You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@subversion.apache.org by Max Bowsher <ma...@ukf.net> on 2004/07/23 23:45:50 UTC
Attn: rey4 Bindings mess reversion incomplete
The trunk perl bindings are still broken on OS X and Cygwin, because there
is still an interdependency between libsvn_swig_perl and swig_core. The move
of the current_pool variable was not reverted.
Max.
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org
Re: rey4 Bindings mess reversion incomplete
Posted by Russell Yanofsky <re...@columbia.edu>.
Max Bowsher wrote:
> ...
> However, this doesn't work on Cygwin, and making libsvn_swig_perl
> into a static library isn't a solution, because it needs to link to a
> shared libswigpl.
What is the actual error you get with the patch? libsvn_swig_perl shouldn't
have a link step since its a static library. And all the perl modules that
depend on it already link to libswigpl.
> Please, can we put Perl's current_pool back in libsvn_swig_perl?
>
> Otherwise, this interdependency is *really* going to mess up ever
> getting the perl bindings to work on Cygwin.
We can revert the change, but I'd like to know how the static library
solution fails on cygwin.
> Could you explain what the gains of the big bindings shakeup begun in
> r10233 are? I'm not familiar with the Windows build, so it's hard for
> me to understand what the motivation was.
The main motivation was to simplify the windows build. Putting the swig
runtime into the core modules allowed us to remove the clunky
gen_swig_runtime.py script and other apparatus for building the swig runtime
separately. Ben Reser also said it could help with multiple versions of perl
on the same machine (
http://www.contactor.se/~dast/svn/archive-2004-07/0451.shtml ).
r10233 moved the swig runtime into core modules on windows
r10236 got the same thing working on unix
r10374 reverted the effects of r10233 and r10236 on unix due to
incompatibilities
- Russ
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org
Re: rey4 Bindings mess reversion incomplete
Posted by Max Bowsher <ma...@ukf.net>.
Russell Yanofsky wrote:
> Max Bowsher wrote:
>> Russell Yanofsky wrote:
>>> Max Bowsher wrote:
>>>> Russell Yanofsky wrote:
>>>>> What if you apply the patch at
>>>>> http://www.contactor.se/~dast/svn/archive-2004-07/0586.shtml ?
>>>>
>>>> No good. Darwin.
>>>
>>> I don't understand. Why doesn't it work on Darwin?
>>
>> Because dynamic library libsvn_swig_perl cannot link to loadable
>> module core.
>
> If the patch works, it should make libsvn_swig_perl into a static library.
First, let me apologize - this does work on darwin. I misunderstood the
rules involving bundles and dylibs.
However, this doesn't work on Cygwin, and making libsvn_swig_perl into a
static library isn't a solution, because it needs to link to a shared
libswigpl.
Please, can we put Perl's current_pool back in libsvn_swig_perl?
Otherwise, this interdependency is *really* going to mess up ever getting
the perl bindings to work on Cygwin.
Could you explain what the gains of the big bindings shakeup begun in r10233
are? I'm not familiar with the Windows build, so it's hard for me to
understand what the motivation was.
> (Also, your replies have been accidentally sent to dev@apr.apache.org, not
> the subversion list)
Oops, sorry, that would be Outlook Express being dumb, and me not watching
what it's doing carefully enough.
Max.
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org
Re: rey4 Bindings mess reversion incomplete
Posted by Russell Yanofsky <re...@columbia.edu>.
Max Bowsher wrote:
> Russell Yanofsky wrote:
>> Max Bowsher wrote:
>>> Russell Yanofsky wrote:
>>>> What if you apply the patch at
>>>> http://www.contactor.se/~dast/svn/archive-2004-07/0586.shtml ?
>>>
>>> No good. Darwin.
>>
>> I don't understand. Why doesn't it work on Darwin?
>
> Because dynamic library libsvn_swig_perl cannot link to loadable
> module core.
If the patch works, it should make libsvn_swig_perl into a static library.
- Russ
(Also, your replies have been accidentally sent to dev@apr.apache.org, not
the subversion list)
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org
Re: rey4 Bindings mess reversion incomplete
Posted by Branko Čibej <br...@xbc.nu>.
[moved to the correct dev list -- careful there, Max]
Max Bowsher wrote:
>Russell Yanofsky wrote:
>
>
>>Max Bowsher wrote:
>>
>>
>>>Russell Yanofsky wrote:
>>>
>>>
>>>>Max Bowsher wrote:
>>>>
>>>>
>>>>>The trunk perl bindings are still broken on OS X and Cygwin, because
>>>>>there is still an interdependency between libsvn_swig_perl and
>>>>>swig_core. The move of the current_pool variable was not reverted.
>>>>>
>>>>>Max.
>>>>>
>>>>>
>>>>What if you apply the patch at
>>>>http://www.contactor.se/~dast/svn/archive-2004-07/0586.shtml ?
>>>>
>>>>
>>>No good. Darwin.
>>>
>>>
>>I don't understand. Why doesn't it work on Darwin?
>>
>>
>
>Because dynamic library libsvn_swig_perl cannot link to loadable module
>core.
>
>
>
>>>>I'm reluntant
>>>>to move the current_pool variable back to libsvn_swig_perl, because
>>>>it makes the swig code messier and requires that we build
>>>>libsvn_swig_perl as a dll on windows.
>>>>
>>>>
>>>Why is that bad?
>>>
>>>
>>It's not really bad, just different than what we do with
>>
>>
>libsvn_swig_python
>
>
>>and libsvn_swig_java.
>>
>>
>
>On unix we build libsvn_swig_* as shared libraries. Why does the the Windows
>build not do that too?
>
>
Because the Windows build does not build DLLs in general, yet, and I
don't want to make too many exceptions. Besides, on Unix we can build
both static and shared libs, I believe.
-- Brane
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org
Re: rey4 Bindings mess reversion incomplete
Posted by Max Bowsher <ma...@ukf.net>.
Russell Yanofsky wrote:
> Max Bowsher wrote:
>> Russell Yanofsky wrote:
>>> Max Bowsher wrote:
>>>> The trunk perl bindings are still broken on OS X and Cygwin, because
>>>> there is still an interdependency between libsvn_swig_perl and
>>>> swig_core. The move of the current_pool variable was not reverted.
>>>>
>>>> Max.
>>>
>>> What if you apply the patch at
>>> http://www.contactor.se/~dast/svn/archive-2004-07/0586.shtml ?
>>
>> No good. Darwin.
>
> I don't understand. Why doesn't it work on Darwin?
Because dynamic library libsvn_swig_perl cannot link to loadable module
core.
>>> I'm reluntant
>>> to move the current_pool variable back to libsvn_swig_perl, because
>>> it makes the swig code messier and requires that we build
>>> libsvn_swig_perl as a dll on windows.
>>
>> Why is that bad?
>
> It's not really bad, just different than what we do with
libsvn_swig_python
> and libsvn_swig_java.
On unix we build libsvn_swig_* as shared libraries. Why does the the Windows
build not do that too?
Max.
Re: rey4 Bindings mess reversion incomplete
Posted by Russell Yanofsky <re...@columbia.edu>.
Max Bowsher wrote:
> Russell Yanofsky wrote:
>> Max Bowsher wrote:
>>> The trunk perl bindings are still broken on OS X and Cygwin, because
>>> there is still an interdependency between libsvn_swig_perl and
>>> swig_core. The move of the current_pool variable was not reverted.
>>>
>>> Max.
>>
>> What if you apply the patch at
>> http://www.contactor.se/~dast/svn/archive-2004-07/0586.shtml ?
>
> No good. Darwin.
I don't understand. Why doesn't it work on Darwin?
>> I'm reluntant
>> to move the current_pool variable back to libsvn_swig_perl, because
>> it makes the swig code messier and requires that we build
>> libsvn_swig_perl as a dll on windows.
>
> Why is that bad?
It's not really bad, just different than what we do with libsvn_swig_python
and libsvn_swig_java.
- Russ
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org
Re: rey4 Bindings mess reversion incomplete
Posted by Max Bowsher <ma...@ukf.net>.
Russell Yanofsky wrote:
> Max Bowsher wrote:
>> The trunk perl bindings are still broken on OS X and Cygwin, because
>> there is still an interdependency between libsvn_swig_perl and
>> swig_core. The move of the current_pool variable was not reverted.
>>
>> Max.
>
> What if you apply the patch at
> http://www.contactor.se/~dast/svn/archive-2004-07/0586.shtml ?
No good. Darwin.
> I'm reluntant
> to move the current_pool variable back to libsvn_swig_perl, because it
makes
> the swig code messier and requires that we build libsvn_swig_perl as a dll
> on windows.
Why is that bad?
Max.
Re: rey4 Bindings mess reversion incomplete
Posted by Russell Yanofsky <re...@columbia.edu>.
Max Bowsher wrote:
> The trunk perl bindings are still broken on OS X and Cygwin, because
> there is still an interdependency between libsvn_swig_perl and
> swig_core. The move of the current_pool variable was not reverted.
>
> Max.
What if you apply the patch at
http://www.contactor.se/~dast/svn/archive-2004-07/0586.shtml ? I'm reluntant
to move the current_pool variable back to libsvn_swig_perl, because it makes
the swig code messier and requires that we build libsvn_swig_perl as a dll
on windows.
- Russ
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org