On 22/07/2014 15:02, dev wrote:
I hope this is the right maillist for this sort of question.
not really, the users list would be more appropriate
and then build httpd after checking a few basic dependencies. Must I
unpack
the apr and apr-util sources into ./srclib and use
Hello,
Sorry for the delay,
Here is the version of apache and mod_wsgi:
Apache/2.2.14 (Ubuntu)
mod_wsgi/2.8
Other packages installed : mod_fastcgi/2.4.6 mod_jk/1.2.28
PHP/5.3.2-1ubuntu4.24 with Suhosin-Patch mod_ssl/2.2.14 OpenSSL/0.9.8k
Python/2.6.5
Otherwise , I'll create a new case in mod_wsgi
On Saturday 19 July 2014 20:04:09, Christophe JAILLET wrote:
Using the following regex:
ap_log_.?error.*(_ERR|_EMERG|_CRIT)[^A]*$
many places with missing APLOGNO can be found.
There are some false positives because the [^A]* at the end of the
regex is here to check, in a more or less
Another tidbit of information ...
I attempted to see what was going on by running httpd in the foreground with
strace. When I did that, the tests worked perfectly. I guess this suggests it
is some sort of timing issue. Running httpd in the foreground without trace
exhibited the problem.
I got it to break with these python scripts, as well as with my own test
client and server (browser javascript and node.js). Strange though, I can't
reproduce it with only 2 clients on either of those programs; I have to get
3 clients connected at once. Past that, I see exactly the behavior you
Hi,
shouldn't the #error just a few lines below be updated as well, to be
more explicit than too old ?
CJ
Le 22/07/2014 21:29, rj...@apache.org a écrit :
Author: rjung
Date: Tue Jul 22 19:29:08 2014
New Revision: 1612653
URL: http://svn.apache.org/r1612653
Log:
Clarify comment.
Modified:
On 22.07.2014 22:20, Christophe JAILLET wrote:
Hi,
shouldn't the #error just a few lines below be updated as well, to be
more explicit than too old ?
You are right. But what about instead changing the configure pcre
version test:
Index: configure.in
+1
Le 22/07/2014 23:01, Rainer Jung a écrit :
On 22.07.2014 22:20, Christophe JAILLET wrote:
Hi,
shouldn't the #error just a few lines below be updated as well, to be
more explicit than too old ?
You are right. But what about instead changing the configure pcre
version test:
Index:
Hi Rainer,
On 22.07.2014 23:01, Rainer Jung wrote:
documenting the requirement PCRE = 6.7 and dropping the check (and
error message) for PCRE_DUPNAMES from server/util_pcre.c.
-1.
Please think of non-configure builds;
it doesnt hurt if the code errors out when the requirements do not met.
Gün.