GNOME Bugzilla – Bug 743606
INFO: Test Summary report glib-2.43.3 for AIX
Last modified: 2018-05-24 17:24:18 UTC
glib-2.43.3 build on AIX 5.3 TL7 using gcc 4.7.3 Some compiler messages, etc.. make check returns: ============================================================================ Testsuite summary for glib 2.43.3 ============================================================================ # TOTAL: 3334 # PASS: 3313 # SKIP: 13 # XFAIL: 0 # FAIL: 0 # XPASS: 0 # ERROR: 8 ============================================================================ See glib/tests/test-suite.log Please report to http://bugzilla.gnome.org/enter_bug.cgi?product=glib ============================================================================
Created attachment 295594 [details] make check test-suite log
hi Michael, It will be very difficult for us to fix these problems without access to an AIX box or a more detailed description of the specific failures. Patches are welcome -- particularly if they fix general problems with (for example) POSIX compliance.
Looked at the first reported error: On AIX, as root, you can chdir even when the x bit is clear, but not as a normal user Test code: ret = g_mkdir ("mkdir-test", 0666); g_assert (ret == 0); ret = g_stat ("mkdir-test", &buf); g_assert (ret == 0); g_assert (S_ISDIR (buf.st_mode)); cwd = g_get_current_dir (); path = g_build_filename (cwd, "mkdir-test", NULL); g_free (cwd); ret = g_chdir (path); g_assert (errno == EACCES); g_assert (ret == -1); On line example: root@x064:[/]mkdir xxx root@x064:[/]chmod 666 xxx root@x064:[/]cd xxx root@x064:[/xxx]su michael root@x064:[]id uid=199(michael) gid=1(staff) groups=1(staff) root@x064:[]pwd ksh: : Cannot access parent directories. root@x064:[]cd / root@x064:[/]cd xxx ksh: xxx: Permission denied. root@x064:[/] root@x064:[/xxx]id uid=0(root) gid=0(system) groups=0(system),2(bin),3(sys),7(security),8(cron),10(audit),11(lp) root@x064:[/xxx] If I understand the test code, even though EACCES is not the errno - the return code is still -1. While (perhaps) this test should not be run as root - for these reasons - it might be nice to now what the errno is (hopefully not zero). Question: rather than just report g_assert() failure, is there an accepted way to provide additional output? And/or how can I run only this one group (fileutils) of tests? And then I will come up with what I hope will be accepted as a good patch.
Second failure in fileutils.c - tap-test calls ./fileutils -k --tap and fails with an IOT ../../tap-test[6]: 745530 IOT/Abort trap(coredump) # GLib:ERROR:fileutils.c:853:test_stdio_wrappers: assertion failed: (ret == -1) ERROR: fileutils - missing test plan ERROR: fileutils - exited with status 134 (terminated by signal 6?) root@x064:[/data/prj/gnome/glib/glib-2.43.3/glib/tests] AIX 5.3 does not have a tap interface. Not sure where/how this is handled by fileutils.c - some help and I will look for a patch. p.s. since AIX 6.1 TL7 there is a tap device available, but not configured by default.
(In reply to comment #4) > AIX 5.3 does not have a tap interface. Not that kind of tap. TAP == Test Anything Protocol. That's just part of the test running stuff. The actual bug here is: > # GLib:ERROR:fileutils.c:853:test_stdio_wrappers: assertion failed: (ret == -1) which suggests some call is succeeding under AIX that GLib expects to fail
Actually, except for the EACCES asset failure, it looks as if the ERRORs reported are 4x ../../tap-test failures plus 4x "missing test plan" which seems to be generated whenever any test fails. So, I "NEEDINFO" on how to react to the tap-test. root@x065:[/data/prj/gnome/glib/glib-2.43.3/glib/tests]rm gdate*log root@x065:[/data/prj/gnome/glib/glib-2.43.3/glib/tests]make gdatetime.log PASS: gdatetime 1 /GDateTime/invalid PASS: gdatetime 2 /GDateTime/add_days PASS: gdatetime 3 /GDateTime/add_full PASS: gdatetime 4 /GDateTime/add_hours PASS: gdatetime 5 /GDateTime/add_minutes PASS: gdatetime 6 /GDateTime/add_months PASS: gdatetime 7 /GDateTime/add_seconds PASS: gdatetime 8 /GDateTime/add_weeks PASS: gdatetime 9 /GDateTime/add_years PASS: gdatetime 10 /GDateTime/compare PASS: gdatetime 11 /GDateTime/diff ERROR: gdatetime - missing test plan ERROR: gdatetime - exited with status 134 (terminated by signal 6?) cat gdatetime.log # random seed: R02Sc986f203644b012acddfc80620bd8589 # Start of GDateTime tests # Bug Reference: http://bugzilla.gnome.org/702674 ok 1 /GDateTime/invalid PASS: gdatetime 1 /GDateTime/invalid ok 2 /GDateTime/add_days PASS: gdatetime 2 /GDateTime/add_days ok 3 /GDateTime/add_full PASS: gdatetime 3 /GDateTime/add_full ok 4 /GDateTime/add_hours PASS: gdatetime 4 /GDateTime/add_hours ok 5 /GDateTime/add_minutes PASS: gdatetime 5 /GDateTime/add_minutes ok 6 /GDateTime/add_months PASS: gdatetime 6 /GDateTime/add_months ok 7 /GDateTime/add_seconds PASS: gdatetime 7 /GDateTime/add_seconds ok 8 /GDateTime/add_weeks PASS: gdatetime 8 /GDateTime/add_weeks ok 9 /GDateTime/add_years PASS: gdatetime 9 /GDateTime/add_years ok 10 /GDateTime/compare ** GLib:ERROR:gdatetime.c:208:test_GDateTime_equal: assertion failed (g_date_time_get_utc_offset (dt1) / G_USEC_PER_SEC == (-3 * 3600)): (0 == -10800) PASS: gdatetime 10 /GDateTime/compare ok 11 /GDateTime/diff PASS: gdatetime 11 /GDateTime/diff ../../tap-test[5]: 225520 IOT/Abort trap(coredump) # GLib:ERROR:gdatetime.c:208:test_GDateTime_equal: assertion failed (g_date_time_get_utc_offset (dt1) / G_USEC_PER_SEC == (-3 * 3600)): (0 == -10800) ERROR: gdatetime - missing test plan ERROR: gdatetime - exited with status 134 (terminated by signal 6?) +++++++++++++ ADJUST tap-test ++++++++++++++++ FAKE ../../tap-test #! /bin/sh # run a GTest in tap mode. The test binary is passed as $1 exit 0 $1 -k --tap ++++++++++ Very Different results - but still get missing test plan ++++ root@x065:[/data/prj/gnome/glib/glib-2.43.3/glib/tests]rm gdate*log root@x065:[/data/prj/gnome/glib/glib-2.43.3/glib/tests]make gdatetime.log ERROR: gdatetime - missing test plan root@x065:[/data/prj/gnome/glib/glib-2.43.3/glib/tests]cat gdatetime.log ERROR: gdatetime - missing test plan root@x065:[/data/prj/gnome/glib/glib-2.43.3/glib/tests] Note: this is repeatable (i.e., with tap-test restored to original full/verbose log file is result.
OK - tap is test any protocol ... :) tried running fileutils as 'michael' root@x065:[/data/prj/gnome/glib/glib-2.43.3/glib/tests]chown -R michael . root@x065:[/data/prj/gnome/glib/glib-2.43.3/glib/tests]rm file*log root@x065:[/data/prj/gnome/glib/glib-2.43.3/glib/tests]su michael root@x065:[/data/prj/gnome/glib/glib-2.43.3/glib/tests]make fileutils.log PASS: fileutils 1 /fileutils/build-path PASS: fileutils 2 /fileutils/build-pathv PASS: fileutils 3 /fileutils/build-filename PASS: fileutils 4 /fileutils/build-filenamev PASS: fileutils 5 /fileutils/mkdir-with-parents PASS: fileutils 6 /fileutils/format-size-for-display PASS: fileutils 7 /fileutils/errors PASS: fileutils 8 /fileutils/basename PASS: fileutils 9 /fileutils/dir-make-tmp PASS: fileutils 10 /fileutils/file-open-tmp PASS: fileutils 11 /fileutils/mkstemp PASS: fileutils 12 /fileutils/mkdtemp PASS: fileutils 13 /fileutils/set-contents PASS: fileutils 14 /fileutils/read-link ERROR: fileutils - missing test plan ERROR: fileutils - exited with status 134 (terminated by signal 6?) root@x065:[/data/prj/gnome/glib/glib-2.43.3/glib/tests]cat fileutils.log # random seed: R02S4299bde1078a6d379ded1cfd4f1752cc # Start of fileutils tests ok 1 /fileutils/build-path PASS: fileutils 1 /fileutils/build-path ok 2 /fileutils/build-pathv PASS: fileutils 2 /fileutils/build-pathv ok 3 /fileutils/build-filename PASS: fileutils 3 /fileutils/build-filename ok 4 /fileutils/build-filenamev PASS: fileutils 4 /fileutils/build-filenamev ok 5 /fileutils/mkdir-with-parents PASS: fileutils 5 /fileutils/mkdir-with-parents ok 6 /fileutils/format-size-for-display PASS: fileutils 6 /fileutils/format-size-for-display ok 7 /fileutils/errors PASS: fileutils 7 /fileutils/errors ok 8 /fileutils/basename PASS: fileutils 8 /fileutils/basename ok 9 /fileutils/dir-make-tmp PASS: fileutils 9 /fileutils/dir-make-tmp ok 10 /fileutils/file-open-tmp PASS: fileutils 10 /fileutils/file-open-tmp ok 11 /fileutils/mkstemp PASS: fileutils 11 /fileutils/mkstemp ok 12 /fileutils/mkdtemp PASS: fileutils 12 /fileutils/mkdtemp ok 13 /fileutils/set-contents PASS: fileutils 13 /fileutils/set-contents ** GLib:ERROR:fileutils.c:853:test_stdio_wrappers: assertion failed: (ret == -1) ok 14 /fileutils/read-link PASS: fileutils 14 /fileutils/read-link ../../tap-test[5]: 737428 IOT/Abort trap(coredump) # GLib:ERROR:fileutils.c:853:test_stdio_wrappers: assertion failed: (ret == -1) ERROR: fileutils - missing test plan ERROR: fileutils - exited with status 134 (terminated by signal 6?) +840 static void +841 test_stdio_wrappers (void) +842 { +843 GStatBuf buf; +844 gchar *cwd, *path; +845 gint ret; +846 struct utimbuf ut; +847 GError *error = NULL; +848 +849 g_remove ("mkdir-test/test-create"); +850 g_rmdir ("mkdir-test"); +851 +852 ret = g_stat ("mkdir-test", &buf); +853 g_assert (ret == -1); +854 ret = g_mkdir ("mkdir-test", 0666); +855 g_assert (ret == 0); +856 ret = g_stat ("mkdir-test", &buf); +857 g_assert (ret == 0); +858 g_assert (S_ISDIR (buf.st_mode)); +859 +860 cwd = g_get_current_dir (); +861 path = g_build_filename (cwd, "mkdir-test", NULL); +862 g_free (cwd); +863 ret = g_chdir (path); +864 g_assert (errno == EACCES); +865 g_assert (ret == -1); So, as root the test at line 853 passed, because it was the test at 864 that was reported first - and $1 -k --tap (which is, in this case, ./fileutils -k --tap) seems to be ending in a signal 6 #define SIGABRT 6 /* (*) abort process */ Suggestions?
OK - the reason for the failure at line 853 is because there is a core file in mkdir-test, rmdir does not work when it is not empty! :) So, as root, first cleanup then run test (I have added a fprintf statement, which is why the failure is at 865 rather than 864) root@x065:[/data/prj/gnome/glib/glib-2.43.3/glib/tests]id uid=0(root) gid=0(system) groups=0(system),2(bin),3(sys),7(security),8(cron),10(audit),11(lp) root@x065:[/data/prj/gnome/glib/glib-2.43.3/glib/tests]ls -lR mkdir-test mkdir-test: total 1948 -rw-r--r-- 1 michael system 1987072 Jan 27 20:54 core root@x065:[/data/prj/gnome/glib/glib-2.43.3/glib/tests]rm -rf mkdir-test root@x065:[/data/prj/gnome/glib/glib-2.43.3/glib/tests]./fileutils /fileutils/build-path: OK /fileutils/build-pathv: OK /fileutils/build-filename: OK /fileutils/build-filenamev: OK /fileutils/mkdir-with-parents: OK /fileutils/format-size-for-display: OK /fileutils/errors: OK /fileutils/basename: OK /fileutils/dir-make-tmp: OK /fileutils/file-open-tmp: OK /fileutils/mkstemp: OK /fileutils/mkdtemp: OK /fileutils/set-contents: OK /fileutils/read-link: OK /fileutils/stdio-wrappers: test_stdio_wrappers of mkdir-test: -1 ** GLib:ERROR:fileutils.c:865:test_stdio_wrappers: assertion failed: (errno == EACCES) IOT/Abort trap(coredump) root@x065:[/data/prj/gnome/glib/glib-2.43.3/glib/tests]ls -lR mkdir-test mkdir-test: total 1944 -rw-r--r-- 1 root system 1983064 Jan 28 2015 core root@x065:[/data/prj/gnome/glib/glib-2.43.3/glib/tests] root@x065:[/data/prj/gnome/glib/glib-2.43.3/glib/tests]rm -rf mkdir-test Now after cleanup, run as michael - and runs as expected. root@x065:[/data/prj/gnome/glib/glib-2.43.3/glib/tests]su michael michael@x065:[/data/prj/gnome/glib/glib-2.43.3/glib/tests]id uid=199(michael) gid=1(staff) groups=1(staff) michael@x065:[/data/prj/gnome/glib/glib-2.43.3/glib/tests]id michael@x065:[/data/prj/gnome/glib/glib-2.43.3/glib/tests]./fileutils /fileutils/build-path: OK /fileutils/build-pathv: OK /fileutils/build-filename: OK /fileutils/build-filenamev: OK /fileutils/mkdir-with-parents: OK /fileutils/format-size-for-display: OK /fileutils/errors: OK /fileutils/basename: OK /fileutils/dir-make-tmp: OK /fileutils/file-open-tmp: OK /fileutils/mkstemp: OK /fileutils/mkdtemp: OK /fileutils/set-contents: OK /fileutils/read-link: OK /fileutils/stdio-wrappers: test_stdio_wrappers of mkdir-test: -1 OK michael@x065:[/data/prj/gnome/glib/glib-2.43.3/glib/tests] It is possible the other errors reported are the result of mkdir-test not being empty!
OK - Gdatetime failure is because AIX 5.3 TL7 does not haev International Time variables such as "Europe/Amsterdam" so there is nothing to resolve #ifdef G_OS_UNIX tz = g_time_zone_new ("America/Recife"); #elif defined G_OS_WIN32 tz = g_time_zone_new ("E. South America Standard Time"); #endif Note: I tested ASIS on an AIX 7.1 system - with ICU4C fileset installed, and it still failed - so I expect it will need a compile on such a system to succeed. Test to come later.
Additional INFO: I have recompiled on a different system with AIX 6.1 as base and see 6 errors, rather than 8. One complication - that I need to verify on the AIX 5.3 system - is that several programs do not find/use ./glib/.libs/libglib-2.0.a, i.e., programs such as glib/tests/.libs/test-spawn-echo fail to load because the library does not load. Only after this command: root@lp01:[/data/prj/gnome/glib-2.43.3]cp ./glib/.libs/libglib-2.0.a /opt/lib would all programs load. So, beside the bugs/issues that are due to AIX behavior being different than Linux behavior a number of programs may be using the wrong libglib-2.0.a during testing. After doing 'rm /opt/lib/libglib-2.0.a' again I get the following diff -u re: test-spawn-echo (seems to be the only program not loading) + ERROR: spawn-multithreaded + ========================== + + # random seed: R02S8adff1393f67adfffd002a88800ed1fa + exec(): 0509-036 Cannot load program /data/prj/gnome/glib-2.43.3/glib/tests/.libs/test-spawn-echo because of the following errors: + 0509-150 Dependent module libglib-2.0.a(libglib-2.0.so.0) could not be loaded. + 0509-022 Cannot load module libglib-2.0.a(libglib-2.0.so.0). + 0509-026 System error: A file or directory in the path name does not exist. + exec(): 0509-036 Cannot load program /data/prj/gnome/glib-2.43.3/glib/tests/.libs/test-spawn-echo because of the following errors: + 0509-150 Dependent module libglib-2.0.a(libglib-2.0.so.0) could not be loaded. + 0509-022 Cannot load module libglib-2.0.a(libglib-2.0.so.0). + 0509-026 System error: A file or directory in the path name does not exist. + ** + GLib:ERROR:spawn-multithreaded.c:77:test_spawn_sync_multithreaded_instance: assertion failed (arg == stdout_str): ("thread 5" == "") + exec(): 0509-036 Cannot load program /data/prj/gnome/glib-2.43.3/glib/tests/.libs/test-spawn-echo because of the following errors: + 0509-150 Dependent module libglib-2.0.a(libglib-2.0.so.0) could not be loaded. + 0509-022 Cannot load module libglib-2.0.a(libglib-2.0.so.0). + 0509-026 System error: A file or directory in the path name does not exist. + # Start of gthread tests + exec(): 0509-036 Cannot load program /data/prj/gnome/glib-2.43.3/glib/tests/.libs/test-spawn-echo because of the following errors: + 0509-150 Dependent module libglib-2.0.a(libglib-2.0.so.0) could not be loaded. + 0509-022 Cannot load module libglib-2.0.a(libglib-2.0.so.0). + 0509-026 System error: A file or directory in the path name does not exist. + exec(): 0509-036 Cannot load program /data/prj/gnome/glib-2.43.3/glib/tests/.libs/test-spawn-echo because of the following errors: + 0509-150 Dependent module libglib-2.0.a(libglib-2.0.so.0) could not be loaded. + 0509-022 Cannot load module libglib-2.0.a(libglib-2.0.so.0). + 0509-026 System error: A file or directory in the path name does not exist. + exec(): 0509-036 Cannot load program /data/prj/gnome/glib-2.43.3/glib/tests/.libs/test-spawn-echo because of the following errors: + 0509-150 Dependent module libglib-2.0.a(libglib-2.0.so.0) could not be loaded. + 0509-022 Cannot load module libglib-2.0.a(libglib-2.0.so.0). + 0509-026 System error: A file or directory in the path name does not exist. + exec(): 0509-036 Cannot load program /data/prj/gnome/glib-2.43.3/glib/tests/.libs/test-spawn-echo because of the following errors: + 0509-150 Dependent module libglib-2.0.a(libglib-2.0.so.0) could not be loaded. + 0509-022 Cannot load module libglib-2.0.a(libglib-2.0.so.0). + 0509-026 System error: A file or directory in the path name does not exist. + exec(): 0509-036 Cannot load program /data/prj/gnome/glib-2.43.3/glib/tests/.libs/test-spawn-echo because of the following errors: + 0509-150 Dependent module libglib-2.0.a(libglib-2.0.so.0) could not be loaded. + 0509-022 Cannot load module libglib-2.0.a(libglib-2.0.so.0). + 0509-026 System error: A file or directory in the path name does not exist. + exec(): 0509-036 Cannot load program /data/prj/gnome/glib-2.43.3/glib/tests/.libs/test-spawn-echo because of the following errors: + 0509-150 Dependent module libglib-2.0.a(libglib-2.0.so.0) could not be loaded. + 0509-022 Cannot load module libglib-2.0.a(libglib-2.0.so.0). + 0509-026 System error: A file or directory in the path name does not exist. + exec(): 0509-036 Cannot load program /data/prj/gnome/glib-2.43.3/glib/tests/.libs/test-spawn-echo because of the following errors: + 0509-150 Dependent module libglib-2.0.a(libglib-2.0.so.0) could not be loaded. + 0509-022 Cannot load module libglib-2.0.a(libglib-2.0.so.0). + 0509-026 System error: A file or directory in the path name does not exist. + exec(): 0509-036 Cannot load program /data/prj/gnome/glib-2.43.3/glib/tests/.libs/test-spawn-echo because of the following errors: + 0509-150 Dependent module libglib-2.0.a(libglib-2.0.so.0) could not be loaded. + 0509-022 Cannot load module libglib-2.0.a(libglib-2.0.so.0). + 0509-026 System error: A file or directory in the path name does not exist. + exec(): 0509-036 Cannot load program /data/prj/gnome/glib-2.43.3/glib/tests/.libs/test-spawn-echo because of the following errors: + 0509-150 Dependent module libglib-2.0.a(libglib-2.0.so.0) could not be loaded. + 0509-022 Cannot load module libglib-2.0.a(libglib-2.0.so.0). + 0509-026 System error: A file or directory in the path name does not exist. + exec(): 0509-036 Cannot load program /data/prj/gnome/glib-2.43.3/glib/tests/.libs/test-spawn-echo because of the following errors: + 0509-150 Dependent module libglib-2.0.a(libglib-2.0.so.0) could not be loaded. + 0509-022 Cannot load module libglib-2.0.a(libglib-2.0.so.0). + 0509-026 System error: A file or directory in the path name does not exist. + exec(): 0509-036 Cannot load program /data/prj/gnome/glib-2.43.3/glib/tests/.libs/test-spawn-echo because of the following errors: + 0509-150 Dependent module libglib-2.0.a(libglib-2.0.so.0) could not be loaded. + 0509-022 Cannot load module libglib-2.0.a(libglib-2.0.so.0). + 0509-026 System error: A file or directory in the path name does not exist. + exec(): 0509-036 Cannot load program /data/prj/gnome/glib-2.43.3/glib/tests/.libs/test-spawn-echo because of the following errors: + exec(): 0509-036 Cannot load libglib-/data/prj/gnome/glib-2. could not be loaded. + test-spawn-echo because of the following errors: + 0509-022 Cannot load module 0509-150 Dependent module libglib-2.0.a(libglib-2.0.so.0) could not be loaded. + 0509-022 Cannot load module libglib-2.0.a(libglib-2.0.so.0). + 0509-026 System error 0A file or directory : A file or directory in the path name does not exist. + ../../tap-test[5]: 5898470 IOT/Abort trap(coredump) + # GLib:ERROR:spawn-multithreaded.c:77:test_spawn_sync_multithreaded_instance: assertion failed (arg == stdout_str): ("thread 5" == "") + ERROR: spawn-multithreaded - missing test plan + ERROR: spawn-multithreaded - exited with status 134 (terminated by signal 6?) + + ERROR: spawn-singlethread + ========================= + + # random seed: R02S69867a5430c879e894823d174739d7ae + exec(): 0509-036 Cannot load program /data/prj/gnome/glib-2.43.3/glib/tests/.libs/test-spawn-echo because of the following errors: + 0509-150 Dependent module libglib-2.0.a(libglib-2.0.so.0) could not be loaded. + 0509-022 Cannot load module libglib-2.0.a(libglib-2.0.so.0). + 0509-026 System error: A file or directory in the path name does not exist. + ** + GLib:ERROR:spawn-singlethread.c:178:test_spawn_sync: assertion failed (arg == stdout_str): ("thread 1" == "") + # Start of gthread tests + ../../tap-test[5]: 13041918 IOT/Abort trap(coredump) + # GLib:ERROR:spawn-singlethread.c:178:test_spawn_sync: assertion failed (arg == stdout_str): ("thread 1" == "") + ERROR: spawn-singlethread - missing test plan + ERROR: spawn-singlethread - exited with status 134 (terminated by signal 6?) + PASS: testing ============= ** I shall look deeper into root-cause for the other errors
root-cause for ERROR with tz = g_time_zone_new ("America/Recife"); For AIX the TZDIR is "/usr/share/lib/zoneinfo" diff -ur a/gtimezone.c b/gtimezone.c --- a/gtimezone.c 2014-12-03 07:03:32.000000000 +0000 +++ b/gtimezone.c 2015-02-01 17:13:37.000000000 +0000 @@ -410,7 +410,11 @@ tzdir = getenv ("TZDIR"); if (tzdir == NULL) + # ifndef _AIX tzdir = "/usr/share/zoneinfo"; + # else + tzdir = "/usr/share/lib/zoneinfo"; + # endif if (*identifier == ':') identifier ++; Unfortunately, this does not work on AIX 5.3 - because of a later test. 1) AIX 5.3 fileset bos.rte.date does not have info for "America/Recife" 2) Changing the string to "Brazil/East" als fails, but later with (.libs/gdatetime:659596): GLib-CRITICAL **: init_zone_from_iana_info: assertion 'size >= sizeof (struct tzhead) && memcmp (header, "TZif", 4) == 0' failed so I am assuming that TZif != 4 (not sure how to determine that quickly) It could/would have worked because - I think - because gmtoff is -3 hours: root@x065:[/data/prj/gnome/glib/glib-2.43.3/glib/tests]zdump -c 2016 -v Brazil/East | grep 2015 Brazil/East Sun Feb 1 13:47:33 2015 CUT Brazil/East Sat Feb 14 03:59:59 2015 GMT = Sat Feb 14 00:29:59 2015 CUT isdst=0 gmtoff = -12600 Brazil/East Sat Feb 14 04:00:00 2015 GMT = Sat Feb 14 00:30:00 2015 CUT isdst=0 gmtoff = -12600 Brazil/East Fri Oct 23 04:59:59 2015 GMT = Fri Oct 23 01:29:59 2015 CUT isdst=0 gmtoff = -12600 Brazil/East Fri Oct 23 05:00:00 2015 GMT = Fri Oct 23 01:30:00 2015 CUT isdst=0 gmtoff = -12600 ** One down for later than AIX 5.3 **
gvariant is 'failing' regularly, but it is also passing (see attachment gvariant-odd-results.text) By default AIX is building a 32-bit application, and by default there is "only" 256MByte for the combined .data, .bss and .stack area. Is it possible that the combined tests are nearing this borderline - because sometimes it passes all, sometimes it does not - and I have seen it fail in different tests. Ideas?
Created attachment 295898 [details] results of running gvariant test manually - sometimes SUCCESS, sometimes not
The mappedfile error seems to occur because the user_cache_dir is not made by default if it does not exist. Maybe I missed something elsewhere - but even after having unpacked, and rebuilt - several times - $HOME/.cache is not made in advance and this test fails when it cannot create a file in /.cache So, please check my 'mappedfiles-research.text (trying to keep the wall of text down a bit) - maybe I missed something there. Michael
Created attachment 295899 [details] CLI testing, grep, etc. to determine a possible root-cause for mappedfile error
The last error I found in the original test-suite.log was: # End of skip-all tests # End of misc tests # Start of forking tests # testing:ERROR:testing.c:103:test_fork_fail: code should not be reached ok 21 /forking/fail assertion PASS: testing 21 /forking/fail assertion ok 22 /forking/patterns PASS: testing 22 /forking/patterns Seems to be resolved - no idea why. root@x065:[/data/prj/gnome/glib/glib-2.43.3/glib/tests]make testing.log PASS: testing 1 /random-generator/rand-1 PASS: testing 2 /random-generator/rand-2 PASS: testing 3 /random-generator/random-conversions PASS: testing 4 /misc/assertions PASS: testing 5 /misc/test-data PASS: testing 6 /misc/primetoul PASS: testing 7 /misc/fatal-log-handler PASS: testing 8 /misc/expected-messages PASS: testing 9 /misc/dash-p PASS: testing 10 /misc/nonfatal SKIP: testing 11 /misc/skip # SKIP Skipped should count as passed, not failed PASS: testing 12 /misc/skip-all PASS: testing 13 /misc/fail PASS: testing 14 /misc/incomplete PASS: testing 15 /misc/timeout PASS: testing 16 /misc/expected-messages/expect-error PASS: testing 17 /misc/expected-messages/skip-debug PASS: testing 18 /misc/dash-p/child PASS: testing 19 /misc/dash-p/child/sub PASS: testing 20 /misc/dash-p/child/sub2 PASS: testing 21 /forking/fail assertion PASS: testing 22 /forking/patterns PASS: testing 23 /trap_subprocess/fail PASS: testing 24 /trap_subprocess/no-such-test PASS: testing 25 /trap_subprocess/patterns I will continue this, but need some suggestions for gvariant, while I suspect the mappedfile test will succeed if I create $HOME/.cache before 'make check' after a fresh make. The error re: gtimezone.c - I will attach a patch - even though this is perhaps too simple - as it creates a new error on AIX 5.3, but allows gdatetime test to pass on AIX 6.1
Created attachment 295900 [details] fixes default directory name for TZDIR for AIX A rather simple patch - as it only fixes part of the problem on AIX 5.3, because the timezone information on AIX 5.3 does not pass a later test. On AIX 5.3, with a correct TZDIR the glib/tests/gdatatime test fails with: (/data/prj/gnome/glib/glib-2.43.3/glib/tests/.libs/lt-gdatetime:676014): GLib-CRITICAL **: init_zone_from_iana_info: assertion 'memcmp (header, "TZif", 4) == 0' failed And as this is a test in the library there is a chance of aborts for applications when ICU timezone info is used (on AIX 5.3). Note: for debugging I modified the original test to two single tests: g_return_if_fail (size >= sizeof (struct tzhead)); g_return_if_fail (memcmp (header, "TZif", 4) == 0); so that it is very clear that the failure for AIX 5.3 is coming from the memcmp() test.
From http://www.tablespace.net/papers/unix_time.html Versions of AIX prior to 6.1 include an Olsen-like database in /usr/share/lib/zoneinfo but they do not use it for system time conversions. These files are not technically TZif files as they are not of the same format but do use zic and zdump utilities to create and display the contents of the compiled files. (Note that AIX 6.1 and later systems actually use Olsen TZif files.) So, for AIX 5.3 the test will always fail. Is this a good reason to stop all further testing? Yes, if all further tests depend on this, NO, if they do not. Question: rather than hard-code this default, is it perhaps something to put into ./configure? What is more 'portable' versus more maintainable (doubtful there will be many other exceptions come up anno 2015+) regards, Michael p.s. Not much more INFO I can think to give. If something additional is desired, please ask.
-- GitLab Migration Automatic Message -- This bug has been migrated to GNOME's GitLab instance and has been closed from further activity. You can subscribe and participate further through the new bug through this link to our GitLab instance: https://gitlab.gnome.org/GNOME/glib/issues/985.