4.6/intl/configure: { echo "configure: error: installation or configuration problem: C compiler cannot create executables." 1>&2; exit 1; }
So gcc can't write for some reason ...?
Then open the configure script, find that line and see what it tries to do before bailing out with that message. Maybe you can repeat the offending command on the command line see its output.
Bastiaan.
Bastiaan:
On 18 Nov 2015, at 14:21, Schneider schneidt@mail.nih.gov wrote:
The error message comes from
gpc346u4-sources/gcc-3.4.6/intl/configure: { echo "configure: error: installation or configuration problem: C compiler cannot create executables." 1>&2; exit 1; }
So gcc can't write for some reason ...?
Then open the configure script, find that line and see what it tries to do before bailing out with that message. Maybe you can repeat the offending command on the command line see its output.
Having switched back to Xcode 7 that error is gone (see previous email).
Tom
Thomas D. Schneider, Ph.D. Senior Investigator National Institutes of Health National Cancer Institute Center for Cancer Research Gene Regulation and Chromosome Biology Laboratory Molecular Information Theory Group Frederick, Maryland 21702-1201 https://schneider.ncifcrf.gov (current link) https://alum.mit.edu/www/toms (permanent link)
Hi Tom,
Could you post the output of
gcc -v
and check whether you can build a basic helloworld.c on this machine?
Peter
Peter::
Could you post the output of gcc -v and check whether you can build a basic helloworld.c on this machine?
===== % uname -mprsv === Darwin 14.5.0 Darwin Kernel Version 14.5.0: Tue Sep 1 21:23:09 PDT 2015; root:xnu-2782.50.1~1/RELEASE_X86_64 x86_64 i386 ===== % which gcc ====== /usr/bin/gcc ===== % gcc -v ========= Configured with: --prefix=/Applications/Xcode.app/Contents/Developer/usr --with-gxx-include-dir=/usr/include/c++/4.2.1 Apple LLVM version 7.0.0 (clang-700.1.76) Target: x86_64-apple-darwin14.5.0 Thread model: posix ===== % hello.c ======== //C hello world example #include <stdio.h>
int main() { printf("Hello happy world\n"); return 0; } ===== % gcc hello.c ==== ===== % run a.out: ===== Hello happy world ======================
Thanks!
Tom
Thomas D. Schneider, Ph.D. Senior Investigator National Institutes of Health National Cancer Institute Center for Cancer Research Gene Regulation and Chromosome Biology Laboratory Molecular Information Theory Group Frederick, Maryland 21702-1201 https://schneider.ncifcrf.gov (current link) https://alum.mit.edu/www/toms (permanent link)
Hi Tom,
I don't know whether anyone has ever built GPC with Clang before.
I was going to suggest trying to built stock GCC 3.4.6 first, because the problems building GPC are generally always to do with the GCC steps. However, my very first Google hit searhing for 'sdk xcode gcc' turned up this; https://github.com/devernay/xcodelegacy
Comments there include "Many components were removed in recent versions of Xcode..."
and most important maybe "GCC 4.0, GCC 4.2 and LLVM GCC 4.2 cannot compile for OS X 10.10 or newer."
It looks as though support for vintage GCC compilers is just not there in 10.10. Probably only options are to downgrade OS X, or use a Linux box or Windows/MinGW.
Peter
This is probably not an issue for you yet (on Yosemite) but under OS X 10.11, El Capitan, writing to /usr is no longer allowed. At least you’ll know this can be an issue if you get problems after upgrading to El Capitan.
Good luck, Bastiaan.
Bastiaan Veelo wrote on Wed, 18 Nov 2015:
This is probably not an issue for you yet (on Yosemite) but under OS X 10.11, El Capitan, writing to /usr is no longer allowed. At least you’ll know this can be an issue if you get problems after upgrading to El Capitan.
Writing to /usr, apart from everything under /usr/local, is something that was never supported under OS X (although it was possible with root permissions), and in fact should never be done on any Unix platform (unless you want to hack the system for one reason or another, which could be perfectly legitimate).
All of those files and directories "belong" to the operating system and could be changed/overwritten with any OS upgrade.
Jonas
On 18 Nov 2015, at 14:05, Jonas Maebe jonas.maebe@elis.ugent.be wrote:
Bastiaan Veelo wrote on Wed, 18 Nov 2015:
This is probably not an issue for you yet (on Yosemite) but under OS X 10.11, El Capitan, writing to /usr is no longer allowed. At least you’ll know this can be an issue if you get problems after upgrading to El Capitan.
Writing to /usr, apart from everything under /usr/local, is something that was never supported under OS X (although it was possible with root permissions), and in fact should never be done on any Unix platform (unless you want to hack the system for one reason or another, which could be perfectly legitimate).
All of those files and directories "belong" to the operating system and could be changed/overwritten with any OS upgrade.
Indeed. Still, most TeX-based utilities relied on a symbolic link to the TeX installation to be present in /bin. https://tug.org/mactex/UpdatingForElCapitan.pdf
Bastiaan.
-
Bastiaan Veelo -
Jonas Maebe -
Ken Linder -
Peter -
Prof Abimbola Olowofoyeku -
Schneider -
Schneider, Thomas (NIH/NCI) [E]