Date: 12/29/98
- Next message: miri <email protected>: "[PHP-DEV] Bug #1011: Inconsistent header for apache_php3_module_main"
- Previous message: Sascha Schumann: "Re: [PHP-DEV] Another latest CVS problem"
- In reply to: Sascha Schumann: "Re: [PHP-DEV] Another latest CVS problem"
- Next in thread: Sascha Schumann: "Re: [PHP-DEV] Another latest CVS problem"
- Reply: Sascha Schumann: "Re: [PHP-DEV] Another latest CVS problem"
- Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]
On Tue, 29 Dec 1998, Sascha Schumann wrote:
> Nope - as I said before, _php3_regcomp doesn't call efree directly. I
> don't think it's a gdb problem, because the parameters to _php3_regcomp
> are not bogus, but show that an old version of functions/reg.c was used to
> compile the binary which you are now debugging.
Thanks again for the followup. After working with the existing tree
and getting nowhere, I finally decided a couple of hours ago to take a
large step back. I downloaded Apache 1.3.3, got a fresh copy of the
PHP3 current CVS tree, and started over. This is a bit more tedious
than just installing PHP3 because I have to integrate PHP2 and FP98
server extensions into the daemon also, and I was hesitant to touch
all that unless I couldn't make it work any other way. I still don't
know what was wrong with the old tree; I've upgraded this way for a
long time.
The new daemon is mostly stable on the very latest PHP3 tree. I am
still occasionally getting Document Contains No Data errors; this
might be every 20-30 page loads or so when running as a full preforked
daemon. I've not been able to find where it might be failing because
when I run under gdb with -X option, about every other page load I get
an error similar to one of the following below. PHP3 and Apache are
both compiled with -g -O flags.
Though I don't work much at this level any more, this looks pretty
flakey to me, and somehow I don't think these are causing the empty
documents (???). I have quite a list of 'directory' lines in .gdbinit
to find the source files, but I don't think this is from the source
tree. Do any of the developers see this when debugging, or have any
idea what causes it and/or how to avoid it? Thanks for any insight
you may have.
-- Michael
Starting program: /usr/local/apache/src/./httpd -X -d /usr/local/apache
-f conf/httpd.conf
Program received signal SIGSEGV, Segmentation fault.
0x401395a9 in __libc_free (mem=0x5000000)
(gdb) bt
#0 0x401395a9 in __libc_free (mem=0x5000000)
#1 0x82b6cf0 in ?? ()
#2 0x25b9 in ?? ()
Program received signal SIGSEGV, Segmentation fault.
0x401395a9 in __libc_free (mem=0x5000000)
(gdb) bt
#0 0x401395a9 in __libc_free (mem=0x5000000)
#1 0x82b6da8 in ?? ()
#2 0x24c9 in ?? ()
-- PHP Development Mailing List http://www.php.net/ To unsubscribe send an empty message to php-dev-unsubscribe <email protected> For help: php-dev-help <email protected>
- Next message: miri <email protected>: "[PHP-DEV] Bug #1011: Inconsistent header for apache_php3_module_main"
- Previous message: Sascha Schumann: "Re: [PHP-DEV] Another latest CVS problem"
- In reply to: Sascha Schumann: "Re: [PHP-DEV] Another latest CVS problem"
- Next in thread: Sascha Schumann: "Re: [PHP-DEV] Another latest CVS problem"
- Reply: Sascha Schumann: "Re: [PHP-DEV] Another latest CVS problem"
- Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]

