I was going through failing tests today and noticed some which failed
because of memory leaks using code like the following:
RETURN_RT_STRING(s, 0);
}
For non-unicode mode this is functionally identical to:
RETURN_STRING(s, 0);
}
However for unicode mode, this binary value is translated to
Just for clarity's sake, I'm officially agreeing to this and I'm
fine with the choice of Pierre as maintainer. Pierre and whoever else,
let me know when and where to deliver my current work in progress,
and where to forward the patches I've already received. When there's
a 2.0.35 release to be mad
Hi!
Without much ado, here is a piece of problematic code:
Please note that the only difference between f1 and f2 that the first
records the return value of f3() into a variable. However, this changes
the output:
array(1) {
["foo"]=>
&string(3) "bar"
}
array(1) {
["foo"]=>
string
The i variable inside insertionsort() is used by the swap() macro and is
actually needed. The other two issues are indeed valid and patches to
fix them have just been applied to CVS, thanks for pointing them out.
Ilia
Ron Korving wrote:
Nice to see RC2 :)
I noticed that few of my earlier obse
I will be contributing improvements to the php.net website. [EMAIL PROTECTED]
asked me to request CVS access so I can directly contribute.
--
PHP Internals - PHP Runtime Development Mailing List
To unsubscribe, visit: http://www.php.net/unsub.php
Hi
currently, iconv detection doesnt work on OS X (with 5_1), as it looks
for iconv.so instead of iconv.dylib.
The patch by andrei from last week broke that.
http://cvs.php.net/viewcvs.cgi/php-src/acinclude.m4?r1=1.332.2.10&r2=1.332.2.11&pathrev=PHP_5_1
reverting it does help, but AFAICT and
