You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@apr.apache.org by Paul Querna <ch...@force-elite.com> on 2005/07/19 15:02:32 UTC
[VOTE] APR 1.2.0
-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1
Good Afternoon,
APR 1.2.0 is ready for testing and voting for release.
Download in your favorite archive format at:
http://people.apache.org/~pquerna/dev/apr-1.2.0/
apr-1.2.0.tar.bz2: FE 46 07 E9 28 77 4E 80 10 6F 71 19 26 18 8A 5D
apr-1.2.0.tar.gz: 56 7A 8F 00 75 BD DD 00 60 A4 F0 AD DC D8 BF DA
apr-1.2.0.zip: 96 00 50 6F 5C EC 0F 2B 59 ED 9A ED 39 A2 80 B5
There is no corresponding APR-Util tag or release at this time.
APR-Util will get their own release when they are ready.
- -Paul
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.2.4 (Darwin)
iD8DBQFC3Pno94h19kJyHwARAk0sAKCORDRBJfkfbnMUvdKaL7g1n1/YRQCgoj8n
ITR9cu8B3PJaaAlnbQe6gdo=
=9Hun
-----END PGP SIGNATURE-----
Re: [VOTE] APR 1.2.0
Posted by Graham Leggett <mi...@sharp.fm>.
Paul Querna said:
> APR 1.2.0 is ready for testing and voting for release.
+1 on Solaris v2.8 (normal and Solaris PKG), MacOSX 10.3, and Fedora Core
3 (normal and RPM).
Regards,
Graham
--
Re: [VOTE] APR 1.2.0
Posted by Joe Orton <jo...@redhat.com>.
On Tue, Jul 19, 2005 at 03:02:32PM +0200, Paul Querna wrote:
> Good Afternoon,
>
> APR 1.2.0 is ready for testing and voting for release.
+1 for release, testall runs in both apr and apr-util using that tarball
with the apr-util 1.1.2 release tarball on:
PASS: RHEL4/i686 RHEL3/i686 FC3/i686 RHEL4/ppc RawHide/ppc RHEL4/x86_64 FC4/i686
(and agreed it doesn't matter at all that apr/apr-util version numbers
get out of synch)
joe
Re: [VOTE] APR 1.2.0
Posted by Mladen Turk <mt...@apache.org>.
Paul Querna wrote:
> -----BEGIN PGP SIGNED MESSAGE-----
> Hash: SHA1
>
> Good Afternoon,
>
> APR 1.2.0 is ready for testing and voting for release.
>
+1.
Tested on WINXP/i386, WINXP/amd64, SLES9/amd64
Regards,
Mladen.
Re: [VOTE] APR 1.2.0
Posted by Sander Temme <sa...@temme.net>.
-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1
On Jul 19, 2005, at 3:02 PM, Paul Querna wrote:
> APR 1.2.0 is ready for testing and voting for release.
>
I'm facing problems with the test suite on both FreeBSD 5.2.1 and
Darwin 8.2.0.
The Darwin output:
[sctemme@MonaLisa] apr-1.2.0 $ make test
(cd test && make check)
for prog in testlockperf testshmproducer testshmconsumer
testmutexscope testall ; do \
./$prog; \
if test $? = 255; then \
echo "$prog failed"; \
break; \
fi; \
done
dyld: Library not loaded: /usr/local/apr/lib/libapr-1.0.dylib
Referenced from: /Volumes/Files/projects/asf/apr-1.2.0/test/./
testlockperf
Reason: image not found
dyld: Library not loaded: /usr/local/apr/lib/libapr-1.0.dylib
Referenced from: /Volumes/Files/projects/asf/apr-1.2.0/test/./
testshmproducer
Reason: image not found
dyld: Library not loaded: /usr/local/apr/lib/libapr-1.0.dylib
Referenced from: /Volumes/Files/projects/asf/apr-1.2.0/test/./
testshmconsumer
Reason: image not found
dyld: Library not loaded: /usr/local/apr/lib/libapr-1.0.dylib
Referenced from: /Volumes/Files/projects/asf/apr-1.2.0/test/./
testmutexscope
Reason: image not found
dyld: Library not loaded: /usr/local/apr/lib/libapr-1.0.dylib
Referenced from: /Volumes/Files/projects/asf/apr-1.2.0/test/./testall
Reason: image not found
I assume it tries to link against an installed libapr, and I have not
made install (and IMHO shouldn't have to).
FreeBSD:
[sctemme@bagheera] apr-1.2.0 $ make test
(cd test && make check)
for prog in testlockperf testshmproducer testshmconsumer
testmutexscope testall ; do ./$prog; if test $? = 255; then echo
"$prog failed"; break; fi; done
This program won't work on this platform because there is no support
for threads.
*** Error code 254
Stop in /local0/asf/apr-1.2.0/test.
*** Error code 1
Stop in /local0/asf/apr-1.2.0.
The configure script (correctly) detects that this build should not
be threaded, but apparently the test suite does not pick up that clue
phone.
S.
- --
sander@temme.net http://www.temme.net/sander/
PGP FP: 51B4 8727 466A 0BC3 69F4 B7B8 B2BE BC40 1529 24AF
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.1 (Darwin)
iD8DBQFC3QaDnjkrwtLH+RIRAqzzAJ9R6P3k80+Jb/hVraQZupuVH92rHACeIj6L
DFX2BVaYPEKNwyAQYEyw6+0=
=DiJr
-----END PGP SIGNATURE-----
Re: [VOTE] APR 1.2.0
Posted by Henry Jen <he...@ztune.net>.
Paul Querna wrote:
> Henry Jen wrote:
>
>>
>>-1 for Win32, the condvars deadlock is a serious bug. I knew this is not
>>news, but as the patch had been available for quite a while, is it
>>possible to get it fixed?
>
>
> No.
>
> I will not commit such a platform specific patch. Anyone who actually
> compiles APR on win32 that wants to do it?
>
I tried to convince our team members to use/contribute to apr as much as
we can, the recent apr_dbd_sqlite3 driver by Rick Keiner is one example.
Cannot make progress on such a serious bug will make me much harder to
pitch the idea of a broader collaboration between jxta-c and apr project.
Please someone looks into the patch.
Thanks,
Henry
Re: [VOTE] APR 1.2.0
Posted by E Holyat <eh...@yahoo.com>.
I did some testing without the extra mutex, and it looks good to go. Due to the UNIX documentation for the pthread signalling(that Henry pointed out), the behavior is the same.
"William A. Rowe, Jr." <wr...@rowe-clan.net> wrote:At 04:56 PM 7/20/2005, Henry Jen wrote:
>Good catch, you also need to reset cond->signalled.
>I adapted your changes on the "else if" also the static INLINE internal function and created a new patch.
This is definitely coolness; do you two agree this is the complete
patch for locks/win32/thread_cond.c? If so, I'll commit this flavor
in a couple hours after you both look it over one last time.
I'm afraid, never having used condvars, that I'm not 100% certain
about the expected behavior, so I'm not the best reviewer. Thanks
to you both for spending cycles examining this code.
FYI - in the future be careful to tag [PATCH apr-1.2] or something
in the subject so that patches aren't lost in the noise of a longer
conversation.
Bill
---------------------------------
Start your day with Yahoo! - make it your home page
Re: [VOTE] APR 1.2.0
Posted by "William A. Rowe, Jr." <wr...@rowe-clan.net>.
At 04:56 PM 7/20/2005, Henry Jen wrote:
>Good catch, you also need to reset cond->signalled.
>I adapted your changes on the "else if" also the static INLINE internal function and created a new patch.
This is definitely coolness; do you two agree this is the complete
patch for locks/win32/thread_cond.c? If so, I'll commit this flavor
in a couple hours after you both look it over one last time.
I'm afraid, never having used condvars, that I'm not 100% certain
about the expected behavior, so I'm not the best reviewer. Thanks
to you both for spending cycles examining this code.
FYI - in the future be careful to tag [PATCH apr-1.2] or something
in the subject so that patches aren't lost in the noise of a longer
conversation.
Bill
Re: [VOTE] APR 1.2.0
Posted by Henry Jen <he...@ztune.net>.
E Holyat wrote:
> I don't use the broadcast functionality, but, it looks like that is
> broken too. signal_all is never reset to 0 after the broadcast function
> sets it to 1. Here is an untested fix
>
> cond->num_waiting--;/*we have the lock(s)*/
> if (cond->signal_all) {
> if (cond->num_waiting == 0) {
> + cond->signal_all=0;
> ResetEvent(cond->event);
> }
> done=1;
> }
> else if (cond->signalled) {
> cond->signalled = 0;
> ResetEvent(cond->event);
> done=1;
> }
Good catch, you also need to reset cond->signalled.
I adapted your changes on the "else if" also the static INLINE internal
function and created a new patch.
Cheers,
Henry
Re: [VOTE] APR 1.2.0
Posted by E Holyat <eh...@yahoo.com>.
I don't use the broadcast functionality, but, it looks like that is broken too. signal_all is never reset to 0 after the broadcast function sets it to 1. Here is an untested fix
cond->num_waiting--;/*we have the lock(s)*/
if (cond->signal_all) {
if (cond->num_waiting == 0) {
+ cond->signal_all=0;
ResetEvent(cond->event);
}
done=1;
}
else if (cond->signalled) {
cond->signalled = 0;
ResetEvent(cond->event);
done=1;
}
Bill Stoddard <bi...@wstoddard.com> wrote:
E Holyat wrote:
> Here is a patch for win32 that has been tested extensively for a few
> months now. I submitted it to bugzilla
>
Based on a quick code review, I'm +1 for committing this patch.
Bill
---------------------------------
Yahoo! Mail
Stay connected, organized, and protected. Take the tour
Re: [VOTE] APR 1.2.0
Posted by Bill Stoddard <bi...@wstoddard.com>.
E Holyat wrote:
> Here is a patch for win32 that has been tested extensively for a few
> months now. I submitted it to bugzilla
>
Based on a quick code review, I'm +1 for committing this patch.
Bill
Re: [VOTE] APR 1.2.0
Posted by Henry Jen <he...@ztune.net>.
E Holyat wrote:
> Here is a patch for win32 that has been tested extensively for a few
> months now. I submitted it to bugzilla
>
> The previous patch addressed only the unlock being called more than
> once.
>
> This attachment avoids race conditions that the previous patch
> doesn't. This patch also fixes the multiple calls to unlock. This
> patch also consolidates the the duplicate efforts in
> apr_thread_cond_wait and apr_thread_cond_timedwait
Seems like there are couple patches around, here is the one I submitted
on Jun 02, this patch is used by jxta-c project.
> Hi,
>
> We encountered the issue as reported in Bug
> 27654(http://issues.apache.org/bugzilla/show_bug.cgi?id=27654), too
> bad I did not check the issue database earlier until I figured it out
> after 3 days.
>
> The patch in the issues database should be working, is there a reason
> for not committing it?
>
> Anyway, I have created another patch(see attachment) to address the
> same issue and which would also fix bug 34336 as it totally removing
> the cond->mutex. I believe using the mutex passed in to ensure mutual
> access should be sufficient.
>
> Unless the different threads can use different mutex for the same
> cond, but I don't see that as a valid usage and nor can I think of a
> use case for that.
>
> Please review it, comments are welcome. I would like to see the patch
> be committed to head and also back port to 0.9.x release. The patch
> had been tested for both apr-0.9.6 and SVN trunk with 'patch -Np0' in
> the apr folder.
Re: [VOTE] APR 1.2.0
Posted by E Holyat <eh...@yahoo.com>.
Here is a patch for win32 that has been tested extensively for a few months now. I submitted it to bugzilla
The previous patch addressed only the unlock being called more than once.
This attachment avoids race conditions that the previous patch doesn't. This patch also fixes the multiple calls to unlock. This patch also consolidates the the duplicate efforts in apr_thread_cond_wait and apr_thread_cond_timedwait
Cliff Woolley <jw...@virginia.edu> wrote:
On Wed, 20 Jul 2005, Paul Querna wrote:
> > -1 for Win32, the condvars deadlock is a serious bug. I knew this is not
> > news, but as the patch had been available for quite a while, is it
> > possible to get it fixed?
>
> No.
>
> I will not commit such a platform specific patch. Anyone who actually
> compiles APR on win32 that wants to do it?
I'll set up a build on my windows box at work this afternoon and check it
out if OtherBill doesn't beat me to it.
--Cliff
__________________________________________________
Do You Yahoo!?
Tired of spam? Yahoo! Mail has the best spam protection around
http://mail.yahoo.com
Re: [VOTE] APR 1.2.0
Posted by Cliff Woolley <jw...@virginia.edu>.
On Wed, 20 Jul 2005, Paul Querna wrote:
> > -1 for Win32, the condvars deadlock is a serious bug. I knew this is not
> > news, but as the patch had been available for quite a while, is it
> > possible to get it fixed?
>
> No.
>
> I will not commit such a platform specific patch. Anyone who actually
> compiles APR on win32 that wants to do it?
I'll set up a build on my windows box at work this afternoon and check it
out if OtherBill doesn't beat me to it.
--Cliff
Re: [VOTE] APR 1.2.0
Posted by Paul Querna <ch...@force-elite.com>.
Henry Jen wrote:
> Paul Querna wrote:
>
>> -----BEGIN PGP SIGNED MESSAGE-----
>> Hash: SHA1
>>
>> Good Afternoon,
>>
>> APR 1.2.0 is ready for testing and voting for release.
>>
>> Download in your favorite archive format at:
>> http://people.apache.org/~pquerna/dev/apr-1.2.0/
>>
>> apr-1.2.0.tar.bz2: FE 46 07 E9 28 77 4E 80 10 6F 71 19 26 18 8A 5D
>> apr-1.2.0.tar.gz: 56 7A 8F 00 75 BD DD 00 60 A4 F0 AD DC D8 BF DA
>> apr-1.2.0.zip: 96 00 50 6F 5C EC 0F 2B 59 ED 9A ED 39 A2 80 B5
>>
>>
>> There is no corresponding APR-Util tag or release at this time.
>> APR-Util will get their own release when they are ready.
>>
>> - -Paul
>
>
> -1 for Win32, the condvars deadlock is a serious bug. I knew this is not
> news, but as the patch had been available for quite a while, is it
> possible to get it fixed?
No.
I will not commit such a platform specific patch. Anyone who actually
compiles APR on win32 that wants to do it?
Re: [VOTE] APR 1.2.0
Posted by Henry Jen <he...@ztune.net>.
Paul Querna wrote:
> -----BEGIN PGP SIGNED MESSAGE-----
> Hash: SHA1
>
> Good Afternoon,
>
> APR 1.2.0 is ready for testing and voting for release.
>
> Download in your favorite archive format at:
> http://people.apache.org/~pquerna/dev/apr-1.2.0/
>
> apr-1.2.0.tar.bz2: FE 46 07 E9 28 77 4E 80 10 6F 71 19 26 18 8A 5D
> apr-1.2.0.tar.gz: 56 7A 8F 00 75 BD DD 00 60 A4 F0 AD DC D8 BF DA
> apr-1.2.0.zip: 96 00 50 6F 5C EC 0F 2B 59 ED 9A ED 39 A2 80 B5
>
>
> There is no corresponding APR-Util tag or release at this time.
> APR-Util will get their own release when they are ready.
>
> - -Paul
-1 for Win32, the condvars deadlock is a serious bug. I knew this is not
news, but as the patch had been available for quite a while, is it
possible to get it fixed?
Cheers,
Henry
APR 1.2.0 build broken on NetWare (Re: [VOTE] APR 1.2.0)
Posted by Brad Nicholes <bn...@novell.com>.
I'm not sure how to fix it. There was a reason why I had to compile
the Ldap support into a separate library before linking the whole thing
together but sitting here at ApacheCon, its not coming to me right at
the moment. Any how, my concern is that even if I can fix the apr build
files to not require the prebuilt apuldap library, ldap support on
NetWare would be broken anyway due to the fact that there was a good
reason why I patched the make files to build the library in the first
place. I still think that it would be easier to release at least a
1.2.0 version of APU with the ldap patches folded in and anything else
that might be good to go.
Brad
>>> On Wednesday, July 20, 2005 at 11:54 am, in message
<42...@force-elite.com>, Paul Querna <ch...@force-elite.com>
wrote:
> Brad Nicholes wrote:
>> -1 Netware. I am having build problems because of the complicated
way
>> that apr and apr-util build together on NetWare. The problems are
>> solved if we release a new apr-util as well. I had to change the
way
>> NetWare builds the ldap portion of apr-util and without the updates
that
>> are in trunk for apr-util, the ldap library is missing when it
comes
>> time to link the whole thing together.
>>
>> Brad
>
> Dang, that sucks.
>
> Any idea how hard it is to fix? I do sort of consider this a show
> stopper to APR (and hence, for httpd).
>
> Any chance it can be fixed this week?
>
> Thanks,
>
> Paul
Re: [VOTE] APR 1.2.0
Posted by Paul Querna <ch...@force-elite.com>.
Brad Nicholes wrote:
> -1 Netware. I am having build problems because of the complicated way
> that apr and apr-util build together on NetWare. The problems are
> solved if we release a new apr-util as well. I had to change the way
> NetWare builds the ldap portion of apr-util and without the updates that
> are in trunk for apr-util, the ldap library is missing when it comes
> time to link the whole thing together.
>
> Brad
Dang, that sucks.
Any idea how hard it is to fix? I do sort of consider this a show
stopper to APR (and hence, for httpd).
Any chance it can be fixed this week?
Thanks,
Paul
Re: [VOTE] APR 1.2.0
Posted by Brad Nicholes <bn...@novell.com>.
-1 Netware. I am having build problems because of the complicated way
that apr and apr-util build together on NetWare. The problems are
solved if we release a new apr-util as well. I had to change the way
NetWare builds the ldap portion of apr-util and without the updates that
are in trunk for apr-util, the ldap library is missing when it comes
time to link the whole thing together.
Brad
>>> On Tuesday, July 19, 2005 at 7:02 am, in message
<42...@force-elite.com>, Paul Querna <ch...@force-elite.com>
wrote:
> -----BEGIN PGP SIGNED MESSAGE-----
> Hash: SHA1
>
> Good Afternoon,
>
> APR 1.2.0 is ready for testing and voting for release.
>
> Download in your favorite archive format at:
> http://people.apache.org/~pquerna/dev/apr-1.2.0/
>
> apr-1.2.0.tar.bz2: FE 46 07 E9 28 77 4E 80 10 6F 71 19 26 18 8A 5D
> apr-1.2.0.tar.gz: 56 7A 8F 00 75 BD DD 00 60 A4 F0 AD DC D8 BF DA
> apr-1.2.0.zip: 96 00 50 6F 5C EC 0F 2B 59 ED 9A ED 39 A2 80 B5
>
>
> There is no corresponding APR-Util tag or release at this time.
> APR-Util will get their own release when they are ready.
>
> - -Paul
> -----BEGIN PGP SIGNATURE-----
> Version: GnuPG v1.2.4 (Darwin)
>
> iD8DBQFC3Pno94h19kJyHwARAk0sAKCORDRBJfkfbnMUvdKaL7g1n1/YRQCgoj8n
> ITR9cu8B3PJaaAlnbQe6gdo=
> =9Hun
> -----END PGP SIGNATURE-----
Re: [VOTE] APR 1.2.0
Posted by Justin Erenkrantz <ju...@erenkrantz.com>.
On Tue, Jul 19, 2005 at 03:02:32PM +0200, Paul Querna wrote:
> -----BEGIN PGP SIGNED MESSAGE-----
> Hash: SHA1
>
> Good Afternoon,
>
> APR 1.2.0 is ready for testing and voting for release.
I've told Paul this in person; but for the list:
Everything passes except for testsock on darwin 8.2.0 (10.4.2).
It appears that we're horking on the kqueue with the non-blocking sockets and
'sockchild read' goes into an infinite loop. I'm fairly sure this is a
regression.
I'm investigating further. -- justin