Re: [Vala] Vala broken on non-gnu systems
Those defines should be defined by vala internally? All systems can be identified by __linux__ __FreeBSD__ ... do you think this should be defined by vala? Autoconf is imho not the right place for this. On 23/03/2011, at 12:41, Marco Trevisan (Treviño) m...@3v1n0.net wrote: Il giorno mer, 16/03/2011 alle 14.05 +0100, pancake ha scritto: /* Result of the benchmark: $ time ./a.out 1 real0m0.225s user0m0.223s sys 0m0.000s $ time ./a.out 0 real0m0.380s user0m0.377s sys 0m0.000s */ Well, since the difference seems sensible, it would be better if valac would export something like HAVE_POSIX when possible (or HAVE_{LINUX,BSD,WIN} ... ) and use them in VAPIs (and in vala files). I guess this would allow some code optinizations without a build system... ___ vala-list mailing list vala-list@gnome.org http://mail.gnome.org/mailman/listinfo/vala-list ___ vala-list mailing list vala-list@gnome.org http://mail.gnome.org/mailman/listinfo/vala-list
Re: [Vala] Vala broken on non-gnu systems
Il giorno mer, 16/03/2011 alle 14.05 +0100, pancake ha scritto: /* Result of the benchmark: $ time ./a.out 1 real0m0.225s user0m0.223s sys 0m0.000s $ time ./a.out 0 real0m0.380s user0m0.377s sys 0m0.000s */ Well, since the difference seems sensible, it would be better if valac would export something like HAVE_POSIX when possible (or HAVE_{LINUX,BSD,WIN} ... ) and use them in VAPIs (and in vala files). I guess this would allow some code optinizations without a build system... ___ vala-list mailing list vala-list@gnome.org http://mail.gnome.org/mailman/listinfo/vala-list
Re: [Vala] Vala broken on non-gnu systems
On 16 March 2011 13:49, Anatol Pomozov anatol.pomo...@gmail.com wrote: On Tue, Mar 15, 2011 at 10:46 PM, Alexander Kojevnikov alexan...@kojevnikov.com wrote: On 16 March 2011 12:07, Anatol Pomozov anatol.pomo...@gmail.com wrote: While you are here you might want to fix tests/testrunner.sh that also uses GNU extension (xargs -r). See line 167. Which Unix are you using? macosx 10.6.4 I filed this as bgo#644884 [0], please test the attached patch and comment in the bug. [0] https://bugzilla.gnome.org/show_bug.cgi?id=644884 ___ vala-list mailing list vala-list@gnome.org http://mail.gnome.org/mailman/listinfo/vala-list
Re: [Vala] Vala broken on non-gnu systems
Hi On Tue, Mar 15, 2011 at 11:43 PM, Alexander Kojevnikov alexan...@kojevnikov.com wrote: On 16 March 2011 13:49, Anatol Pomozov anatol.pomo...@gmail.com wrote: On Tue, Mar 15, 2011 at 10:46 PM, Alexander Kojevnikov alexan...@kojevnikov.com wrote: On 16 March 2011 12:07, Anatol Pomozov anatol.pomo...@gmail.com wrote: While you are here you might want to fix tests/testrunner.sh that also uses GNU extension (xargs -r). See line 167. Which Unix are you using? macosx 10.6.4 I filed this as bgo#644884 [0], please test the attached patch and comment in the bug. [0] https://bugzilla.gnome.org/show_bug.cgi?id=644884 Current HEAD is broken on macosx so I checked out revision before the Jurg's change $ git checkout a2d2120b~ applied your patch and it fixes the issue with xargs -r. But now tests are broken because of other error: make check-TESTS TEST: Building... FAIL cc1: error: -Wuninitialized is not supported without -O error: cc exited with status 256 Compilation failed: 1 error(s), 0 warning(s) make[4]: *** [check-TESTS] Error 1 make[3]: *** [check-am] Error 2 make[2]: *** [check] Error 2 make[1]: *** [check-recursive] Error 1 make: *** [check] Error 2 ___ vala-list mailing list vala-list@gnome.org http://mail.gnome.org/mailman/listinfo/vala-list
Re: [Vala] Vala broken on non-gnu systems
Hi On Wed, Mar 16, 2011 at 1:37 AM, Alexander Kojevnikov alexan...@kojevnikov.com wrote: On 16 March 2011 16:33, Anatol Pomozov anatol.pomo...@gmail.com wrote: But now tests are broken because of other error: make check-TESTS TEST: Building... FAIL cc1: error: -Wuninitialized is not supported without -O error: cc exited with status 256 Compilation failed: 1 error(s), 0 warning(s) make[4]: *** [check-TESTS] Error 1 make[3]: *** [check-am] Error 2 make[2]: *** [check] Error 2 make[1]: *** [check-recursive] Error 1 make: *** [check] Error 2 This should be fixed by the patch in the other bug I mentioned, please give it a try: https://bugzilla.gnome.org/show_bug.cgi?id=644876 Thanks for your patches. It is much better now. Now only dbus tests fail with /dbus/basic-types/server:FAIL Failed to start message bus: launchd's environment variable DBUS_LAUNCHD_SESSION_BUS_SOCKET is empty, but should contain a socket path. EOF in dbus-launch reading address from bus daemon ** CRITICAL **: file /Users/anatol/sources/vala/tests/_test/main.c: line 19824: uncaught error: Error connecting: No such file or directory (g-io-error-quark, 1) aborting... dbus.basic_types.server.check: line 4: 74545 Abort trap ./test /dbus/basic-types/server kill: usage: kill [-s sigspec | -n signum | -sigspec] pid | jobspec ... or kill -l [sigspec] It looks similar to this issue https://trac.macports.org/ticket/26712 ___ vala-list mailing list vala-list@gnome.org http://mail.gnome.org/mailman/listinfo/vala-list
Re: [Vala] Vala broken on non-gnu systems
/* Result of the benchmark: $ time ./a.out 1 real0m0.225s user0m0.223s sys 0m0.000s $ time ./a.out 0 real0m0.380s user0m0.377s sys 0m0.000s */ #include string.h #include sys/types.h #define ITERS 10 int slen = ; char *str; static inline size_t mystrnlen(const char* str, size_t maxlen) { char* p = memchr(str, 0, maxlen); if (p == NULL) return maxlen; else return (p - str); } int test_memchr () { int i; int ret = 0; for (i=0; i ITERS; i++) { ret += mystrnlen (str, slen); } return ret; } void test_strnlen () { int i; int ret = 0; for (i=0; i ITERS; i++) { ret += strnlen (str, slen); } return ret; } main(int argc, char **argv) { if (argc1) { str = malloc (slen+1); memset (str, 'A', slen); str[slen] = 0; switch (atoi (argv[1])) { case 0: test_memchr (); break; case 1: test_strnlen (); break; } } else printf (Usage: memchr [0|1]\n); } On 03/16/11 09:59, Anatol Pomozov wrote: Hi On Wed, Mar 16, 2011 at 1:37 AM, Alexander Kojevnikov alexan...@kojevnikov.com wrote: On 16 March 2011 16:33, Anatol Pomozovanatol.pomo...@gmail.com wrote: But now tests are broken because of other error: make check-TESTS TEST: Building... FAIL cc1: error: -Wuninitialized is not supported without -O error: cc exited with status 256 Compilation failed: 1 error(s), 0 warning(s) make[4]: *** [check-TESTS] Error 1 make[3]: *** [check-am] Error 2 make[2]: *** [check] Error 2 make[1]: *** [check-recursive] Error 1 make: *** [check] Error 2 This should be fixed by the patch in the other bug I mentioned, please give it a try: https://bugzilla.gnome.org/show_bug.cgi?id=644876 Thanks for your patches. It is much better now. Now only dbus tests fail with /dbus/basic-types/server:FAIL Failed to start message bus: launchd's environment variable DBUS_LAUNCHD_SESSION_BUS_SOCKET is empty, but should contain a socket path. EOF in dbus-launch reading address from bus daemon ** CRITICAL **: file /Users/anatol/sources/vala/tests/_test/main.c: line 19824: uncaught error: Error connecting: No such file or directory (g-io-error-quark, 1) aborting... dbus.basic_types.server.check: line 4: 74545 Abort trap ./test /dbus/basic-types/server kill: usage: kill [-s sigspec | -n signum | -sigspec] pid | jobspec ... or kill -l [sigspec] It looks similar to this issue https://trac.macports.org/ticket/26712 ___ vala-list mailing list vala-list@gnome.org http://mail.gnome.org/mailman/listinfo/vala-list ___ vala-list mailing list vala-list@gnome.org http://mail.gnome.org/mailman/listinfo/vala-list
Re: [Vala] Vala broken on non-gnu systems
On Tue, 2011-03-15 at 20:05 +0100, pancake wrote: The commit a2s2120.. dated at 2011-02-10 adds the use of strnlen function in glib-2.0.vapi. This function is a gnu extension on libc. This is not glib and this function is only available in GNU LibC. Do the build breaks on OSX, Windows and probably all BSD systems. This is now fixed in master. strnlen is in POSIX 2008 but Mac OS X doesn't appear to support it yet. Jürg ___ vala-list mailing list vala-list@gnome.org http://mail.gnome.org/mailman/listinfo/vala-list
Re: [Vala] Vala broken on non-gnu systems
Hi On Tue, Mar 15, 2011 at 12:05 PM, pancake panc...@nopcode.org wrote: The commit a2s2120.. dated at 2011-02-10 adds the use of strnlen function in glib-2.0.vapi. This function is a gnu extension on libc. This is not glib and this function is only available in GNU LibC. Do the build breaks on OSX, Windows and probably all BSD systems. It would be great if someone with commit access (Jurg?) fixes that. While you are here you might want to fix tests/testrunner.sh that also uses GNU extension (xargs -r). See line 167. ___ vala-list mailing list vala-list@gnome.org http://mail.gnome.org/mailman/listinfo/vala-list
Re: [Vala] Vala broken on non-gnu systems
On 16 March 2011 12:07, Anatol Pomozov anatol.pomo...@gmail.com wrote: While you are here you might want to fix tests/testrunner.sh that also uses GNU extension (xargs -r). See line 167. Which Unix are you using? FreeBSD's xarg works fine with or without the -r option: -r Compatibility with GNU xargs. The GNU version of xargs runs the utility argument at least once, even if xargs input is empty, and it supports a -r option to inhibit this behavior. The FreeBSD version of xargs does not run the utility argument on empty input, but it supports the -r option for command-line compatibil- ity with GNU xargs, but the -r option does nothing in the FreeBSD version of xargs. On a side note, there are other issues with the test runner under FreeBSD, they should be fixed by the patch attached to bgo#644876 [0] [0] https://bugzilla.gnome.org/show_bug.cgi?id=644876 ___ vala-list mailing list vala-list@gnome.org http://mail.gnome.org/mailman/listinfo/vala-list
Re: [Vala] Vala broken on non-gnu systems
Hi On Tue, Mar 15, 2011 at 10:46 PM, Alexander Kojevnikov alexan...@kojevnikov.com wrote: On 16 March 2011 12:07, Anatol Pomozov anatol.pomo...@gmail.com wrote: While you are here you might want to fix tests/testrunner.sh that also uses GNU extension (xargs -r). See line 167. Which Unix are you using? macosx 10.6.4 FreeBSD's xarg works fine with or without the -r option: -r Compatibility with GNU xargs. The GNU version of xargs runs the utility argument at least once, even if xargs input is empty, and it supports a -r option to inhibit this behavior. The FreeBSD version of xargs does not run the utility argument on empty input, but it supports the -r option for command-line compatibil- ity with GNU xargs, but the -r option does nothing in the FreeBSD version of xargs. On a side note, there are other issues with the test runner under FreeBSD, they should be fixed by the patch attached to bgo#644876 [0] [0] https://bugzilla.gnome.org/show_bug.cgi?id=644876 ___ vala-list mailing list vala-list@gnome.org http://mail.gnome.org/mailman/listinfo/vala-list