glib2 @2.78.4 -x11 -xauth +no_x11 +quartz fails to build on OSX 10.4 Tiger

[pkoshevoy (Pavel Koshevoy)]

$ sudo port clean glib2
--->  Cleaning glib2

$ sudo port upgrade glib2 -x11 -xauth +no_x11 +quartz
--->  Computing dependencies for glib2
--->  Fetching archive for glib2
--->  Attempting to fetch glib2-2.78.4_0+quartz.darwin_8.ppc.tbz2 from http://packages.macports.org/glib2
--->  Attempting to fetch glib2-2.78.4_0+quartz.darwin_8.ppc.tbz2 from http://mirror.fcix.net/macports/packages/glib2
--->  Attempting to fetch glib2-2.78.4_0+quartz.darwin_8.ppc.tbz2 from http://ywg.ca.packages.macports.org/mirror/macports/packages/glib2
--->  Fetching distfiles for glib2
--->  Verifying checksums for glib2
--->  Extracting glib2
--->  Applying patches to glib2
--->  Configuring glib2
Warning: Configuration logfiles contain indications of -Wimplicit-function-declaration; check that features were not accidentally disabled:
  statx: found in build/meson-logs/meson-log.txt
  __sync_swap: found in build/meson-logs/meson-log.txt
--->  Building glib2
Error: Failed to build glib2: command execution failed
Error: See /opt/local/var/macports/logs/_opt_local_var_macports_sources_rsync.macports.org_macports_release_tarballs_ports_devel_glib2/glib2/main.log for details.
Error: Follow https://guide.macports.org/#project.tickets if you believe there is a bug.

[pkoshevoy (Pavel Koshevoy)]

/opt/local/var/macports/logs/_opt_local_var_macports_sources_rsync.macports.org_macports_release_tarballs_ports_devel_glib2/glib2/main.log

[kencu (Ken)]

 ../glib-2.78.4/gio/gosxcontenttype.m:357:30: error: 'kUTTypeApplication' undeclared (first use in this function); did you mean 'GApplication'?

you'll not get very far at all trying to build -x11 +quartz on Tiger.

the quartz variants are usually built to support current MacOS systems, and they require newer Apple feature sets, SDKs, and technologies, for the most part.

I suggest you stick to +x11 and you'll get much further.

[pkoshevoy (Pavel Koshevoy)]

I have no use for x11 support in any of the ports, I always explicitly disable it (-x11 -xauth +no_x11 +quartz). glib2 used to build on tiger with +quartz, but now it's doesn't. Unfortunately, glib2 is a dependency for some of the ports I actually need (libass ... or one of its dependencies). I would rather have an older version of glib2 that builds on tiger, than the latest version that breaks tiger compatibility and therefore prevents upgrade of the ports that I actually need.

I suspect the patch to fix that build failure might be a trivial 1-liner -- just add

#include <LaunchServices/UTCoreTypes.h>

to gio/gosxcontenttype.m

[pkoshevoy (Pavel Koshevoy)]

I added #include <LaunchServices/UTCoreTypes.h> to gio/gosxcontenttype.m and ran (under sudo su -)

cd "/opt/local/var/macports/build/_opt_local_var_macports_sources_rsync.macports.org_macports_release_tarballs_ports_devel_glib2/glib2/work/build" && /opt/local/bin/ninja -j1 --verbose -v

that built for a while and then failed on glib/gfileutils.c implicit declaration of strnlen I replaced it with memchr and tried again... this time if failed on glib/gmem.c implicit declaration of posix_memalign

macmini-g4-125:/opt/local/var/macports/build/_opt_local_var_macports_sources_rsync.macports.org_macports_release_tarballs_ports_devel_glib2/glib2/work/build root# cd "/opt/local/var/macports/build/_opt_local_var_macports_sources_rsync.macports.org_macports_release_tarballs_ports_devel_glib2/glib2/work/build" && /opt/local/bin/ninja -j1 --verbose -v 
[1/532] /opt/local/bin/gcc-mp-7 -Iglib/libglib-2.0.0.dylib.p -Iglib -I../glib-2.78.4/glib -I. -I../glib-2.78.4 -I/opt/local/include -fvisibility=hidden -fdiagnostics-color=always -std=gnu99 -O2 -g -D_GNU_SOURCE -fno-strict-aliasing -DG_DISABLE_CAST_CHECKS -Wduplicated-branches -Wimplicit-fallthrough -Wmisleading-indentation -Wmissing-field-initializers -Wnonnull -Wunused -Wno-unused-parameter -Wno-pedantic -Wno-format-zero-length -Wno-variadic-macros -Werror=format=2 -Werror=init-self -Werror=missing-include-dirs -Werror=pointer-arith -Werror=unused-result -Wstrict-prototypes -Wno-bad-function-cast -Wno-declaration-after-statement -Werror=implicit-function-declaration -Werror=missing-prototypes -Werror=pointer-sign -pipe -Os -fstrict-aliasing -Wno-deprecated-declarations -D__DARWIN_NON_CANCELABLE=1 -arch ppc '-DG_LOG_DOMAIN="GLib"' -DGLIB_COMPILATION -MD -MQ glib/libglib-2.0.0.dylib.p/gmem.c.o -MF glib/libglib-2.0.0.dylib.p/gmem.c.o.d -o glib/libglib-2.0.0.dylib.p/gmem.c.o -c ../glib-2.78.4/glib/gmem.c
FAILED: glib/libglib-2.0.0.dylib.p/gmem.c.o 
/opt/local/bin/gcc-mp-7 -Iglib/libglib-2.0.0.dylib.p -Iglib -I../glib-2.78.4/glib -I. -I../glib-2.78.4 -I/opt/local/include -fvisibility=hidden -fdiagnostics-color=always -std=gnu99 -O2 -g -D_GNU_SOURCE -fno-strict-aliasing -DG_DISABLE_CAST_CHECKS -Wduplicated-branches -Wimplicit-fallthrough -Wmisleading-indentation -Wmissing-field-initializers -Wnonnull -Wunused -Wno-unused-parameter -Wno-pedantic -Wno-format-zero-length -Wno-variadic-macros -Werror=format=2 -Werror=init-self -Werror=missing-include-dirs -Werror=pointer-arith -Werror=unused-result -Wstrict-prototypes -Wno-bad-function-cast -Wno-declaration-after-statement -Werror=implicit-function-declaration -Werror=missing-prototypes -Werror=pointer-sign -pipe -Os -fstrict-aliasing -Wno-deprecated-declarations -D__DARWIN_NON_CANCELABLE=1 -arch ppc '-DG_LOG_DOMAIN="GLib"' -DGLIB_COMPILATION -MD -MQ glib/libglib-2.0.0.dylib.p/gmem.c.o -MF glib/libglib-2.0.0.dylib.p/gmem.c.o.d -o glib/libglib-2.0.0.dylib.p/gmem.c.o -c ../glib-2.78.4/glib/gmem.c
In file included from ../glib-2.78.4/glib/gmem.c:31:0:
./config.h:358:0: warning: "__GCC_HAVE_SYNC_COMPARE_AND_SWAP_4" redefined
 #define __GCC_HAVE_SYNC_COMPARE_AND_SWAP_4
 
<built-in>: note: this is the location of the previous definition
../glib-2.78.4/glib/gmem.c: In function 'g_aligned_alloc':
../glib-2.78.4/glib/gmem.c:672:11: error: implicit declaration of function 'posix_memalign' [-Werror=implicit-function-declaration]
   errno = posix_memalign (&res, alignment, real_size);
           ^~~~~~~~~~~~~~
cc1: some warnings being treated as errors
ninja: build stopped: subcommand failed.

[pkoshevoy (Pavel Koshevoy)]

The changes I've tried so far:

# (cd /opt/local/var/macports/build/_opt_local_var_macports_sources_rsync.macports.org_macports_release_tarballs_ports_devel_glib2/glib2/work/glib-2.78.4/gio; diff gosxcontenttype.m~ gosxcontenttype.m)
26a27
> #include <LaunchServices/UTCoreTypes.h>

# (cd /opt/local/var/macports/build/_opt_local_var_macports_sources_rsync.macports.org_macports_release_tarballs_ports_devel_glib2/glib2/work/glib-2.78.4/glib; diff gfileutils.c~ gfileutils.c)3003c3003
<   g_assert (dir == NULL || strnlen (dir, buffer_size) < buffer_size);
---
>   g_assert (dir == NULL || memchr (dir, '\0', buffer_size) != NULL);

[kencu (Ken)]

I just rebuilt glib2 2.78.4_0 (normal default x11 version) and it builds and installs on Tiger PPC for me without any trouble.

As x11 is the only windowing system that has any chance of being useful on Tiger, that is what everyone uses. So once again, I recommend you use that instead of the largely useless quartz variant on Tiger.

If you insist on staying on quartz on Tiger despite this, and if you get a patch sorted out that works and passes some kind of testing, we will look at it.

[pkoshevoy (Pavel Koshevoy)]

It appears that strnlen is provided in /opt/local/include/LegacySupport/string.h, and posix_memalign is provided in /opt/local/include/LegacySupport/stdlib.h So, I don't think I should be running into those build errors. Any idea why I'm seeing these errors when I just rerun the ninja command?

cd "/opt/local/var/macports/build/_opt_local_var_macports_sources_rsync.macports.org_macports_release_tarballs_ports_devel_glib2/glib2/work/build" && /opt/local/bin/ninja -j1 --verbose -v

[pkoshevoy (Pavel Koshevoy)]

I ran it as root ... when macports builds stuff -- does it set anything in env like CPATH, or LIBPATH?

[ryandesign (Ryan Carsten Schmidt)]

Replying to pkoshevoy:

when macports builds stuff -- does it set anything in env like CPATH, or LIBPATH?

Yes, MacPorts sets the CPATH environment variable to the value of the MacPorts compiler.cpath option and its sets the LIBRARY_PATH environment variable to the value of the MacPorts compiler.library_path option. There are numerous other environment variables that MacPorts sets in various ways. You can see the full list by examining a port's main.log file or using the debug mode switch (-d).