Re: [PHP-DEV] 5.0.5RC1 rolled

2005-08-31 Thread steve
I was a little early in posting the bt -- it seems that that there must be an exception tree, or multiple exceptions -- I have several successive cont for which I got a bt. This is all from a single request. Sometimes, between call_user_function and call_user_function_ex the parameters like functi

[PHP-DEV] Re: Re was: [PHP-DEV] 5.0.5RC1 rolled

2005-08-31 Thread steve
This also happens in 5.1.0RC1, btw: Breakpoint 1, zend_throw_exception_internal (exception=0xa19baa4) at /root/webserver_software_tmp_fastcgi_apache2/php-5.1.0RC1/Zend/zend_exceptions.c:39 39 if (exception != NULL) { (gdb) bt #0 zend_throw_exception_internal (exception=0xa19baa4) at

[PHP-DEV] Re: Re was: [PHP-DEV] 5.0.5RC1 rolled

2005-08-30 Thread steve
On 8/30/05, Marcus Boerger <[EMAIL PROTECTED]> wrote: > in 15 you see that all parameters are bork! If that is true, then 10, 15, and 24 (all call_user_function_ex) are borked. 24 would be the earliest. Maybe something is being cleaned up at the start of shutdown but before the user shutdown funct

Re: [PHP-DEV] Re: Re was: [PHP-DEV] 5.0.5RC1 rolled

2005-08-30 Thread steve
This was with: /usr/bin/valgrind --tool=memcheck --leak-check=yes -v --trace-children=yes And it didn't seem to add anything while the errorlog was being written. # more log.pid17870 ==17870== Memcheck, a memory error detector for x86-linux. ==17870== Copyright (C) 2002-2004, and GNU GPL'd, by

Re: [PHP-DEV] Re: Re was: [PHP-DEV] 5.0.5RC1 rolled

2005-08-30 Thread steve
Yea, I noticed that too... #!/bin/bash export PHPRC export PHP_FCGI_CHILDREN exec /usr/bin/valgrind --tool=memcheck --trace-children=yes \ --log-file=/usr/local/apache2/fastcgi-bin/log /usr/local/apache2/fastcgi-bin/php505.fcgi $@ This one works. :) However... it didn't seem to find much. -- PH

Re: [PHP-DEV] Re: Re was: [PHP-DEV] 5.0.5RC1 rolled

2005-08-30 Thread Marcus Boerger
Hello steve, Tuesday, August 30, 2005, 11:32:33 PM, you wrote: > :) > Can valgrind attach to a process like gdb? I somehow doubt it... In > that case, I must admit, I've never used valgrind before. How can I > set up valgrind to check a fast-cgi process? > I tried changing the php5.fcgi file to

[PHP-DEV] Re: Re was: [PHP-DEV] 5.0.5RC1 rolled

2005-08-30 Thread steve
:) Can valgrind attach to a process like gdb? I somehow doubt it... In that case, I must admit, I've never used valgrind before. How can I set up valgrind to check a fast-cgi process? I tried changing the php5.fcgi file to php505.fcgi and created this as php5.fcgi: #/bin/bash /usr/bin/valgrind -

[PHP-DEV] Re was: [PHP-DEV] 5.0.5RC1 rolled

2005-08-30 Thread Marcus Boerger
Hello steve, #15 0x081ccec4 in call_user_function_ex (function_table=0x1, object_pp=0x1, function_name=0x1, retval_ptr_ptr=0x1, param_count=1, params=0x1, no_separation=1, symbol_table=0x1) at /root/webserver_software_tmp_fastcgi_apache2/php-5.0.5RC2/Zend/zend_execute_API.c:559 #16 0x081e53eb

Re: [PHP-DEV] 5.0.5RC1 rolled

2005-08-30 Thread steve
Thats what I was looking for, thanks. I used that and no exception was thrown in 5.0.4, but there is one in 5.0.5RC2. It is in a shutdown function, which is why I couldn't figure out where it was happening. Not to mention, it was in an AJAX call. The shutdown function is defined in JPSpan (http://s

Re: [PHP-DEV] 5.0.5RC1 rolled

2005-08-29 Thread Marcus Boerger
Hello steve, for a c backtrace set a breakpoint on function zend_throw_exception_internal. regards marcus Monday, August 29, 2005, 11:24:16 PM, you wrote: > I get this error in the error log, on every page view in this release: > PHP Fatal error: Exception thrown without a stack frame in Unk

Re: [PHP-DEV] 5.0.5RC1 rolled

2005-08-29 Thread steve
I get this error in the error log, on every page view in this release: PHP Fatal error: Exception thrown without a stack frame in Unknown on line 0 It seems to happen after the page is complete, since from the webpage it renders fine. I found a bug reference about throwing exceptions in __destr

Re: [PHP-DEV] 5.0.5RC1 rolled

2005-08-28 Thread Wez Furlong
Fixed in CVS. (just noticed that I could have just merged the config.m4 back from head rather than re-investigating it, oh well.) --Wez. On 8/28/05, Wez Furlong <[EMAIL PROTECTED]> wrote: > Thanks for checking; I'm working on a fix now. > > --Wez. > > On 8/26/05, Mike Robinson <[EMAIL PROTECTE

Re: [PHP-DEV] 5.0.5RC1 rolled

2005-08-28 Thread Wez Furlong
Thanks for checking; I'm working on a fix now. --Wez. On 8/26/05, Mike Robinson <[EMAIL PROTECTED]> wrote: > On my woody box, configure finished fine but with... > > Generating files > updating cache ./config.cache > creating ./config.status > creating php5.spec > creating main/build-defs.h > cr

RE: [PHP-DEV] 5.0.5RC1 rolled

2005-08-26 Thread Mike Robinson
On my woody box, configure finished fine but with... Generating files updating cache ./config.cache creating ./config.status creating php5.spec creating main/build-defs.h creating scripts/phpize creating scripts/man1/phpize.1 sed: can't read ./scripts/man1/phpize.1.in: No such file or directory cr

Re: [PHP-DEV] 5.0.5RC1 rolled

2005-08-22 Thread Wez Furlong
I think I saw a configure related problem with the sqlite extension recently. I won't be able to check on this until thursday, maybe wednesday if things go well during the week. If someone else has time before I do, can they try building sqlite as a shared extension to verify if there was a proble

[PHP-DEV] 5.0.5RC1 rolled

2005-08-22 Thread Zeev Suraski
www.php.net/~zeev/php-5.0.5RC1.tar.gz www.php.net/~zeev/php-5.0.5RC1.tar.bz2 If there are no stoppers I'd like to release it towards the end of the week. Zeev -- PHP Internals - PHP Runtime Development Mailing List To unsubscribe, visit: http://www.php.net/unsub.php