[AD] [ alleg-Bugs-1629060 ] 4.2.1 isnt binary compatible with 4.2.0

["SourceForge.net" <noreply@xxxxxxxxxx>]

Bugs item #1629060, was opened at 2007-01-06 00:06
Message generated for change (Comment added) made by thresh_alt
You can respond by visiting: 
https://sourceforge.net/tracker/?func=detail&atid=105665&aid=1629060&group_id=5665

Please note that this message will contain a full copy of the comment thread,
including the initial issue submission, for this request,
not just the latest update.
Category: None
Group: None
Status: Open
Resolution: Works For Me
Priority: 5
Private: No
Submitted By: thresh (thresh_alt)
Assigned to: Nobody/Anonymous (nobody)
Summary: 4.2.1 isnt binary compatible with 4.2.0

Initial Comment:
built allegro-4.2.1 today and i got that with, for example, atanks game:
thresh@xxxxxxxxxx ~ $ /usr/games/atanks
/usr/games/atanks: symbol lookup error: /usr/lib/liballeg.so.4.2: undefined symbol: _blender_trans24

http://pastebin.ca/307011 -- thats the symbols changed from 4.2.0 to 4.2.1.
looks like they all moved to _unsharable.a library
but the programs already compiled with 4.2.0 do not include that new archive library and the shared library doesn't provide them now.


----------------------------------------------------------------------

>Comment By: thresh (thresh_alt)
Date: 2007-04-10 15:32

Message:
Logged In: YES 
user_id=1684677
Originator: YES

OK, i found what was the problem.

Enabling mmx/asm in configure results in placing _blender_trans24 and
other symbols to unsharable.a library thus preventing clients built with
older (built without asm/mmx) allegro to launch.

Should this be considered as a bug? in my opinion, yes, as it seems
allegro contains broken asm code that could not be shared.

----------------------------------------------------------------------

Comment By: thresh (thresh_alt)
Date: 2007-01-07 15:16

Message:
Logged In: YES 
user_id=1684677
Originator: YES

It is ALT Linux Sisyphus unstable branch. Currently, i'm the maintainer of
liballegro in the distribution.
In fact, i only mentioned atanks as an example of going on.
I don't know if it was an erroneus build of 4.2.0 or it's 4.2.1 that's
built strangely. In fact, in the spec file i just bumped the version and
rebuilt the library in current environment.

So, if you all don't have that behaviour i experience here, i should think
that it's something wrong with local builds, so CLOSED/WFM.

Thanks, nice to see such a rapid responce.

----------------------------------------------------------------------

Comment By: Evert Glebbeek (eglebbk)
Date: 2007-01-07 03:19

Message:
Logged In: YES 
user_id=807362
Originator: NO

Just tested this on my own system with one of my own programs: installed
4.2.0, compiled, then upgraded to current SVN. Works fine.
I agree with Peter; it's probably a problem with the atanks binary. It's
possible the package maintainers for the distribution somehow broke
something with a custom patch; what Linux distribution is showing this
problem?

----------------------------------------------------------------------

Comment By: Peter Wang (tjaden)
Date: 2007-01-06 01:44

Message:
Logged In: YES 
user_id=28616
Originator: NO

I can't reproduce it with atanks-2.0-beta.  The included binary didn't
work on my system (Slackware 11.0) with a message "floating point
exception" so I built the source on my system using Allegro 4.2.0, then
upgraded to Allegro 4.2.1.  The binary worked fine.  I believe the original
binary must have been built with a somehow "incorrect" 4.2.0. 

----------------------------------------------------------------------

Comment By: Evert Glebbeek (eglebbk)
Date: 2007-01-06 00:12

Message:
Logged In: YES 
user_id=807362
Originator: NO

Not good.
Thanks for reporting; we'll try fixing it as soon as possible.

----------------------------------------------------------------------

You can respond by visiting: 
https://sourceforge.net/tracker/?func=detail&atid=105665&aid=1629060&group_id=5665