Re: [Gnash-dev] Revert the subshape change until regressions are fixed?

[Petter Reinholdtsen]

Gah, trying again, this time with the CC still in place.  It was lost
in editing by mistake.

On Thu, Jan 16, 2014 at 01:30:19PM +0100, Petter Reinholdtsen wrote:
> Cc to Bastiaan Jacques, to make sure the author of the subshape
> changes are aware of my proposal to revert the change until it is no
> longer introducing regressions.
> 
> [Petter Reinholdtsen 2014-01-10]
> > I had a look at http://sv.gnu.org/bugs/?40315 and the comments there,
> > and the indication that the subshape changes was causing the problem.
> > The subshape changes also seem to be behind the problems in bug
> > #39989, and I decided to try to see if I could revert the subshape
> > changes and get the flash files working.
> > 
> > I managed to do this using the git history, and the patch to remove
> > the code is attached.  It might contain a bit more than it should, so
> > please review and see if parts of it is safe to apply.
> > 
> > Perhaps the changes in this patch should be applied to master and the
> > problematic changes moved to a branch until these bugs in the subshape
> > change are fixed?  It should be easier to fix now that two test cases
> > are available.  And perhaps some test suite tests could be created to
> > detect these problems automatically?
> 
> Bastiaan, the subshape changes introduced crash bugs when using two
> flash files.  These worked in the last release.  Do you have any idea
> how to fix this, or is it better to revert the change (and perhaps put
> it into a separate branch) and integrate it into the master branch
> when the regressions are fixed?
> 
> I tried to find my way around the changes but currently lack the
> mental bandwidth to wrap my head around it, and am unable to fix the
> crash problems myself. :(
> 
> The complete mailing list thread can be found via
> <URL: http://lists.gnu.org/archive/html/gnash-dev/2014-01/msg00004.html >
> if you are not currently subscribed.

-- 
Happy hacking
Petter Reinholdtsen