Edit report at https://bugs.php.net/bug.php?id=63867&edit=1
ID: 63867 Updated by: php-bugs@lists.php.net Reported by: mamfelt at gmail dot com Summary: make test core dumps on bug63377.php -Status: Feedback +Status: No Feedback Type: Bug Package: Testing related Operating System: AIX 6.1 PHP Version: 5.3.20 New Comment: No feedback was provided. The bug is being suspended because we assume that you are no longer experiencing the problem. If this is not the case and you are able to provide the information that was requested earlier, please do so and change the status of the bug back to "Open". Thank you. Previous Comments: ------------------------------------------------------------------------ [2012-12-28 13:54:10] mamfelt at gmail dot com Adding: $ LDR_CNTRL="MAXDATA=0x8000000" $ export LDR_CNTRL $ make test and system moves on, so not a bug - a linker flag needs to be added: -bmaxdata:0x80000000 This becomes an FYI for others. ------------------------------------------------------------------------ [2012-12-28 13:49:49] fel...@php.net Hi, I've fixed an issue in this test. (bug63377.phpt) Please, apply the change manually and try again: http://git.php.net/?p=php-src.git;a=commitdiff;h=56d9edbbb93a2a338211b2e7c05cf52347a963ce ------------------------------------------------------------------------ [2012-12-28 13:43:44] mamfelt at gmail dot com Description: ------------ make test fails (fatal error) on test 63377 and does not continue with remaining tests. AIX 6.1, two different TL levels, one with 2G RAM and 1.5G paging space one with 9G RAM, 512 MByte paging space michael@x054:[/data/prj/php/php-5.3.20]ulimit -a time(seconds) unlimited file(blocks) unlimited data(kbytes) unlimited stack(kbytes) 4194304 memory(kbytes) unlimited coredump(blocks) 0 nofiles(descriptors) 2000 threads(per process) unlimited processes(per user) unlimited -------- I am aware that AIX has a different HW memory model (10 or 11 256M Byte memory segments, so it may lie in the 3072M limit used in the test. However, I do not really see the system grab that much memory before the test crashes. I will assume it is hard to spot using topas/nmon. The test is also running a loop that is allocating? and writing? 1024 x 1024 x 100 'a' characters (or 22 x 102400k) Some suggestions, read instructions, for: 1) setting up a situation that I can debug this test to see what is failing 2) instructions for the letting make test continue, even though this has failed. 3) instructions on how to skip this test for the time being. Test script: --------------- # ./configure ... # make # make test Expected result: ---------------- that "make test" completes normally, asking to send test results, etc. Actual result: -------------- PASS ZE2 type hinting [tests/lang/type_hints_002.phpt] PASS ZE2 type hinting [tests/lang/type_hints_003.phpt] PASS Bug #46897: ob_flush() should fail to flush unerasable buffers [tests/output/bug46897.phpt] TEST 630/9330 [tests/output/bug63377.phpt] Fatal error: Out of memory (allocated 213647360) (tried to allocate 209715545 bytes) in /data/prj/php/php-5.3.20/run-tests.php on line 1818 ------------------------------------------------------------------------ -- Edit this bug report at https://bugs.php.net/bug.php?id=63867&edit=1