Magnus Määttä wrote:
I get _alot_ of memory leaks.. My php_test_results file is 132KB large.
I have uploaded the output of valgrind for a simple phpinfo() script to
http://www.sebastian-bergmann.de/valgrind.txt
HTH,
Sebastian
--
Sebastian Bergmann
http://sebastian-bergmann.de/
On Sun, Feb 09, 2003 at 08:05:14PM -, Ilia Alshanetsky wrote :
iliaa Sun Feb 9 15:05:14 2003 EDT
Modified files:
/php4/ext/standardfile.c
Log:
Added feature request #14097 (option allowing file() command not to include
line endings in it's
On Mon, 10 Feb 2003, Stephen Thorne wrote:
Just a thought - have you tried passing $hnd by reference?
array($hnd, open_session)
no, but my example is get from session/tests/003.phpt
So if the original example/test is wrong...
Otherwise when you try and set state in the class for it to be
On Mon, 10 Feb 2003, Markus Fischer wrote:
On Sun, Feb 09, 2003 at 08:05:14PM -, Ilia Alshanetsky wrote :
iliaa Sun Feb 9 15:05:14 2003 EDT
Modified files:
/php4/ext/standard file.c
Log:
Added feature request #14097 (option allowing
Hi Shane,
Looks good to me. I will test it later this week, but it doesn't look
like there will be any problems.
--Wez.
On Sun, 9 Feb 2003, Shane Caraveo wrote:
The attached patch adds 2 optional parameters to proc_open(). The
first, cwd, and the second is an array for executing with a
Translating the documentation,
please review my application(cysoft)
--
PHP Development Mailing List http://www.php.net/
To unsubscribe, visit: http://www.php.net/unsub.php
beside that there's no need for escaping the double quotes in his example,
shouldn't the transsid parser recognize them ? i mean i can come up with other
examples where escaping them is necessary.
http://bugs.php.net/?id=22111edit=1
waiting for any rfcs or isos.
regards,
Harald Radi
--
nme - we
On February 10, 2003 03:34 am, you wrote:
On Sun, Feb 09, 2003 at 08:05:14PM -, Ilia Alshanetsky wrote :
iliaa Sun Feb 9 15:05:14 2003 EDT
Modified files:
/php4/ext/standard file.c
Log:
Added feature request #14097 (option allowing file() command not
On Mon, 10 Feb 2003, Ilia A. wrote:
We could turn the existing one into a flags parameter:
1 - use include path
2 - include new line
4 - skip_blank_lines
This would not break BC and we don't need multiple optional
parameters.
Do we want to create
According to a manual I found on the web for call's to SQLPrepare
(referenced in ext/odbc/php_odbc.c) you can implant ? in place of
variables for statement execution. PearDB makes reference to this
on:
http://pear.php.net/manual/en/core.db.tut_execute.php
So my question is, is PHP's ODBC
In the spirit of code rather than discussions that lead nowhere, here is
a contribution that removes the filter code from PHP's sapi layer for
Apache httpd-2.0. Until Zend can cleanly support streamy input, PHP
should probably just use this method. Of course, this will not solve
the threadsafety
On Mon, 10 Feb 2003, Justin Erenkrantz wrote:
In the spirit of code rather than discussions that lead nowhere, here is
a contribution that removes the filter code from PHP's sapi layer for
Apache httpd-2.0. Until Zend can cleanly support streamy input, PHP
should probably just use this
Attached is the proposed solution.
Ilia
Index: ext/standard/file.c
===
RCS file: /repository/php4/ext/standard/file.c,v
retrieving revision 1.299
diff -u -3 -p -r1.299 file.c
--- ext/standard/file.c 9 Feb 2003 20:43:05 -
Hey Ilia,
Lets also have either an enum or some real #define'd constants for the
values used in the C code.
eg:
REGISTER_LONG_CONSTANT() and the code that checks for a number should
both be using a symbolic constant rather than a hard-coded number.
--Wez.
On Mon, 10 Feb 2003, Ilia A. wrote:
The patch appears to work correctly with only one 'interesting' bug. When the
virtual() function is used to include a php file or php files are included
via the use of mod_include, random binary data is dumped on screen. This is
particular to PHP files, as virtual() function can successfully
Here is the final revision of the patch.
Ilia
Index: ext/standard/file.c
===
RCS file: /repository/php4/ext/standard/file.c,v
retrieving revision 1.300
diff -u -3 -p -r1.300 file.c
--- ext/standard/file.c 9 Feb 2003 23:11:23 -
--On Monday, February 10, 2003 3:44 PM -0500 Ilia A.
[EMAIL PROTECTED] wrote:
The patch appears to work correctly with only one 'interesting'
bug. When the virtual() function is used to include a php file or
php files are included via the use of mod_include, random binary
data is dumped on
Hi,
the current implementation of uniqid set the more entropy default
true for CYGWIN and false for the rest. CYGWIN must use more
entropy because it does not produce new values after usleep(1)
necessarily. However usleep(1) should nowadays be very slow
compared to whatever php_combined_lcg()
On Mon, 10 Feb 2003, Marcus Börger wrote:
Hi,
the current implementation of uniqid set the more entropy default
true for CYGWIN and false for the rest. CYGWIN must use more
entropy because it does not produce new values after usleep(1)
necessarily. However usleep(1) should nowadays be very
At 23:08 10.02.2003, Sascha Schumann wrote:
No, some users might depend on the return format (think of
database entries).
That was one of the reasons i only changed it for CYGWIN.
--
PHP Development Mailing List http://www.php.net/
To unsubscribe, visit: http://www.php.net/unsub.php
Well, to me, calling the code that flushes the headers and the output
buffers twice doesn't kill us. Unless someone can come up with a better way
to not call these two functions in main/main.c::php_request_shutdown, the
patch suffices for me. MOD_PHP4_H is undefined in main.c, so it's not
If you need a define for it, we can add one..?
--Jani
On Mon, 10 Feb 2003, Joseph Tate wrote:
Well, to me, calling the code that flushes the headers and the output
buffers twice doesn't kill us. Unless someone can come up with a better way
to not call these two functions in
The old Apache seems to puke on that stuff you added
some time ago and now merged to the PHP_4_3 branch.
Apache version: IBM HTTP Server 1.3.19.3 (Apache 1.3.20)
It propably needs some #ifdefs around it?
--Jani
-- Forwarded message --
Date: 11 Feb
On Tue, 11 Feb 2003, Jani Taskinen wrote:
The old Apache seems to puke on that stuff you added
some time ago and now merged to the PHP_4_3 branch.
Apache version: IBM HTTP Server 1.3.19.3 (Apache 1.3.20)
It propably needs some #ifdefs around it?
Thanks for the heads
On Tue, 11 Feb 2003, Jani Taskinen wrote:
The old Apache seems to puke on that stuff you added
some time ago and now merged to the PHP_4_3 branch.
Apache version: IBM HTTP Server 1.3.19.3 (Apache 1.3.20)
For anyone who cares, these packages can be downloaded for
free
This is really a support question and is being asked on the wrong list.
But in light of that... PHP's odbc_prepare function should work with
the '?' option. It really isn't the one deciding this though, as it's
more your ODBC Driver. As far as PEAR is concerned, I have little
insight into
Im currently trying to secure a php program where I authenticate a user
against an md5 string stored in a mysql database..
My problem is that since this program consists out of more that one pages, I
am using a session and I register the md5 string and username as session
variables.
As I
Can anyone point me to a possible solution for this?
1. Use SSL.
2. Throw away an existing session id, if a user authenticated
successfully (e.g. destroy the old session, and copy the
data into a new one).
3. Provide a logout button which destroys the session.
-
On Mon, 10 Feb 2003 15:44:22 -0500, Ilia A. wrote:
The patch appears to work correctly with only one 'interesting' bug. When
the virtual() function is used to include a php file or php files are
included via the use of mod_include, random binary data is dumped on
screen. This is particular to
Justin Erenkrantz wrote:
Until Zend can cleanly support streamy input, PHP should probably just
use this method.
I was under the impression it could, at least someone told me so (Zeev,
IIRC, at a Conference last year).
If not, this would surely be a nice feature for the Zend Engine 2.
On Mon, Feb 10, 2003 at 09:10:44PM -0800, Ian Holsman wrote:
On Mon, 10 Feb 2003 15:44:22 -0500, Ilia A. wrote:
The patch appears to work correctly with only one 'interesting' bug. When
the virtual() function is used to include a php file or php files are
included via the use of
31 matches
Mail list logo