From Bugzilla Helper: User-Agent: Mozilla/5.0 (X11; U; Linux i686; en-US; rv:1.4) Gecko/20030830 Debian/1.4-3.he-1 Description of problem: Fails test which seems to related to glibc-2.3.2-77 (iconv) ===================================================================== TIME END 2003-09-09 03:44:06 ===================================================================== TEST RESULT SUMMARY --------------------------------------------------------------------- Exts skipped : 54 Exts tested : 33 --------------------------------------------------------------------- Number of tests : 476 Tests skipped : 113 (23.7%) Tests warned : 0 ( 0.0%) Tests failed : 1 ( 0.2%) Tests passed : 362 (76.1%) --------------------------------------------------------------------- Time taken : 226 seconds ===================================================================== ===================================================================== FAILED TEST SUMMARY --------------------------------------------------------------------- Bug #16069 [ext/iconv/tests/bug16069.phpt] ===================================================================== make: *** [test] Error 1 ++ find .. -name '*.diff' -type f -print + cat ../ext/iconv/tests/bug16069.diff Version-Release number of selected component (if applicable): php-4.3.2-7 How reproducible: Always Steps to Reproduce: 1. Try to recompile again current Actual Results: Fails to recompile Additional info: System is up to current from up2date.
php-4.3.2-8.ent turned that test off, though it would be interesting to know whether this really is a bug in the glibc iconv.
Gimme few hours. I'll install a clean beta2 and start the iteration from there.
Duh! Mystical enought the buiding works aOK if i use command like LC_ALL=fi_FI@euro rpmbuild -ba php.spec I really don't know enought about php and glibc to make any assumtions. The system was stil 'up to date EL3-beta'.
Ah, good catch. What is your $LANG setting normally? I also had to unset TZ to fix some build failures, must be some env vars are getting leaked which weren't previously or something.
My env for these 'test-setups' are just pretty much defaultwhich would be LANG=en_US.UTF-8 LANGVAR=en_US.UTF-8 (those seems to be only related ones) The spec-file at least claims to be setting 'LANG=C' as usually Actually 'LC_ALL=fi_FI' builds just fine too, but for example 'LC_ALL=fi_FI.UTF-8' does not. So it seems to be UTF-8 related.
Shouldn't this be closed already?
Yes, this was fixed for the final release. Thanks for the report.