You are viewing a plain text version of this content. The canonical link for it is here.
Posted to apreq-dev@httpd.apache.org by Joe Schaefer <jo...@sunstarsys.com> on 2004/11/30 19:12:22 UTC

towards a 2.05 release

I've tackled the few final ABI & build system
changes that were in STATUS.  The rest look to me
like issues that can be addressed in a future release. 

I'd like to start the release candidate process soon, so 
anyone with a serious itch needs to have a look at the 
current trunk to see what can be done about it.

Yes, I'm still pressing infra@ to fix the rsync snapshots.  
For those of you who haven't installed svn, please be patient 
(or just install svn!).  Sometimes I put a tarball at
http://cvs.apache.org/~joes/ , so you can always check there 
while infra@ finds the necessary tuits.

Thanks!

-- 
Joe Schaefer


Re: towards a 2.05 release

Posted by Christian Hansen <ch...@ngmedia.com>.
Christian Hansen wrote:
> I'll try to go backwards and see if i can compile any earlier revisions.

I have tried to compile revisions 2004-11-28 and 2004-11-29 same errors.


[Tue Dec 07 07:27:00 2004] [debug] mod_apreq_big_request_test.c(48): [client 
127.0.0.1] initializing request
[Tue Dec 07 07:27:09 2004] [notice] child pid 649 exit signal Bus error (10), 
possible coredump in /Users/chansen/src/apreq/apreq/env/t

Core was generated by `/usr/local/apache2/bin/httpd'.
#0  0x0047ac0c in ?? ()
(gdb) bt
#0  0x0047ac0c in ?? ()
#1  0x006d142c in ?? ()
#2  0x0106ad24 in ?? ()
#3  0x0066cdf8 in ?? ()
#4  0x0001eeb8 in hash_mutex ()
#5  0x0001f648 in hash_mutex ()
#6  0x0000793c in vt_db_close (dbm=0x198c650) at apr_dbm_berkeleydb.c:206
#7  0x00002680 in utf7_wctomb () at apr_buckets_eos.c:38
#8  0x00018b04 in hash_mutex ()
#9  0x00008f3c in big5_wctomb () at sdbm.c:476
#10 0x000090c8 in ksc5601_mbtowc () at sdbm.c:535
#11 0x00009360 in johab_hangul_mbtowc () at sdbm_hash.c:52
#12 0x000097ec in apu__sdbm_delpair (pag=0x0, key={dptr = 0x0, dsize = 239288}) 
at sdbm_pair.c:219
#13 0x0000aefc in iso2022_jp2_mbtowc () at apr_uri.c:342

I guess next step is to compile apache in debug mode.

Regards
Christian


Re: towards a 2.05 release

Posted by Christian Hansen <ch...@ngmedia.com>.
Joe Schaefer wrote:
> Christian Hansen <ch...@ngmedia.com> writes:
> 
> 
>>Joe Schaefer wrote:
>>
>>>Which tests, please?  It is never ok for the trunk
>>>to be failing tests; just be sure the tests fail on a fresh svn checkout.
>>
>>All tests fails
> 
> 
> Sorry, I can't reproduce it.  See if you have any better
> luck with this tarball, which includes the buildconf patch:
> 
>     http://svn.apache.org/~joes/libapreq2-2.05-osx-test-2004-12-06.tar.gz
> 
> This is rolled with 
> 
>      % make release
> 
> so if there's a difference, it's probably more libtool/auto* 
> breakage on OSX (maybe gmake is required with glibtool, dunno).
> 

All tests fails with same errors as with trunk. This is very strange since the
snapshot released 2004-11-29 works.

I don't think the buidtools are broken since i have built quite a few packages,
something around 200 with the help of port (darwinports).

I'll try to go backwards and see if i can compile any earlier revisions.

I'm very grateful for the help i have received so far.


Regards
Christian


Re: towards a 2.05 release

Posted by Paul Querna <ch...@force-elite.com>.
Justin Erenkrantz wrote:
> --On Wednesday, March 9, 2005 1:17 PM -0600 "William A. Rowe, Jr." 
> <wr...@rowe-clan.net> wrote:
> 
>> Bingo.
>>
>> I see any perl-wrapper as surviving very nicely in mod_perl, given
>> that it would be reflecting functions exposed from the core distro.
>>
>> As they mention, the perl wrappers still need refreshing.
> 
> 
> AFAICT, that module requires libapreq2.  So, we'd have to bring it all 
> into our tree.  So, my -1 for trunk still holds until after 2.2.x is 
> branched off and on its way.  I just don't think it is in a state where 
> it'd be easy to bring into our tree and not have it delay the beta 
> process unduly.
> 
> IMHO, it can wait for 2.4.  -- justin
> 

I agree with Justin.  This might be a good thing for the future, but the 
timing isn't right to do it now.

I think the focus should be on getting a 2.2 beta out first, and revisit 
this once 2.2 is branched.

-Paul

Re: towards a 2.05 release

Posted by Justin Erenkrantz <ju...@erenkrantz.com>.
--On Wednesday, March 9, 2005 1:12 PM -0800 Paul Querna 
<ch...@force-elite.com> wrote:

> I don't see how an optional module that depends on an external library is
> that helpful compared to the current situation. (Module bundled with the
> Library.)
>
> It makes the most sense to me to include the library and module if we do
> this.  I don't think including both should be done at this time.

That is precisely my feelings as well.  -- justin

Re: towards a 2.05 release

Posted by Paul Querna <ch...@force-elite.com>.
Joe Schaefer wrote:
> Justin Erenkrantz <ju...@erenkrantz.com> writes:
> 
> [...]
> 
> 
>>AFAICT, that module requires libapreq2.  So, we'd have to bring it all
>>into our tree.  
> 
> 
> Why would you need to do that?  I don't know what Bill's expecting,
> but it'd be a total surprise to me if he were suggesting making
> mod_apreq2 a mandatory module for 2.2 (ie, always installed).  OTOH,
> C developers that want to use libapreq2 for their apache modules
> aren't well served unless apreq becomes a core feature of the server.
> Which might be a good idea for 2.4, or not.
> 
> So if I were the one making the proposal for 2.2 inclusion, I'd
> offer to include those 4 C files referenced earlier, and configuring
> httpd to treat them as optional module.
> 

I don't see how an optional module that depends on an external library 
is that helpful compared to the current situation. (Module bundled with 
the Library.)

It makes the most sense to me to include the library and module if we do 
this.  I don't think including both should be done at this time.

-Paul

Re: towards a 2.05 release

Posted by Joe Schaefer <jo...@sunstarsys.com>.
Justin Erenkrantz <ju...@erenkrantz.com> writes:

[...]

> AFAICT, that module requires libapreq2.  So, we'd have to bring it all
> into our tree.  

Why would you need to do that?  I don't know what Bill's expecting,
but it'd be a total surprise to me if he were suggesting making
mod_apreq2 a mandatory module for 2.2 (ie, always installed).  OTOH,
C developers that want to use libapreq2 for their apache modules
aren't well served unless apreq becomes a core feature of the server.
Which might be a good idea for 2.4, or not.

So if I were the one making the proposal for 2.2 inclusion, I'd
offer to include those 4 C files referenced earlier, and configuring
httpd to treat them as optional module.

-- 
Joe Schaefer


Re: towards a 2.05 release

Posted by Justin Erenkrantz <ju...@erenkrantz.com>.
--On Wednesday, March 9, 2005 1:17 PM -0600 "William A. Rowe, Jr." 
<wr...@rowe-clan.net> wrote:

> Bingo.
>
> I see any perl-wrapper as surviving very nicely in mod_perl, given
> that it would be reflecting functions exposed from the core distro.
>
> As they mention, the perl wrappers still need refreshing.

AFAICT, that module requires libapreq2.  So, we'd have to bring it all into 
our tree.  So, my -1 for trunk still holds until after 2.2.x is branched 
off and on its way.  I just don't think it is in a state where it'd be easy 
to bring into our tree and not have it delay the beta process unduly.

IMHO, it can wait for 2.4.  -- justin

Re: towards a 2.05 release

Posted by "William A. Rowe, Jr." <wr...@rowe-clan.net>.
At 12:07 PM 3/9/2005, Joe Schaefer wrote:
>Justin Erenkrantz <ju...@erenkrantz.com> writes:
>
>> In general, I'm not really comfortable with adding a large amount of
>> perl code to our tree.  
>
>I think Bill's assuming some familiarity with how the httpd-apreq project
>is organized.  The actual code I assume he's talkabout dropping into 
>the httpd tree is our apache2 module, which is only a couple of C files.
>If you're looking for something to review, I'd suggest
>
>  <URL: http://svn.apache.org/repos/asf/httpd/
>        apreq/branches/multi-env-unstable/module/apache2/ >

Bingo.

I see any perl-wrapper as surviving very nicely in mod_perl, given
that it would be reflecting functions exposed from the core distro.

As they mention, the perl wrappers still need refreshing.

Bill


Re: towards a 2.05 release

Posted by Joe Schaefer <jo...@sunstarsys.com>.
Justin Erenkrantz <ju...@erenkrantz.com> writes:

> --On Wednesday, March 9, 2005 10:00 AM -0600 "William A. Rowe, Jr."
> <wr...@rowe-clan.net> wrote:
>
>> With the approach of httpd 2.1-beta (in anticipation of 2.2 GA)
>> I'd like to propose the httpd project integrate apreq into the
>> core distribution.  This project has evolved considerably since
>> it was first considered.
>>
>> Comment or vote?
>
> In general, I'm not really comfortable with adding a large amount of
> perl code to our tree.  

I think Bill's assuming some familiarity with how the httpd-apreq project
is organized.  The actual code I assume he's talkabout dropping into 
the httpd tree is our apache2 module, which is only a couple of C files.
If you're looking for something to review, I'd suggest

  <URL: http://svn.apache.org/repos/asf/httpd/
        apreq/branches/multi-env-unstable/module/apache2/ >

-- 
Joe Schaefer


Re: towards a 2.05 release

Posted by Justin Erenkrantz <ju...@erenkrantz.com>.
--On Wednesday, March 9, 2005 10:00 AM -0600 "William A. Rowe, Jr." 
<wr...@rowe-clan.net> wrote:

> With the approach of httpd 2.1-beta (in anticipation of 2.2 GA)
> I'd like to propose the httpd project integrate apreq into the
> core distribution.  This project has evolved considerably since
> it was first considered.
>
> Comment or vote?

In general, I'm not really comfortable with adding a large amount of perl 
code to our tree.  httpd is a C project not a perl project - I'm not 
overjoyed at the prospect of having to maintain perl code.  I, of course, 
am from the camp that anything in our tree I am collectively responsible 
for.

Additionally, I think merging apreq should be something for 2.4 not 2.2. 
Adding it right *now* would cause a lot of delays.  For example, from a 
quick glance, they use automake - which is a big no-no.  So, it'd take a 
lot of effort to get it integrated in our build system.  I really think 
that effort would best be placed in getting a solid 2.2 release out sooner.

If there is still a lot of interest, this can be something we merge in 
right after we branch 2.2.x off into 2.3.x.  And can be an impetus for 
2.4.x's release.

So, -0 in concept.  And, -1 for doing it today - it's not in a state that 
can just slide into our tree at all without issues.  -- justin

Re: towards a 2.05 release

Posted by "William A. Rowe, Jr." <wr...@rowe-clan.net>.
With the approach of httpd 2.1-beta (in anticipation of 2.2 GA)
I'd like to propose the httpd project integrate apreq into the
core distribution.  This project has evolved considerably since
it was first considered.

Comment or vote?

Bill

At 08:49 AM 3/9/2005, Eli Marmor wrote:
>Following the major patches, API changes, modifications and improvements
>since 2.04, is it possible to release a pre-2.05 (or a candidate to
>become 2.05) so we can start testing it and find bugs or problems under
>specific platforms, environments and configurations?
>
>I also want to use this opportunity to thank Joe, Stas, Max, Randy,
>Bojan, and all the other great developers who brought libapreq2 to its
>current status.
>
>-- 
>Eli Marmor
>marmor@netmask.it
>Netmask (El-Mar) Internet Technologies Ltd.
>__________________________________________________________
>Tel.:   +972-9-766-1020          8 Yad-Harutzim St.
>Fax.:   +972-9-766-1314          P.O.B. 7004
>Mobile: +972-50-5237338          Kfar-Saba 44641, Israel


Re: towards a 2.05 release

Posted by "William A. Rowe, Jr." <wr...@rowe-clan.net>.
With the approach of httpd 2.1-beta (in anticipation of 2.2 GA)
I'd like to propose the httpd project integrate apreq into the
core distribution.  This project has evolved considerably since
it was first considered.

Comment or vote?

Bill

At 08:49 AM 3/9/2005, Eli Marmor wrote:
>Following the major patches, API changes, modifications and improvements
>since 2.04, is it possible to release a pre-2.05 (or a candidate to
>become 2.05) so we can start testing it and find bugs or problems under
>specific platforms, environments and configurations?
>
>I also want to use this opportunity to thank Joe, Stas, Max, Randy,
>Bojan, and all the other great developers who brought libapreq2 to its
>current status.
>
>-- 
>Eli Marmor
>marmor@netmask.it
>Netmask (El-Mar) Internet Technologies Ltd.
>__________________________________________________________
>Tel.:   +972-9-766-1020          8 Yad-Harutzim St.
>Fax.:   +972-9-766-1314          P.O.B. 7004
>Mobile: +972-50-5237338          Kfar-Saba 44641, Israel


Re: towards a 2.05 release

Posted by Eli Marmor <ma...@netmask.it>.
This is exactly why I suggested having a pre-2.05 release.

As Joe wrote, the source is not ready for 2.05, especially when there
is currently a regression in some areas in comparison with 2.04.

My suggestion: Give us a pre-2.05, without a usable Perl, and without a
correct documentation, so we can at least play with the new C API.

While the developers completes the work (Perl/doc/etc.), the rest of us
can port our code to the new C API and find all the last-minute-bugs.

I believe that it can lead to a better 2.05, maybe even without the
"-dev" extension...  ;-)

Joe Schaefer wrote:
> 
> Max Kellermann <ma...@duempel.org> writes:
> 
> > On 2005/03/09 15:49, Eli Marmor <ma...@netmask.it> wrote:
> >> Following the major patches, API changes, modifications and
> >> improvements since 2.04, is it possible to release a pre-2.05 (or a
> >> candidate to become 2.05) so we can start testing it and find bugs
> >> or problems under specific platforms, environments and
> >> configurations?
> >
> > I think we should rather push 2.05 through the door, finally.
> 
> Release a 2.05-dev soonish?  Sure, but the multi-env branch needs
> to become trunk before that happens, and there are a few obstacles
> which currently prevent that:
> 
>    1) sorting out the perl APIs:
>       a)  Providing the missing Table/Upload/Error APIs.
>       b)  Renaming Apache::Request -> Apache2::Request, etc.
>       c)  Documenting the new APR::Request modules, and fixing the old docs.
> 
>    2) Cleaning up the C API docs, and getting doxygen to pull
>       it all together properly (including the perl docs).
> 
>    3) Fix `make release`, and other CPAN-related build stuff.
>       I'm not sure this actually needs to be done in the branch
>       itself, but it is a regression when compared to trunk.
> 
> > What is left to do? I havn't taken time to work on libapreq in the
> > last weeks... let's make a "code review" session in the next few days
> > and collect ideas which should go into the stable API.
> 
> +1 for reviewing the C headers + documentation ASAP.  Anything
> documentation that gets translated into the test suite would be
> ideal.
> 
> > Joe, what is the status of the perl glue?
> 
> Should compile cleanly and test fairly well.  The current tests will
> fail on the missing Upload APIs, and the table APIs are also missing
> some functionality (notably do(), but I don't plan to provide set/add,
> nor will we inherit from APR::Table).

-- 
Eli Marmor
marmor@netmask.it
Netmask (El-Mar) Internet Technologies Ltd.
__________________________________________________________
Tel.:   +972-9-766-1020          8 Yad-Harutzim St.
Fax.:   +972-9-766-1314          P.O.B. 7004
Mobile: +972-50-5237338          Kfar-Saba 44641, Israel

Re: towards a 2.05 release

Posted by Joe Schaefer <jo...@sunstarsys.com>.
Max Kellermann <ma...@duempel.org> writes:

> On 2005/03/09 15:49, Eli Marmor <ma...@netmask.it> wrote:
>> Following the major patches, API changes, modifications and
>> improvements since 2.04, is it possible to release a pre-2.05 (or a
>> candidate to become 2.05) so we can start testing it and find bugs
>> or problems under specific platforms, environments and
>> configurations?
>
> I think we should rather push 2.05 through the door, finally.

Release a 2.05-dev soonish?  Sure, but the multi-env branch needs 
to become trunk before that happens, and there are a few obstacles
which currently prevent that:

   1) sorting out the perl APIs:
      a)  Providing the missing Table/Upload/Error APIs.
      b)  Renaming Apache::Request -> Apache2::Request, etc.
      c)  Documenting the new APR::Request modules, and fixing the old docs.

   2) Cleaning up the C API docs, and getting doxygen to pull
      it all together properly (including the perl docs).

   3) Fix `make release`, and other CPAN-related build stuff.
      I'm not sure this actually needs to be done in the branch 
      itself, but it is a regression when compared to trunk.

> What is left to do? I havn't taken time to work on libapreq in the
> last weeks... let's make a "code review" session in the next few days
> and collect ideas which should go into the stable API.

+1 for reviewing the C headers + documentation ASAP.  Anything
documentation that gets translated into the test suite would be
ideal.

> Joe, what is the status of the perl glue?

Should compile cleanly and test fairly well.  The current tests will
fail on the missing Upload APIs, and the table APIs are also missing 
some functionality (notably do(), but I don't plan to provide set/add, 
nor will we inherit from APR::Table).

-- 
Joe Schaefer


Re: towards a 2.05 release

Posted by Max Kellermann <ma...@duempel.org>.
On 2005/03/09 15:49, Eli Marmor <ma...@netmask.it> wrote:
> Following the major patches, API changes, modifications and
> improvements since 2.04, is it possible to release a pre-2.05 (or a
> candidate to become 2.05) so we can start testing it and find bugs
> or problems under specific platforms, environments and
> configurations?

I think we should rather push 2.05 through the door, finally.

What is left to do? I havn't taken time to work on libapreq in the
last weeks... let's make a "code review" session in the next few days
and collect ideas which should go into the stable API.

Joe, what is the status of the perl glue?

Max


Re: towards a 2.05 release

Posted by Stas Bekman <st...@stason.org>.
Bojan Smojver wrote:
> On Wed, 2005-03-09 at 16:49 +0200, Eli Marmor wrote:
> 
> 
>>I also want to use this opportunity to thank Joe, Stas, Max, Randy,
>>Bojan, and all the other great developers who brought libapreq2 to its
>>current status.
> 
> 
> Huh, I think I'll have to exclude myself from the list, given that I
> haven't done any work at all. I'm just a very happy libapreq2 user.

Certainly not me.

-- 
__________________________________________________________________
Stas Bekman            JAm_pH ------> Just Another mod_perl Hacker
http://stason.org/     mod_perl Guide ---> http://perl.apache.org
mailto:stas@stason.org http://use.perl.org http://apacheweek.com
http://modperlbook.org http://apache.org   http://ticketmaster.com

Re: towards a 2.05 release

Posted by Bojan Smojver <bo...@rexursive.com>.
On Wed, 2005-03-09 at 16:49 +0200, Eli Marmor wrote:

> I also want to use this opportunity to thank Joe, Stas, Max, Randy,
> Bojan, and all the other great developers who brought libapreq2 to its
> current status.

Huh, I think I'll have to exclude myself from the list, given that I
haven't done any work at all. I'm just a very happy libapreq2 user.

-- 
Bojan


Re: towards a 2.05 release

Posted by Eli Marmor <ma...@netmask.it>.
Following the major patches, API changes, modifications and improvements
since 2.04, is it possible to release a pre-2.05 (or a candidate to
become 2.05) so we can start testing it and find bugs or problems under
specific platforms, environments and configurations?

I also want to use this opportunity to thank Joe, Stas, Max, Randy,
Bojan, and all the other great developers who brought libapreq2 to its
current status.

-- 
Eli Marmor
marmor@netmask.it
Netmask (El-Mar) Internet Technologies Ltd.
__________________________________________________________
Tel.:   +972-9-766-1020          8 Yad-Harutzim St.
Fax.:   +972-9-766-1314          P.O.B. 7004
Mobile: +972-50-5237338          Kfar-Saba 44641, Israel

Re: towards a 2.05 release

Posted by Joe Schaefer <jo...@sunstarsys.com>.
Christian Hansen <ch...@ngmedia.com> writes:

> Joe Schaefer wrote:
>> Which tests, please?  It is never ok for the trunk
>> to be failing tests; just be sure the tests fail on a fresh svn checkout.
>
> All tests fails

Sorry, I can't reproduce it.  See if you have any better
luck with this tarball, which includes the buildconf patch:

    http://svn.apache.org/~joes/libapreq2-2.05-osx-test-2004-12-06.tar.gz

This is rolled with 

     % make release

so if there's a difference, it's probably more libtool/auto* 
breakage on OSX (maybe gmake is required with glibtool, dunno).

-- 
Joe Schaefer


Re: towards a 2.05 release

Posted by Christian Hansen <ch...@ngmedia.com>.
Joe Schaefer wrote:
> Christian Hansen <ch...@ngmedia.com> writes:
> 
> 
>>Joe Schaefer wrote:
> 
> 
> [...]
> 
> 
>>>Hmm, here's a patch that lets you do
>>>     % ./buildconf --with-libtool=glibtool
>>>Does it help at all?
>>
>>Yes, works greate, Tested against snapshot released 2004-11-29.
>>Latest  trunk and snapshot failed tests, guess its not in a
>>working state. 
> 
> 
> Which tests, please?  It is never ok for the trunk
> to be failing tests; just be sure the tests fail on 
> a fresh svn checkout.

All tests fails


With trunk, revision 110042:

   244  svn checkout http://svn.apache.org/repos/asf/httpd/apreq/trunk/ apreq
   245  cd apreq
   246  patch < ../buildconf.patch
   247  ./buildconf --with-libtool=glibtool
   248  perl Makefile.PL --with-apache2-apxs=/usr/local/apache2/bin/apxs
   249  make
   250  make test

Failed Test     Stat Wstat Total Fail  Failed  List of Failed
-------------------------------------------------------------------------------
t/big_request.t    0    13    21   22 104.76%  1-21
t/cgi.t                       31   31 100.00%  1-31
t/cookie.t                     5    5 100.00%  1-5
t/request.t                   18   18 100.00%  1-18
Failed 4/4 test scripts, 0.00% okay. 75/75 subtests failed, 0.00% okay.

env/t/logs/error_log:

[notice] child pid 7677 exit signal Bus error (10), possible coredump in 
/Users/chansen/src/apreq/apreq/env/t
[debug] mod_apreq_big_request_test.c(48): [client 127.0.0.1] initializing request
[debug] mod_apreq_big_request_test.c(48): [client 127.0.0.1] initializing request
[error] server reached MaxClients setting, consider raising the MaxClients setting
[notice] child pid 7678 exit signal Bus error (10), possible coredump in 
/Users/chansen/src/apreq/apreq/env/t

[error] [client 127.0.0.1] 
/Users/chansen/src/apreq/apreq/env/t/cgi-bin/test_cgi: error: 
/Users/chansen/src/apreq/apreq/env/t/cgi-bin/.libs/test_cgi does not exist
[error] [client 127.0.0.1] This script is just a wrapper for test_cgi.
[error] [client 127.0.0.1] See the libtool documentation for more information.
[error] [client 127.0.0.1] Premature end of script headers: test_cgi

[debug] mod_apreq_cookie_test.c(46): [client 127.0.0.1] initializing request
[debug] mod_apreq_cookie_test.c(46): [client 127.0.0.1] initializing request
[notice] child pid 7905 exit signal Bus error (10), possible coredump in 
/Users/chansen/src/apreq/apreq/env/t


With snapshot 2004-11-29:

     5  tar -zxf libapreq2-2.05-snapshot-2004-11-29.tar.gz
     6  cd libapreq2-2.05
     7  patch < ../buildconf.patch
     8  patch < ../Makefile.am.patch env/Makefile.am
     9  ./buildconf --with-libtool=glibtool
    10  perl Makefile.PL --with-apache2-apxs=/usr/local/apache2/bin/apxs
    11  make
    12  make test

t/big_request....ok
t/cgi............ok
t/cookie.........ok
t/request........ok
All tests successful.

t/apreq/big_input....ok
t/apreq/cgi..........ok
t/apreq/cookie.......ok
t/apreq/inherit......ok
t/apreq/request......ok
t/apreq/upload.......ok
All tests successful.

docs/Cookie.....ok
docs/Error......ok
docs/Request....ok
docs/Table......ok
docs/Upload.....ok
All tests successful.


OS: Darwin Kernel Version 7.6.0
Apache: 2.0.50
perl: 5.8.5
mod_perl: 1.9917
Apache::Test: 1.16
ExtUtils::MakeMaker: 6.21
ExtUtils::XSBuilder: 0.27
Test::More: 0.47

libtool: 1.5.2
autoconf: 2.59
automake: 1.7.9


Regards
Christian Hansen



Re: towards a 2.05 release

Posted by Joe Schaefer <jo...@sunstarsys.com>.
Christian Hansen <ch...@ngmedia.com> writes:

> Joe Schaefer wrote:

[...]

>> Hmm, here's a patch that lets you do
>>      % ./buildconf --with-libtool=glibtool
>> Does it help at all?
>
> Yes, works greate, Tested against snapshot released 2004-11-29.
> Latest  trunk and snapshot failed tests, guess its not in a
> working state. 

Which tests, please?  It is never ok for the trunk
to be failing tests; just be sure the tests fail on 
a fresh svn checkout.

Many thanks for the doc patch!

-- 
Joe Schaefer


Re: towards a 2.05 release

Posted by Christian Hansen <ch...@ngmedia.com>.
Joe Schaefer wrote:
> Christian Hansen <ch...@ngmedia.com> writes:
> 
> 
>>A minor thing with buildconf is that on OS X the GNU buildtools are
>>prfixed with a 'g':
>>
>>libtool = glibtool
>>libtoolize = glibtoolize
>>
> 
> 
> Hmm, here's a patch that lets you do
> 
>      % ./buildconf --with-libtool=glibtool
> 
> Does it help at all?  

Yes, works greate, Tested against snapshot released 2004-11-29. Latest 
trunk and snapshot failed tests, guess its not in a working state.

> If so, is it enough?

Sure is!


Thanks!
Christian

Re: towards a 2.05 release

Posted by Joe Schaefer <jo...@sunstarsys.com>.
Christian Hansen <ch...@ngmedia.com> writes:

> A minor thing with buildconf is that on OS X the GNU buildtools are
> prfixed with a 'g':
>
> libtool = glibtool
> libtoolize = glibtoolize
>

Hmm, here's a patch that lets you do

     % ./buildconf --with-libtool=glibtool

Does it help at all?  If so, is it enough?


Index: buildconf
===================================================================
--- buildconf	(revision 106989)
+++ buildconf	(working copy)
@@ -19,7 +19,7 @@
 #            checked-out version of the source code.
 
 perl="perl"
-libtool="libtoolize --automake --copy --force"
+libtool="libtool"
 
 while test $# -gt 0 
 do
@@ -33,11 +33,15 @@
   --with-perl=*)
   perl=$optarg
   ;;
+  --with-libtool=*)
+  libtool=$optarg
+  ;;
   esac
 
   shift
 done
 
+libtoolize="${libtool}ize --automake --copy --force"
 
 echo "removing stale config files"
 touch .deps
@@ -46,12 +50,12 @@
 rm -rf autom4te*.cache
 
 echo "checking buildconf prereqs";
-$perl build/version_check.pl libtool   || exit 1
+$perl build/version_check.pl libtool $libtool  || exit 1
 $perl build/version_check.pl autoconf  || exit 1
 $perl build/version_check.pl automake  || exit 1
 $perl build/version_check.pl ExtUtils::XSBuilder || exit 1
 
-echo "running $libtool" && $libtool && \
+echo "running $libtoolize" && $libtoolize && \
 echo "running aclocal"  && aclocal && \
 echo "running autoconf" && autoconf && \
 echo "running autoheader" && autoheader && \

-- 
Joe Schaefer


Re: towards a 2.05 release

Posted by Christian Hansen <ch...@ngmedia.com>.
Joe Schaefer wrote:
> Christian Hansen <ch...@ngmedia.com> writes:
> 
> 
>>This message [1] still aplies on MacOS X, after:
>>
>>ditto env/.libs env/t/.libs
>>
>>all tests passes.
> 
> 
> Remove env/t/.libs and try this patch. (you'll need to
> rerun buildconf).
> 
> 
> Index: env/Makefile.am
> ===================================================================
> --- env/Makefile.am	(revision 106989)
> +++ env/Makefile.am	(working copy)
> @@ -45,6 +45,7 @@
>  
>  run_tests : t/TEST
>  	-cp -f test_cgi t/cgi-bin
> +	-cp -a .libs t
>  	MAKE=$(MAKE) @PERL@ t/TEST
>  else
>  
>

Works greate if i change the flags, my cp does not have -a option.

Index: env/Makefile.am
===================================================================
--- env/Makefile.am     (revision 109715)
+++ env/Makefile.am     (working copy)
@@ -45,6 +45,7 @@

  run_tests : t/TEST
         -cp -f test_cgi t/cgi-bin
+       -cp -Rp .libs t
         MAKE=$(MAKE) @PERL@ t/TEST
  else



A minor thing with buildconf is that on OS X the GNU buildtools are 
prfixed with a 'g':

libtool = glibtool
libtoolize = glibtoolize


Regards
Christian Hansen


Re: towards a 2.05 release

Posted by Joe Schaefer <jo...@sunstarsys.com>.
Christian Hansen <ch...@ngmedia.com> writes:

> This message [1] still aplies on MacOS X, after:
>
> ditto env/.libs env/t/.libs
>
> all tests passes.

Remove env/t/.libs and try this patch. (you'll need to
rerun buildconf).


Index: env/Makefile.am
===================================================================
--- env/Makefile.am	(revision 106989)
+++ env/Makefile.am	(working copy)
@@ -45,6 +45,7 @@
 
 run_tests : t/TEST
 	-cp -f test_cgi t/cgi-bin
+	-cp -a .libs t
 	MAKE=$(MAKE) @PERL@ t/TEST
 else
 

-- 
Joe Schaefer


Re: towards a 2.05 release

Posted by Christian Hansen <ch...@ngmedia.com>.
Joe Schaefer wrote:
> I've tackled the few final ABI & build system
> changes that were in STATUS.  The rest look to me
> like issues that can be addressed in a future release. 
> 
> I'd like to start the release candidate process soon, so 
> anyone with a serious itch needs to have a look at the 
> current trunk to see what can be done about it.

This message [1] still aplies on MacOS X, after:

ditto env/.libs env/t/.libs

all tests passes.

[1] http://article.gmane.org/gmane.comp.apache.apreq/2350


Regards,

Christian


Re: towards a 2.05 release

Posted by Joe Schaefer <jo...@sunstarsys.com>.
Geoffrey Young <ge...@modperlcookbook.org> writes:

> recent reports indicated that the config process wasn't smart enough
> to figure out apr-1-config and so on for 2.1/2.2.

Can you confirm the reports?  2.1/2.2 is pretty much the only 
configuration I use during testing, so I'm surprised to hear 
that it doesn't actually work.

-- 
Joe Schaefer


Re: towards a 2.05 release

Posted by Geoffrey Young <ge...@modperlcookbook.org>.
> I'd like to start the release candidate process soon, so 
> anyone with a serious itch needs to have a look at the 
> current trunk to see what can be done about it.

recent reports indicated that the config process wasn't smart enough to
figure out apr-1-config and so on for 2.1/2.2.  if you point me to where you
would like the fix to live I'll find some time for it.

--Geoff