#40189 [Asn]: endless loop in zlib.inflate stream filter

ID: 40189 Updated by: cellog@php.net Reported By: cellog@php.net Status: Assigned Bug Type: Zlib Related Operating System: kubuntu linux PHP Version: 5CVS-2007-01-22 (CVS) Assigned To: pollita New Comment: last detail: I have a 64-bit machine, not sure if that matters or not, but there you go Previous Comments: ------------------------------------------------------------------------ [2007-01-23 23:57:55] cellog@php.net not sure what this cvs update -f business was. Forget that. Use this to reproduce the bug: cd pecl/phar cvs update -D "Jan 22 01:00:00 2007 UTC" ln -s ~/php5/ext/phar ~/pecl/phar cd ~/php5 ./configure --enable-debug --enable-phar --with-zlib make cli cd ext/phar/tests ~/php5/sapi/cli/php phar_ctx_001.phpt ~/php5/sapi/cli/php phar_ctx_001.phpt or ~/php5/sapi/cli/php phar_oo_compressed_001.phpt ~/php5/sapi/cli/php phar_oo_compressed_001.phpt (sometimes it only happened on the 2nd run for me) ------------------------------------------------------------------------ [2007-01-22 14:44:59] cellog@php.net cvs update -f "Jan 22 02:07:44 2007 UTC" I committed a workaround for the failure, completely forgetting you needed the file for this bug, sorry. This is the patch that needs to be reversed: http://cvs.php.net/viewvc.cgi/pecl/phar/phar.c?r1=1.135&r2=1.136 ------------------------------------------------------------------------ [2007-01-22 09:46:45] tony2001@php.net I have to say, I can't reproduce it on my Suse either. 4 tests fail, but the ones you mentioned work fine. Btw, phar_oo_compressed_001.phpt fails because of several leaks: 016+ /local/dev/php/5_2/main/streams/streams.c(386) : Stream of type 'MEMORY' 0x40203094 (path:(null)) was not closed 017+ /local/dev/php/5_2/main/streams/streams.c(386) : Stream of type 'TEMP' 0x402004f8 (path:(null)) was not closed You need to close those streams when destroying the object. ------------------------------------------------------------------------ [2007-01-22 08:55:57] mike@php.net Seems to work fine here... (incidentally same OS and PHP version): mike@honeybadger:~/cvs/pecl/phar/tests$ php -d"phar.require_hash=0" -d"phar.readonly=0" phar_oo_compressed_001.phpt --TEST-- Phar context --SKIPIF-- --INI-- phar.require_hash=0 phar.readonly=0 --FILE-- string(1) "a" bool(false) string(1) "b" bool(false) string(1) "c" bool(false) string(5) "new a" bool(false) string(5) "new b" bool(true) string(1) "c" bool(false) string(5) "new d" bool(false) ===DONE=== --CLEAN-- --EXPECTF-- string(1) "a" bool(false) string(1) "b" bool(false) string(1) "c" bool(false) string(5) "new a" bool(false) string(5) "new b" bool(true) string(1) "c" bool(false) string(5) "new d" bool(false) ===DONE=== mike@honeybadger:~/cvs/pecl/phar/tests$ php -d"phar.require_hash=0" -d"phar.readonly=0" phar_ctx_001.phpt --TEST-- Phar context --SKIPIF-- --INI-- phar.require_hash=0 phar.readonly=0 --FILE-- string(1) "a" bool(false) string(1) "b" bool(false) string(1) "c" bool(false) string(5) "new a" bool(false) string(5) "new b" bool(true) string(1) "c" bool(false) string(5) "new d" bool(false) ===DONE=== --CLEAN-- --EXPECTF-- string(1) "a" bool(false) string(1) "b" bool(false) string(1) "c" bool(false) string(5) "new a" bool(false) string(5) "new b" bool(true) string(1) "c" bool(false) string(5) "new d" bool(false) ===DONE=== Tests that fail here: FAILED TEST SUMMARY --------------------------------------------------------------------- Phar::mapPhar valid file (bzip2) [/home/mike/cvs/pecl/phar/tests/015b.phpt] Phar::mapPhar invalid file (gzipped file length is too short) [/home/mike/cvs/pecl/phar/tests/016.phpt] Phar::mapPhar invalid file (gzipped file length is too short) [/home/mike/cvs/pecl/phar/tests/016b.phpt] Phar: test that refcounting avoids problems with deleting a file [/home/mike/cvs/pecl/phar/tests/refcount1.phpt] ------------------------------------------------------------------------ [2007-01-22 07:17:03] tony2001@php.net Sara, could you look into this? Btw, I can see the very same code in ext/bz2. ------------------------------------------------------------------------ The remainder of the comments for this report are too long. To view the rest of the comments, please view the bug report online at http://bugs.php.net/40189 -- Edit this bug report at http://bugs.php.net/?id=40189&edit=1