Re: [Gnash-dev] A few more build problem patches

[Benjamin Wolsey]

Am Donnerstag, den 12.02.2009, 05:01 -0800 schrieb John Gilmore:
> I found a workaround to my earlier problem; the patch:
> 
> +        std::streamoff crap = ret;   // C++ is badly broken
> +        _stream->seek(_stream->tell() - crap);
>          return false;
>      }
>      

Was already fixed, I hope.

> Now the compile is failing in the kde support because the
> step and jump methods were removed.  I diked them out in both 
> kde.cpp and in klash3.moc (kdesup.h), with this set of changes:

Also fixed in trunk now.

> The KDE GUI doesn't appear to include the new media-saving preferences
> (nor indeed any other preference-setting capability), but I didn't try
> to fix that, never having programmed a KDE GUI before.
> 

The KDE3 GUI is unmaintained. We don't have enough nearly people to work
on all GUIs, so the only ones that don't rely on contributed patches are
GTK and (largely thanks to John Wimer) KDE4. That has a preferences menu
as of a few days ago, including for saving media.

> In the Ming tests I got 101 unexpected failures (I have the Ubuntu
> ming, not the "very latest unpackaged" ming).  Those are appended below.
> There are also huge numbers of errors (and "make check" actually halts
> with an exit code) in astests-*-Runner, many XPASSes as well as FAILs.
> I presume these are well known.
> 

There should be no FAILs at all in the testsuite, but it does depend on
ming version and on having the correct gstreamer packages installed. The
packaged version in most distros is too old to be very useful, and
although there were some attempts to fix support for it, it would make
the testsuite into an awful mess.

> I also fixed a syntax error in testsuite/actionscript.all/XML.as:
> 
>       check(this.childNodes.hasOwnProperty('length'));
> -     check(this.childNodes[0] === this.firstChild);
> -     check(this.childNodes[0] === this.lastChild);
> +     check(this.childNodes[0] == this.firstChild);
> +     check(this.childNodes[0] == this.lastChild);
>  

That's not a syntax error, but it isn't supported by older ming
versions. Working round these errors means not only #ifdefs for the
failing code but also providing different test counts, for each SWF
version, as well as for each ming version, and isn't feasible for a
testsuite.

> With these changes, it builds and installs (but doesn't make it all the
> way through "make check") on a recent x86 Ubuntu Hardy.  I can run it,
> but Firefox 3.0.5 never finds the plugin.  (It found the 0.8.4 plugin,
> which I removed with "apt-get remove mozilla-plugin-gnash", but it
> doesn't find the one I installed in my home dir with "make install-plugin".)
> 

Does it have the correct permissions?

--
The current release of Gnash is 0.8.4
http://www.gnu.org/software/gnash/

Benjamin Wolsey, Software Developer - http://benjaminwolsey.de

signature.asc
Description: Dies ist ein digital signierter Nachrichtenteil