Re: [Fink-devel] new pdb (hopefully) ready for testing

2007-12-06 Thread Alexey Zakhlestin
I entered "%%%" in the search field — it returned me everything… I
guess it is not what is supposed to happen… because none of those
results actually had '%' in their name…

% and _ symbols should be escaped by slashes before querying

On 12/5/07, Benjamin Reed <[EMAIL PROTECTED]> wrote:
> -BEGIN PGP SIGNED MESSAGE-
> Hash: SHA1
>
> There are still some load issues on the server, but I'd like to put out
> a request to see how things stand up to a little actual load from users.
>
> Please let me know if you hit any issues, or see something that doesn't
> look right.
>
> If you'd like to give it a shot, please check out:
>
>   http://ranger.users.finkproject.org/fink/pdb/
>
> If by the time you see this email it says it's down, that means the
> server couldn't take it, in which case, please *don't* let me know if
> you hit any issues.  ;)
>
> - --
> Benjamin Reed a.k.a. Ranger Rick
> Fink, KDE, and Mac OS X development
> http://www.racoonfink.com/
>
> -BEGIN PGP SIGNATURE-
> Version: GnuPG v1.4.6 (Darwin)
> Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org
>
> iD8DBQFHVvxzUu+jZtP2Zf4RAh6MAJ9GVU+TBs/wjKpXfUfgwB6W4PjpbgCfeK1M
> MzJFTRp5nm7oLDGsjcZnFQI=
> =Loaz
> -END PGP SIGNATURE-
>
> -
> SF.Net email is sponsored by: The Future of Linux Business White Paper
> from Novell.  From the desktop to the data center, Linux is going
> mainstream.  Let it simplify your IT future.
> http://altfarm.mediaplex.com/ad/ck/8857-50307-18918-4
> ___
> Fink-devel mailing list
> Fink-devel@lists.sourceforge.net
> http://news.gmane.org/gmane.os.apple.fink.devel
>


-- 
Alexey Zakhlestin
http://blog.milkfarmsoft.com/
-
SF.Net email is sponsored by: The Future of Linux Business White Paper
from Novell.  From the desktop to the data center, Linux is going
mainstream.  Let it simplify your IT future.
http://altfarm.mediaplex.com/ad/ck/8857-50307-18918-4
___
Fink-devel mailing list
Fink-devel@lists.sourceforge.net
http://news.gmane.org/gmane.os.apple.fink.devel

Re: [Fink-devel] new pdb (hopefully) ready for testing

2007-12-06 Thread Alexey Zakhlestin
On 12/6/07, Benjamin Reed <[EMAIL PROTECTED]> wrote:
> Alexey Zakhlestin wrote:
> > I entered "%%%" in the search field — it returned me everything… I
> > guess it is not what is supposed to happen… because none of those
> > results actually had '%' in their name…
> >
> > % and _ symbols should be escaped by slashes before querying
>
> Considering it's not using sql, I'm kind of surprised that worked...

that's even funnier, yup ;)

-- 
Alexey Zakhlestin
http://blog.milkfarmsoft.com/
-
SF.Net email is sponsored by: The Future of Linux Business White Paper
from Novell.  From the desktop to the data center, Linux is going
mainstream.  Let it simplify your IT future.
http://altfarm.mediaplex.com/ad/ck/8857-50307-18918-4
___
Fink-devel mailing list
Fink-devel@lists.sourceforge.net
http://news.gmane.org/gmane.os.apple.fink.devel

Re: [Fink-devel] new pdb (hopefully) ready for testing

2007-12-06 Thread Benjamin Reed
-BEGIN PGP SIGNED MESSAGE-
Hash: SHA1

Alexey Zakhlestin wrote:
> I entered "%%%" in the search field — it returned me everything… I
> guess it is not what is supposed to happen… because none of those
> results actually had '%' in their name…
> 
> % and _ symbols should be escaped by slashes before querying

Considering it's not using sql, I'm kind of surprised that worked...

- --
Benjamin Reed a.k.a. Ranger Rick
Fink, KDE, and Mac OS X development
http://www.racoonfink.com/

-BEGIN PGP SIGNATURE-
Version: GnuPG v1.4.6 (Darwin)
Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org

iD8DBQFHV/UQUu+jZtP2Zf4RAhs9AJ9dAGl/7eW6QH2DY0NFBr3hawpwuACffQkG
1R61smq4vHFuI8QAlOoJeOQ=
=jLbW
-END PGP SIGNATURE-

-
SF.Net email is sponsored by: The Future of Linux Business White Paper
from Novell.  From the desktop to the data center, Linux is going
mainstream.  Let it simplify your IT future.
http://altfarm.mediaplex.com/ad/ck/8857-50307-18918-4
___
Fink-devel mailing list
Fink-devel@lists.sourceforge.net
http://news.gmane.org/gmane.os.apple.fink.devel

Re: [Fink-devel] new pdb (hopefully) ready for testing

2007-12-06 Thread Alexey Zakhlestin
hmm… do you have some alphanumeric filter there?

three dots "..." gave me the same full resultset… :-/

-- 
Alexey Zakhlestin
http://blog.milkfarmsoft.com/
-
SF.Net email is sponsored by: The Future of Linux Business White Paper
from Novell.  From the desktop to the data center, Linux is going
mainstream.  Let it simplify your IT future.
http://altfarm.mediaplex.com/ad/ck/8857-50307-18918-4
___
Fink-devel mailing list
Fink-devel@lists.sourceforge.net
http://news.gmane.org/gmane.os.apple.fink.devel

Re: [Fink-devel] Xerces-C compilation frustration

2007-12-06 Thread Jürgen Lorenz Simon
Hello Martin,

On 06.12.2007, at 01:16, Martin Costabel wrote:

> Jürgen Lorenz Simon wrote:
>
> I just built xerces-c on a G5/10.4 and on an intel/10.5 without any  
> problem. I have a suspicion:
>
> []
> []
>> c-2.7.0-1002/xerces-c-src_2_7_0/lib -lxerces-c
>> ld: Undefined symbols:
>> __ZN11xercesc_2_712XMLException14loadExceptTextENS_10XMLExcepts5Codes 
>> E
>
> These symbols are defined in libxerces-c.dylib, and this library is  
> linked with with the -lxerces-c flag. Are you absolutely sure that  
> you don't have an older version of this library lying around  
> somewhere where the linker may pick it up, for example in /usr/ 
> local/lib?
>
> -- 
> Martin
>

Spot on, mate! I removed two versions of xerces from /usr/local/lib  
and /usr/lib, and
now the compilaton runs without fail.  If you make it to Bonn,  
Germany, let me buy
you a cold one.

Thanks!


-
SF.Net email is sponsored by: The Future of Linux Business White Paper
from Novell.  From the desktop to the data center, Linux is going
mainstream.  Let it simplify your IT future.
http://altfarm.mediaplex.com/ad/ck/8857-50307-18918-4
___
Fink-devel mailing list
Fink-devel@lists.sourceforge.net
http://news.gmane.org/gmane.os.apple.fink.devel


[Fink-devel] Keeping the tracker moving

2007-12-06 Thread Alexander K. Hansen
-BEGIN PGP SIGNED MESSAGE-
Hash: SHA1

I'm probably a month+ away from having Leopard going (I've asked for
it for Christmas and it's going to be a very busy December and January
for me), and I don't foresee an Intel box, and therefore the ability
for me to look at Intel-only issues, in the near future.

There's at least one package on the tracker right now that is listed
as an Intel-only upgrade, so if somebody could take a look at it (and
others, if present), it would be nice; if anything is 10.5-only (I
haven't noted anything as of yet) that could also benefit from some
attention.  Commit access isn't mandatory.

I'm kind of thinking that I should write up a developer-type document
(on the wiki) dealing with the procedures that I use for evaluating
packages on the tracker, and to get additional contributions in case
I've left something out.  If we have a systematic methodology to
follow, it might be easier for somebody who actually has an Intel box
to step in and follow identical procedures to what I'm doing and
thereby maintain consistency.
-BEGIN PGP SIGNATURE-
Version: GnuPG v1.4.6 (GNU/Linux)
Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org

iD8DBQFHWCsGB8UpO3rKjQ8RAoDpAJ4/CM4uNIPnmERR6D7lMHJQP5cnZwCdFP5c
fiXrMxe8+n0IAjdTOwKc7MQ=
=4Yid
-END PGP SIGNATURE-


-
SF.Net email is sponsored by: The Future of Linux Business White Paper
from Novell.  From the desktop to the data center, Linux is going
mainstream.  Let it simplify your IT future.
http://altfarm.mediaplex.com/ad/ck/8857-50307-18918-4
___
Fink-devel mailing list
Fink-devel@lists.sourceforge.net
http://news.gmane.org/gmane.os.apple.fink.devel


Re: [Fink-devel] new pdb (hopefully) ready for testing

2007-12-06 Thread Benjamin Reed
-BEGIN PGP SIGNED MESSAGE-
Hash: SHA1

Alexey Zakhlestin wrote:
> hmm… do you have some alphanumeric filter there?
> 
> three dots "..." gave me the same full resultset… :-/

Out of curiosity, what would you *expect* it to give you?

=)

I mean, it's basically equivalent to an empty query in the context of
full-text search.  It's not words, and not any part of a package name, etc.

- --
Benjamin Reed a.k.a. Ranger Rick
Fink, KDE, and Mac OS X development
http://www.racoonfink.com/

-BEGIN PGP SIGNATURE-
Version: GnuPG v1.4.6 (Darwin)
Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org

iD8DBQFHWDI8Uu+jZtP2Zf4RAlm2AJ9hC7MZNLlvJ1cACbrYTm1LINXE0QCdHsaU
9+VyH3y0c9PJLrC+85Xs5cM=
=pkZF
-END PGP SIGNATURE-

-
SF.Net email is sponsored by: The Future of Linux Business White Paper
from Novell.  From the desktop to the data center, Linux is going
mainstream.  Let it simplify your IT future.
http://altfarm.mediaplex.com/ad/ck/8857-50307-18918-4
___
Fink-devel mailing list
Fink-devel@lists.sourceforge.net
http://news.gmane.org/gmane.os.apple.fink.devel

Re: [Fink-devel] Keeping the tracker moving

2007-12-06 Thread John Ridgway
On Dec 6, 2007, at 12:01 PM, Alexander K. Hansen wrote:

> -BEGIN PGP SIGNED MESSAGE-
> Hash: SHA1
>
> I'm probably a month+ away from having Leopard going (I've asked for
> it for Christmas and it's going to be a very busy December and January
> for me), and I don't foresee an Intel box, and therefore the ability
> for me to look at Intel-only issues, in the near future.
>
> There's at least one package on the tracker right now that is listed
> as an Intel-only upgrade, so if somebody could take a look at it (and
> others, if present), it would be nice; if anything is 10.5-only (I
> haven't noted anything as of yet) that could also benefit from some
> attention.  Commit access isn't mandatory.
>
> I'm kind of thinking that I should write up a developer-type document
> (on the wiki) dealing with the procedures that I use for evaluating
> packages on the tracker, and to get additional contributions in case
> I've left something out.  If we have a systematic methodology to
> follow, it might be easier for somebody who actually has an Intel box
> to step in and follow identical procedures to what I'm doing and
> thereby maintain consistency.


Writing that up would be a great idea -- for two reasons:
1) Those of us with some time could assist with checking submitted  
packages.
2) Those of us who submit packages could use the document to do our  
own checking -- thus leading, with a little luck, to better initial  
submissions.

Peace
- John



-
SF.Net email is sponsored by: The Future of Linux Business White Paper
from Novell.  From the desktop to the data center, Linux is going
mainstream.  Let it simplify your IT future.
http://altfarm.mediaplex.com/ad/ck/8857-50307-18918-4
___
Fink-devel mailing list
Fink-devel@lists.sourceforge.net
http://news.gmane.org/gmane.os.apple.fink.devel


Re: [Fink-devel] new pdb (hopefully) ready for testing

2007-12-06 Thread Dave Vasilevsky
Looks very cool! A few minor quibbles:

* The 'Sort order' field seems meaningless. The only option is
'Descending', but the results are always sorted ascending by name
anyhow. I'd suggest just scrapping the field, what's the point of
sorting descending-by-name?

* Could you add a way to search only the current distributions? Might
even want to turn that on by default--if Joe User sees a package in
the result list he'd probably assume the package is actually
available, can't trust him to click through to the package page.

Thanks for all your pdb work!
Dave

-
SF.Net email is sponsored by: The Future of Linux Business White Paper
from Novell.  From the desktop to the data center, Linux is going
mainstream.  Let it simplify your IT future.
http://altfarm.mediaplex.com/ad/ck/8857-50307-18918-4
___
Fink-devel mailing list
Fink-devel@lists.sourceforge.net
http://news.gmane.org/gmane.os.apple.fink.devel


Re: [Fink-devel] new pdb (hopefully) ready for testing

2007-12-06 Thread Alexey Zakhlestin
On 12/6/07, Benjamin Reed <[EMAIL PROTECTED]> wrote:
> Alexey Zakhlestin wrote:
> > three dots "..." gave me the same full resultset… :-/
>
> Out of curiosity, what would you *expect* it to give you?
>
> =)
>
> I mean, it's basically equivalent to an empty query in the context of
> full-text search.  It's not words, and not any part of a package name, etc.

I expected to get empty set :-)
except the case, when there is, actually, a package which has three
dots (or percent-sign, or underscore)

I want to be able to find all packages which have underscore or dash
in their names

-- 
Alexey Zakhlestin
http://blog.milkfarmsoft.com/
-
SF.Net email is sponsored by: The Future of Linux Business White Paper
from Novell.  From the desktop to the data center, Linux is going
mainstream.  Let it simplify your IT future.
http://altfarm.mediaplex.com/ad/ck/8857-50307-18918-4
___
Fink-devel mailing list
Fink-devel@lists.sourceforge.net
http://news.gmane.org/gmane.os.apple.fink.devel

Re: [Fink-devel] new pdb (hopefully) ready for testing

2007-12-06 Thread Robert T Wyatt
Dave Vasilevsky wrote:
> Looks very cool! A few minor quibbles:
> 
> * The 'Sort order' field seems meaningless. The only option is
> 'Descending', but the results are always sorted ascending by name
> anyhow. I'd suggest just scrapping the field, what's the point of
> sorting descending-by-name?

It's the only way some of us ever get to be first!

> 
> * Could you add a way to search only the current distributions? Might
> even want to turn that on by default--if Joe User sees a package in
> the result list he'd probably assume the package is actually
> available, can't trust him to click through to the package page.


This is s true. How many times does a user say they see a package in 
the PDB but they can't install it (because it's in a different tree)?

If the user could choose the target tree, then choosing 
unstable should be one of the more obvious 'toggles' as it 
were.


> Thanks for all your pdb work!
> Dave

-
SF.Net email is sponsored by:
Check out the new SourceForge.net Marketplace.
It's the best place to buy or sell services for
just about anything Open Source.
http://sourceforge.net/services/buy/index.php
___
Fink-devel mailing list
Fink-devel@lists.sourceforge.net
http://news.gmane.org/gmane.os.apple.fink.devel


Re: [Fink-devel] Keeping the tracker moving

2007-12-06 Thread Robert T Wyatt
John Ridgway wrote:
> On Dec 6, 2007, at 12:01 PM, Alexander K. Hansen wrote:
> 
>> -BEGIN PGP SIGNED MESSAGE-
>> Hash: SHA1
>>
>> I'm probably a month+ away from having Leopard going (I've asked for
>> it for Christmas and it's going to be a very busy December and January
>> for me), and I don't foresee an Intel box, and therefore the ability
>> for me to look at Intel-only issues, in the near future.
>>
>> There's at least one package on the tracker right now that is listed
>> as an Intel-only upgrade, so if somebody could take a look at it (and
>> others, if present), it would be nice; if anything is 10.5-only (I
>> haven't noted anything as of yet) that could also benefit from some
>> attention.  Commit access isn't mandatory.
>>
>> I'm kind of thinking that I should write up a developer-type document
>> (on the wiki) dealing with the procedures that I use for evaluating
>> packages on the tracker, and to get additional contributions in case
>> I've left something out.  If we have a systematic methodology to
>> follow, it might be easier for somebody who actually has an Intel box
>> to step in and follow identical procedures to what I'm doing and
>> thereby maintain consistency.
> 
> 
> Writing that up would be a great idea -- for two reasons:
> 1) Those of us with some time could assist with checking submitted  
> packages.
> 2) Those of us who submit packages could use the document to do our  
> own checking -- thus leading, with a little luck, to better initial  
> submissions.
> 
> Peace
> - John

+1 Do to my infrequent participation I hesitate to help evaluate tracker 
packages, but if I had something to follow I'd be happy to test stuff 
and provide input.

-
SF.Net email is sponsored by:
Check out the new SourceForge.net Marketplace.
It's the best place to buy or sell services for
just about anything Open Source.
http://sourceforge.net/services/buy/index.php
___
Fink-devel mailing list
Fink-devel@lists.sourceforge.net
http://news.gmane.org/gmane.os.apple.fink.devel


Re: [Fink-devel] Keeping the tracker moving

2007-12-06 Thread Robert T Wyatt
Robert T Wyatt wrote:
> +1 Do to my infrequent participation I hesitate to help evaluate tracker 
> packages, but if I had something to follow I'd be happy to test stuff 
> and provide input.

Due-do. It's been a long couple of days: 
http://ccwf.cc.utexas.edu/~robert/images/utc.gif
(2nd from left looking toward the camera)


-
SF.Net email is sponsored by:
Check out the new SourceForge.net Marketplace.
It's the best place to buy or sell services for
just about anything Open Source.
http://sourceforge.net/services/buy/index.php
___
Fink-devel mailing list
Fink-devel@lists.sourceforge.net
http://news.gmane.org/gmane.os.apple.fink.devel


Re: [Fink-devel] new pdb (hopefully) ready for testing

2007-12-06 Thread Koen van der Drift

On Dec 5, 2007, at 2:30 PM, Benjamin Reed wrote:

> -BEGIN PGP SIGNED MESSAGE-
> Hash: SHA1
>
> There are still some load issues on the server, but I'd like to put  
> out
> a request to see how things stand up to a little actual load from  
> users.
>
> Please let me know if you hit any issues, or see something that  
> doesn't
> look right.


There are a few packages that I don't maintain on the page with all my  
packages:

http://ranger.users.finkproject.org/fink/pdb/browse.php?maintainer=Koen%20van%20der%20Drift

xml-sax-pm
xml-sax-expat-pm

I used to maintain the second one (but it is set as 'none' now), not  
sure why the first one is also there. Maybe it only matches the first  
few characters?

- Koen.

-
SF.Net email is sponsored by:
Check out the new SourceForge.net Marketplace.
It's the best place to buy or sell services for
just about anything Open Source.
http://sourceforge.net/services/buy/index.php
___
Fink-devel mailing list
Fink-devel@lists.sourceforge.net
http://news.gmane.org/gmane.os.apple.fink.devel


Re: [Fink-devel] new pdb (hopefully) ready for testing

2007-12-06 Thread Benjamin Reed
-BEGIN PGP SIGNED MESSAGE-
Hash: SHA1

Koen van der Drift wrote:

> xml-sax-pm
> xml-sax-expat-pm
> 
> I used to maintain the second one (but it is set as 'none' now), not
> sure why the first one is also there. Maybe it only matches the first
> few characters?

well, it's matching *all* versions of the package that are indexed;
according to the pdb, you maintained xml-sax at version 0.13.2 in the
0.8.1 binary release:

http://ranger.users.finkproject.org/fink/pdb/package.php/xml-sax-pm581?rel_id=10.4-powerpc-0.8.1-bindist

CVS agrees:

http://fink.cvs.sourceforge.net/fink/dists/10.4/stable/main/finkinfo/libs/perlmods/xml-sax-pm.info?pathrev=release_0_8_1&view=markup

I'm thinking maybe if you don't specify, we should default to searching
unstable-only or something.

- --
Benjamin Reed a.k.a. Ranger Rick
Fink, KDE, and Mac OS X development
http://www.racoonfink.com/

-BEGIN PGP SIGNATURE-
Version: GnuPG v1.4.6 (Darwin)
Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org

iD8DBQFHWKIAUu+jZtP2Zf4RAl01AJ4kXaKSZT9lztmPzN8PodX+C561/wCeOqqt
KODjq2p6TnAtTbvM6PVZqRY=
=eAba
-END PGP SIGNATURE-

-
SF.Net email is sponsored by:
Check out the new SourceForge.net Marketplace.
It's the best place to buy or sell services for
just about anything Open Source.
http://sourceforge.net/services/buy/index.php
___
Fink-devel mailing list
Fink-devel@lists.sourceforge.net
http://news.gmane.org/gmane.os.apple.fink.devel


[Fink-devel] final octave packaging

2007-12-06 Thread Jack Howarth
   It appears that the Apple c++ compiler on Intel has a bug which
causes the ov_fcn_handle.cc test case to segfault in octave. Likewise
the FSF g++ compiler from gcc42 causes the test_sparse.m testcase to
segfault on powerpc. The gcc43 snapshot compiler doesn't have this
problem. For now, the best approach is just to fork the build. On
i386, the build of octave is done with only the gcc42 compilers.
On powerpc, only the gfortran compiler is used from gcc42. This
produces very good make check results...

powerpc

Summary:

  PASS   3762
  FAIL  8

i386

Summary:

  PASS   3763
  FAIL  7

I have uploaded this updated packaging to the fink tracking entry...

http://sourceforge.net/tracker/index.php?func=detail&aid=1844593&group_id=17203&atid=414256

   Jack

-
SF.Net email is sponsored by:
Check out the new SourceForge.net Marketplace.
It's the best place to buy or sell services for
just about anything Open Source.
http://sourceforge.net/services/buy/index.php
___
Fink-devel mailing list
Fink-devel@lists.sourceforge.net
http://news.gmane.org/gmane.os.apple.fink.devel


Re: [Fink-devel] [Fink-users] xmms-coreaudio 1.0-3 Plugin Not Showing up in xmms 1.2.10-4

2007-12-06 Thread Martin Costabel
Peter O'Gorman wrote:
[]
> So the symbol is still there on 10.5, but is private extern? That is,
> indeed, what nmedit is supposed to do. There are other issues with
> nmedit on Leopard (it refuses to strip global coalesced symbols -
> whatever they are) that mean that I will have to patch libtool.

The nmedit command leaves only one exported symbol, get_oplugin_info. 
This, when called from xmms after loading the plugin via dlopen(), would 
return a structure osx_op, which has osx_about as one of its members. At 
the call of get_oplugin_info(), there appears the undefined symbol error.

For all other plugins, this works, but they are built from the xmms 
package, and this has the old libtool and does not use nmedit. For these 
plugins, the corresponding *_op and *_about symbols are extern. Only the 
libOSX.so plugin is built separately with the xmms-coreaudio package and 
uses nmedit to turn them into private externs.

> I'll try to get around to building xmms.
> 
> In the meantime, does something like this work for you in a patchscript
> (sed script is all on one line)?
> 
> sed -e "[EMAIL PROTECTED]@ -exported_symbols_list
> \$output_objdir/\${libname}-symbols.expsym\'@g" < configure > configure.new
> mv configure.new configure

This doesn't match anything in configure. In fact, there are lines with 
nmedit in them, but they have a format that I don't understand at all, like:

module_expsym_cmds_GCJ='sed -e "s,#.*,," -e "s,^[]*,," -e 
"s,^\(..*\),_&," < $export_symbols > 
$output_objdir/${libname}-symbols.expsym${_S_}$CC -bundle $archargs 
$allow_undefined_flag  -o $lib $libobjs 
$deplibs$compiler_flags${_S_}nmedit -s 
$output_objdir/${libname}-symbols.expsym ${lib}'

This is all one line and there is no space in front of nmedit. I don't 
know how this is transformed into three separate command lines later on, 
but it is.

-- 
Martin



-
SF.Net email is sponsored by:
Check out the new SourceForge.net Marketplace.
It's the best place to buy or sell services for
just about anything Open Source.
http://sourceforge.net/services/buy/index.php
___
Fink-devel mailing list
Fink-devel@lists.sourceforge.net
http://news.gmane.org/gmane.os.apple.fink.devel


Re: [Fink-devel] [Fink-users] xmms-coreaudio 1.0-3 Plugin Not Showing up in xmms 1.2.10-4

2007-12-06 Thread Peter O'Gorman
Martin Costabel wrote:
> Jean-François Mertens wrote:
> []
>> Do not understand fully this argument.
>> I have exactly the same output from the above 'nm -m' command on 10.4;
>> so apparently _ up to what is visible from this output _ the result
>> of the nmedit command is the same on 10.4 and 10.5...
>> (that is what made me ask rather about possible differences in ld).
> 
> If I understand correctly what nmedit -s is supposed to do, it does two 
> things, one visible and the other one invisible (except when it leads to 
> crashes):
> 
> The visible part is that is transforms external symbols into private 
> externals, thus rendering them inaccessible from the outside.
> 
> The invisible part is that it makes sure that other parts of the library 
> that need this symbol still have access to it. And this is where I 
> suspect Leopard's nmedit fails. But I don't know how to test this 
> hypothesis.
> 

So the symbol is still there on 10.5, but is private extern? That is,
indeed, what nmedit is supposed to do. There are other issues with
nmedit on Leopard (it refuses to strip global coalesced symbols -
whatever they are) that mean that I will have to patch libtool.

I'll try to get around to building xmms.

In the meantime, does something like this work for you in a patchscript
(sed script is all on one line)?

sed -e "[EMAIL PROTECTED]@ -exported_symbols_list
\$output_objdir/\${libname}-symbols.expsym\'@g" < configure > configure.new
mv configure.new configure

Peter
-- 
Peter O'Gorman
http://pogma.com

-
SF.Net email is sponsored by:
Check out the new SourceForge.net Marketplace.
It's the best place to buy or sell services for
just about anything Open Source.
http://sourceforge.net/services/buy/index.php
___
Fink-devel mailing list
Fink-devel@lists.sourceforge.net
http://news.gmane.org/gmane.os.apple.fink.devel


Re: [Fink-devel] [Fink-users] xmms-coreaudio 1.0-3 Plugin Not Showing up in xmms 1.2.10-4

2007-12-06 Thread Peter O'Gorman
Martin Costabel wrote:
> Peter O'Gorman wrote:
> []
>> So the symbol is still there on 10.5, but is private extern? That is,
>> indeed, what nmedit is supposed to do. There are other issues with
>> nmedit on Leopard (it refuses to strip global coalesced symbols -
>> whatever they are) that mean that I will have to patch libtool.
> 
> The nmedit command leaves only one exported symbol, get_oplugin_info.
> This, when called from xmms after loading the plugin via dlopen(), would
> return a structure osx_op, which has osx_about as one of its members. At
> the call of get_oplugin_info(), there appears the undefined symbol error.
> 
> For all other plugins, this works, but they are built from the xmms
> package, and this has the old libtool and does not use nmedit. For these
> plugins, the corresponding *_op and *_about symbols are extern. Only the
> libOSX.so plugin is built separately with the xmms-coreaudio package and
> uses nmedit to turn them into private externs.
> 
>> I'll try to get around to building xmms.
>>
>> In the meantime, does something like this work for you in a patchscript
>> (sed script is all on one line)?
>>
>> sed -e "[EMAIL PROTECTED]@ -exported_symbols_list
>> \$output_objdir/\${libname}-symbols.expsym\'@g" < configure >
>> configure.new
>> mv configure.new configure
> 
> This doesn't match anything in configure. In fact, there are lines with
> nmedit in them, but they have a format that I don't understand at all,
> like:
> 
> module_expsym_cmds_GCJ='sed -e "s,#.*,," -e "s,^[]*,," -e
> "s,^\(..*\),_&," < $export_symbols >
> $output_objdir/${libname}-symbols.expsym${_S_}$CC -bundle $archargs
> $allow_undefined_flag  -o $lib $libobjs
> $deplibs$compiler_flags${_S_}nmedit -s
> $output_objdir/${libname}-symbols.expsym ${lib}'
> 
> This is all one line and there is no space in front of nmedit. I don't
> know how this is transformed into three separate command lines later on,
> but it is.

Ah, ok. configure was built with apple's version of gnu libtool on tiger
or panther.

Change the sed to match ${_S_}nmedit instead of ~nmedit and it should
have some effect on the link line.

The symbols that are private extern should still be available to all the
objects in the library itself, so returning the address of osx_about
should work.

Peter
-- 
Peter O'Gorman
http://pogma.com

-
SF.Net email is sponsored by:
Check out the new SourceForge.net Marketplace.
It's the best place to buy or sell services for
just about anything Open Source.
http://sourceforge.net/services/buy/index.php
___
Fink-devel mailing list
Fink-devel@lists.sourceforge.net
http://news.gmane.org/gmane.os.apple.fink.devel