On Wed, Jul 9, 2008 at 1:00 PM, Steve Comstock [EMAIL PROTECTED] wrote:
Hmmm. Seems to me the configure script ought to do the check
for an existing file,
that's in the patch, but it happens at buildconf time fwiw.
I picked up autoconf-2.62 from a GNU mirror, but I was reluctant
to use it;
On Thu, Jul 3, 2008 at 1:41 PM, Steve Comstock [EMAIL PROTECTED] wrote:
ahem...remember that I said 2.2 was a little more complicated than 2.0 on z/OS?
python? I need to install python?
yes and no. In httpd-2.2, python is used to create make dependency
files srclib/apr[-util]/buildoutputs.mk
On Sun, Jun 29, 2008 at 12:01 PM, Steve Comstock
[EMAIL PROTECTED] wrote:
Isn't the correct solution to fix autoconf so it's
aware of z/OS? It doesn't seem right to have multiple
versions of a single function, but rather to have
one function that supports multiple platforms.
GNU autoconf for
On Sun, Jun 29, 2008 at 12:39 PM, Steve Comstock
[EMAIL PROTECTED] wrote:
Would it be best to configure and make apr and apr-util
then use httpd configure with --with-apr=/usr/lpp/zApache ?
I almost always configure and make them together on z/OS. I'm not
aware of any reason why building them
On Wed, Jun 25, 2008 at 1:52 PM, Steve Comstock
[EMAIL PROTECTED] wrote:
and what do you see if you do
ls srclib/apr/passwd/
-rw-r--r-- 1 ADMINS SYS16714 Jun 14 14:36 apr_getpass.c
that means that it didn't get compiled. after make, mine looks like:
$ ls -l srclib/apr/passwd/
On Wed, Jun 25, 2008 at 1:52 PM, Steve Comstock
[EMAIL PROTECTED] wrote:
Greg Ames wrote:
Where did you get your build tools, especially libtool?
OK. I started from this page:
http://people.apache.org/~trawick/apache-2-on-zos.html
which is, admittedly, a little long in the tooth
On Thu, Jun 26, 2008 at 12:45 AM, William A. Rowe, Jr.
[EMAIL PROTECTED] wrote:
Steve Comstock wrote:
And I am told pretty clearly that I will need libtool
and autoconf. Now please remember I am a stranger in
a strange land. There is no definition nor link to any
place to get these tools.
On Fri, Jun 27, 2008 at 10:17 AM, Greg Ames [EMAIL PROTECTED] wrote:
In the mean time, why don't you download the latest httpd 2.0.x
tarball and see if you can build that starting with ./buildconf? That
should get you familiar with the process and check out the toolchain.
Once we get you
On Tue, Jun 24, 2008 at 9:03 PM, Steve Comstock
[EMAIL PROTECTED] wrote:
but wait: I do the gzip to unzip, but do I need to
use pax, in order to convert to EBCDIC, or can I
just use tar?
yes, use pax. I've used this
http://people.apache.org/~gregames/xtar
...script for unpacking and
On Tue, Jun 24, 2008 at 3:56 PM, Steve Comstock
[EMAIL PROTECTED] wrote:
Making all in srclib
Making all in apr
..
WARNING CCN3068 ./dso/unix/dso.c:139 Operation between types void* and
int is not allowed.
WARNING CCN3068 ./dso/unix/dso.c:153 Operation between types const char*
and int
Bojan and Joe,
thanks for the tips. the bugzilla issue below looks like the same basic
problem. I'll read up on the gcc option and see if xlc has something similar.
Greg
- Original Message
From: Bojan Smojver [EMAIL PROTECTED]
To: Joe Orton [EMAIL PROTECTED]
Cc: dev
while testing our recent httpd metadata patches to mod_deflate.c, we
encountered an infinite loop in the new code on AIX. the loop with the
changed code amounts to:
while (!APR_BRIGADE_EMPTY(bb)) {
{
e = APR_BRIGADE_FIRST(bb);
...
if (blah) {
APR_BUCKET_REMOVE(e);
Sander Temme wrote:
On Jan 26, 2007, at 12:04 PM, [EMAIL PROTECTED] wrote:
epoll/Linux only for now. ./buildconf on FreeBSD (minotaur) spewed
tons of
errors; hints appreciated
How do you run ./buildconf? Minotaur has autoconf from ports, and it
has alternative names. Try:
From: Joe Orton [EMAIL PROTECTED]
apr_pollset_remove: speed up the httpd Event MPM when lots of connections
are
in use
I'm getting segfaults in the test suite on ppc (RHEL4) and x86_64 (FC6)
after this change. Reverting the patch fixes the issue.
sorry! reverted in svn. I'll
the current apr_pollset_remove has an O(n**2) scaling issue which is likely to degrade the
perfomance of the httpd Event MPM when it is handling lots of connections. I've been
playing with the following patch off and on but never posted it before due to day job
pressures. it serves pages for
Trevor Hurst wrote:
Hello,
I am in need of building the apr-util source because a
project of mine requires that I build auth-ldap into
my Apache 2.0.44 server.
Creating configure ...
Then when I run ./configure it cannot find the install-sh:
eskimo [361] ./configure
configure: error: cannot find
William A. Rowe, Jr. wrote:
I suggest the following as a friendly compromise to avoid voting...
this sounds like progress.
introduce DISABLE, don't drop ENABLE (especially since your patch would *reverse*
the meaning of the EnableSendfile directive for httpd 2.0.44 built against
a newer apr!!!)
William A. Rowe, Jr. wrote:
I'm very close to an outright veto of such a change, at this time...
forget it.
Not only is the binary interface to sendfile broken now, the source interface is
broken too. Since there was no mmn bump, there is no reasonable way a module
which ships independently
Between 2.0.43 and 2.0.44, the performance of SPECweb99 standard dynamic GETs
has degraded about 7%. Static files are fine, and ab benchmarks of small
dynamic requests are also fine. oprofile + strace shows that the difference
with the SPEC dynamic requests is that sendfile is no longer used
Jeff Trawick wrote:
On Unix, some failures of apr_proc_create() are not noticed in the
calling process and so apr_proc_create() returns APR_SUCCESS even though
it failed.
++1 in concept. It has bugged me for a long time that httpd 2.0 only logs
Premature end of script headers no matter what is
ooops...meant to send this to the list
Greg
Original Message
Subject: Re: cvs commit: apr/include apr_hash.h
At 09:45 AM 1/14/2003, Greg Ames wrote:
William A. Rowe, Jr. wrote:
How do you mean? Are unicies 64 LP architectures?
pls translate for us dummies ^^
long
Brian Pane wrote:
If we could change the apr_atomic_inc/dec functions to use
apr_uint32_t,
this part of the fdqueue code could become a lot simpler.
As far as I know, we can make it work with apr_uint32_t on most
platforms, as long as we declare any inline assembly blocks as
volatile (thanks to
[EMAIL PROTECTED] wrote:
brianp 2003/01/03 10:32:59
Modified:server/mpm/worker fdqueue.c
Log:
Replace most of the mutex locking in the worker MPM's queue info
object with atomic compare-and-swap loops.
A nice New Year's present :-)
Presumably this gets rid of the
have to have a single centralized location for the refcount
and you don't have lifetime issues to worry about anymore.
Greg Ames did some testing on this and says it looks good as far as he can
tell... additional eyes would be appreciated.
in case anyone cares, here's what I did:
* built httpd
[moved to [EMAIL PROTECTED]
Justin Erenkrantz wrote:
On Fri, Jun 28, 2002 at 04:40:11PM -0400, Jeff Trawick wrote:
What is the change in the FreeBSD API?
before the kernel we picked up Wednesday night: bytes to send includes
headers
after the kernel we picked up Wednesday night:
...since Friday, 28-Jun-2002 10:43:24 PDT. This build is basically 2.0.39 with
a patch to apr_sendfile to deal with a change to the FreeBSD sendfile() API. We
had 88 sendfile_it_all asserts pop today, so I decided to put the new build into
production even though there was a fair amount of
Brian Behlendorf wrote:
I'll be rebooting daedalus and icarus again soon to pick up a fix for
this.
What a week.
...and it ain't over yet :-(
Looks like we picked up a FreeBSD 4.6 kernel patch to sendfile() which changes
its API, and causes asserts in sendfile_it_all to pop frequently.
[EMAIL PROTECTED] wrote:
linux sendfile: exit with an error if the file gets smaller in the middle
of
a request.
btw, thanks Justin for analysing the dump from the new trap on daedalus.
Greg
[EMAIL PROTECTED] wrote:
FreeBSD sendfile: return an error if the kernel returns 0 no bytes sent.
This probably means the file became smaller after it was stat()ed.
While testing this on daedalus, I noticed yet another sendfile_it_all assert
dump in /tmp, dated May 10. I think I'll put a
[EMAIL PROTECTED] wrote:
jwoolley02/04/18 23:07:50
Modified:buckets apr_buckets_mmap.c
Log:
Fix the mmap cleanup problem seen on daedalus.
:-) :-)
I missed most of the discussion on this due to e-mail problems. Dang, I'm sure
glad I didn't have to debug this one.
Greg
[EMAIL PROTECTED] wrote:
gregames02/04/01 13:07:13
Modified:buildapr_hints.m4
.STATUS
Log:
FreeBSD: use fcntl cross-process locks by default. Make note of problems
experienced with flock and SysV sems.
I'll update this as appropriate after
[EMAIL PROTECTED] wrote:
-apr_atomic_t apr_atomic_add(apr_atomic_t *mem, apr_uint32_t val)
+apr_int32_t apr_atomic_add(apr_atomic_t *mem, apr_int32_t val)
Ian,
Should we change the 2nd input param to apr_int32_t everywhere? This occured to
me when I used -1 there to implement
[EMAIL PROTECTED] wrote:
gregames02/03/22 13:41:47
Modified:.configure.in
Log:
get rid of the -g compiler flag on OS/390, unless this is a debug or
maintainer mode build.
Does anyone have a problem with doing the same on all platforms? I will commit
such a fix
Bill Stoddard wrote:
(moved from [EMAIL PROTECTED] to [EMAIL PROTECTED])
I hesitate to commit the because I am not sure if apr_atomic_dec will be
portable and
usable on enough OS/hardware combinations.
as we discussed offline, you're right. The return value for apr_atomic_dec
isn't
Jeff Trawick wrote:
[EMAIL PROTECTED] writes:
gregames02/02/15 14:36:40
Modified:buildapr_hints.m4
.configure.in
Log:
provide APR_CC_HINTS. This allows us to specify a default compiler
which the user can override externally (i.e., the
[EMAIL PROTECTED] wrote:
And, you thought there wasn't going to be anything happening with this
item (FreeBSD and threading support)? Ha!
+Status: Justin has tracked down a FreeBSD kernel hacker
+([EMAIL PROTECTED]) and we've isolated some issues
+
Jeff Trawick wrote:
I'm happy with the APR-ized notion of how-did-the-process-exit and
what-is-the-signal-or-exit-code but throwing away the native status is
a real problem.
[...]
missing from patch:
doc in header files, other mpms, apr/threadproc/foo, where foo !=
unix, testing
Victor J. Orlikowski wrote:
Hi all,
Was running some performance tests on AIX...
And oddities popped up between 1.3 and 2.0 (latest CVS of both).
Requests/sec.
no keepalive keepalive
2.0 - prefork 420 590
2.0 - threaded
I assume that this is just a typo. However, I have no way to test it.
But it seems it wasn't tested as well as you might hope before it was
committed.
Greg
---
Index: srclib/apr/include//apr_portable.h
[EMAIL PROTECTED] wrote:
trawick 01/07/25 13:59:29
Modified:network_io/unix sendrecv.c
Log:
fix some issues with apr_sendfile() for FreeBSD
1) checking when to call wait_for_io_or_timeout()
it checked errno without checking rv
it required that we already sent
[EMAIL PROTECTED] wrote:
BTW, the threaded MPM isn't starting up enough child
processes (only 1). StartServers is 3 - is the count of 3 supposed to
include the parent and the unixd child?
no, StartServers and the other process related directives don't include
the parent or our
[EMAIL PROTECTED] wrote:
dreid 01/07/11 07:20:31
Modified:.configure.in
memory/unix apr_sms.c sms_private.h
include apr_pools.h apr_sms.h
Whoops! seems to have broken the build:
/bin/sh /home/gregames/apache/httpd-2.0/srclib/apr/libtool
. Ryan put in an
assert that caught one of my bugs, and vice versa.
I think when Jeff and Greg Ames were debugging www.apache.org, they used
some asserts in httpd in a positive manner.
well, thanks, but most of those are just traps looking for a
particular symptom. We throw them away once
[EMAIL PROTECTED] wrote:
Add a nifty memory/pool debugging option: allocation everything on a page,
then turn the page inaccessible at free time. Causes segfaults on access.
way cool! Thanks.
Greg
Cliff Woolley wrote:
-Original Message-
Ahem. Please don't rush. mass-cleanup can lead to beta delays,
without precautions.
Clearly. That's why I wasn't sure Ryan was even talking about a mass
cleanup when he asked for a full patch.
On the other hand, the nice thing about
45 matches
Mail list logo