>
> Do you have a startup.pl, or anything else that would cause mod_perl
> to run code during initialization? I suspect that's all it takes.
>
Yes, a startup.pl
>
> Well, that makes me feel a bit better. This was getting so weird
> that there was always this feeling in the back of my min
At 8:02 AM +0100 3/22/02, Gerald Richter wrote:
> > I have to have a
>>
>>
>> block in the config file.
>>
>
>Mmh, I never had a need to do this...
Do you have a startup.pl, or anything else that would cause mod_perl
to run code during initialization? I suspect that's all it takes.
>
>>
> Good news. I have it somewhat working.
:-)
> I still can't get any
> directives read on the second pass, but I can get it to run.
:-(
>They
> key items:
>
> I have to have a
>
>
> block in the config file.
>
Mmh, I never had a need to do this...
> embperl_ApacheInitCleanup has to cal
Good news. I have it somewhat working. I still can't get any
directives read on the second pass, but I can get it to run. They
key items:
I have to have a
block in the config file.
embperl_ApacheInitCleanup has to call ap_remove_module
(&embperl_module) every time, not just when mod_per
Life just got more interesting. I was right, but it turns out there
is a solution for this in mod_perl, but it's not working.
perl_section_self_boot has the following comment:
/* make sure this module is re-loaded for the second config read */
However that code is not being executed.
BU
I have a theory which may or may not be correct, but I'm not sure
where to go next.
Embperl is loading and executing just fine the first time.
However the second time the config file is read (by standalone_main)
it is *not* loaded properly. This is true whether or not
standalone_main is execu
At 9:23 PM +0100 3/21/02, Gerald Richter wrote:
>There is one thing we can try, in epapinit.c comment out line 163:
>
>ap_remove_module (&embperl_module) ;
No go.
>
>
>If this all doesn't help, then search for all
>
> cfg -> bUseEnv = -1 ;
If I turn off the directives, I still crash with ot
> >
> >What are the different error message, if any, with and without the
> >AddModule?
>
> Kind of telling.
>
> Cannot add module via name 'embperl.c': not in list of loaded modules
>
Oops, yes it can't work. Sorry I am overseen this.
> It looks to me like the module is being only partially reg
At 3:56 PM +0100 3/21/02, Gerald Richter wrote:
>Embperl_UseEnv on
>AddModule embperl.c
>Embperl_UseEnv on
>
>What are the different error message, if any, with and without the
>AddModule?
Kind of telling.
Cannot add module via name 'embperl.c': not in list of loaded modules
It looks to me like
Looks like didn't get much sleep, since your last mail
>
> I assume all I have to do to not load 1.3.4 is make sure there are no
> references to HTML::Embperl in my config file? No difference.
>
Yes
> I'm definitely using the same apache and mod_perl, because I'm
> testing on the live serv
At 11:19 AM +0100 3/21/02, Gerald Richter wrote:
>So one step back: Does make test work for you? If yes, are you really using
>the same Apache and mod_perl in test and in life? For your last tests, did
>you load 2.0b6 only, or did you also load 1.3.4. Please try to load 2.0
>only. Does this make a
>
> The check for a null pointer worked, but I'm still not sure if we're
> fixing the bug, or working around a symptom.
>
It isn't bad anyway, since it makes Embperl more robust against problems,
where ever they might come from...
> Which leads me to my next problem. I can't seem to get config
That fix got me past
At 7:45 AM +0100 3/21/02, Gerald Richter wrote:
>Most other Perl module won't use Apache config directives, so there is not
>much to compare outside anyway.
The check for a null pointer worked, but I'm still not sure if we're
fixing the bug, or working around a symptom.
Whi
> >hosts. Do you use virtual hosts and have Embperl settings in more then
one
> >virtual host section ?
>
> Yes, but removing them doesn't fix the problem.
I have tested the virtual host problem inbetween and it seems not the source
of problems.
>
> First of all, Embperl is the only PerlModule I
At 8:49 PM +0100 3/20/02, Gerald Richter wrote:
>Ok, this mean something with Apache config goes very wrong (That's also what
>the backtraces tell use). Things are not initilized, when they should be set
>to zero. Jochen reported yesterday some problems with setup inside of virual
>hosts. Do you u
>
> I tracked the problem down, but then I lose it. We're faulting on
> the sAppName value. For whatever reason
> embperl_GetApacheAppName
> is being called and in it
> pDirCfg->AppConfig.sAppName=0x14
> which obviously isn't correct.
>
Ok, this mean something with Apache config goes very wrong
Okay, that's gotten me past the duplicate symbol errors. Now there's
the general problem of just running.
I tracked the problem down, but then I lose it. We're faulting on
the sAppName value. For whatever reason
embperl_GetApacheAppName
is being called and in it
pDirCfg->AppC
> >Never seen this errror message, but the stack backtraces looks like there
> >are more name clashes, so I expect it to work when you only load one of
> >them.
>
> OSX has an option for building shared libraries using multiple name
> spaces. I haven't explored it. It's the default, and typicall
At 8:08 PM +0100 3/19/02, Gerald Richter wrote:
>Never seen this errror message, but the stack backtraces looks like there
>are more name clashes, so I expect it to work when you only load one of
>them.
OSX has an option for building shared libraries using multiple name
spaces. I haven't explor
> At 7:06 PM +0100 3/19/02, Gerald Richter wrote:
> > > Now when I access an HTML::Embperl driven site, I get:
> >>
> >> dyld: /usr/local/apache/bin/httpd multiple definitions of symbol
> >> _EMBPERL_mgGetEscMode
> >
> >Use the appended epmacro.h file and rebuild Embperl (make clean before).
>
> Now when I access an HTML::Embperl driven site, I get:
>
> dyld: /usr/local/apache/bin/httpd multiple definitions of symbol
> _EMBPERL_mgGetEscMode
Use the appended epmacro.h file and rebuild Embperl (make clean before).
This will solve the name clash
Gerald
--
21 matches
Mail list logo
