Re: [orca-list] [solved] Re: Arch+orca stopped talking to me

[Peter Vágner <pvdeejay gmail com>]

Hello,
This has been fixed.
Arch now provides python2-speechd package based off of python2 
speech-dispatcher bindings available with speech-dispatcher 0.7.1 with 
some patches backported from version 0.8. Speech-dispatcher 0.8 appears 
to only provide python 3 bindings.

Greetings

Peter

On 14.03.2013 12:33, Peter Vágner wrote:
> Hello,
> Thanks for the constructive comments.
> I have filed these two issues hoping for a fix like you have suggested.
>
>
>
> https://bugs.archlinux.org/task/34307
> https://bugs.archlinux.org/task/34308
>
> Greetings
>
> Peter
>
> On 13.03.2013 11:16, Dmitri Paduchikh wrote:
> > Actually, I do not think this is a price you pay for living on the
> > bleeding edge but a price of not having proper support of Orca and
> > Speech Dispatcher in Arch Linux.  It seems that nobody checks the
> > consistency of these packages in testing, hence breakage sneaked into
> > stable unnoticed.
> >
> > It even did not need manual checking.  I believe that breakage could be
> > noticed if the dependency relationship between these two packages was
> > specified more clearly.  I am not very familiar with packaging in Arch
> > Linux, so below reasoning may be imprecise, but I believe it's solvable,
> > may be in some other way.
> >
> > Suppose that package of speech-dispatcher makes the following provision
> > in its package specification:
> >
> > Provides       : python2-speechd
> >
> > and dependencies of orca contain:
> >
> > Depends On     : python2-speechd
> >
> > When preparing the package update of Speech Dispatcher the maintainer
> > would change the name of these virtual package to python-speechd in
> > accordance with naming of Python packages in Arch Linux. Then the
> > dependency of orca becomes broken and this can be caught automatically.
> > As I understand such automatic tests are performed regularly in Arch
> > Linux.
> >
> > And of course you could report about this problem to the Arch devel
> > list.  Sorry if you have done that.
> >
> > Kyle wrote:
> >
> > Ky> The problem was indeed caused by the latest speech-dispatcher 
> > shipping only
> > Ky> with python 3.x bindings. It also affects the RSGames client, 
> > which no
> > Ky> longer speaks, but the main problem was the fact that the GNOME 
> > 3.6 version
> > Ky> of Orca stopped speaking. I received a PKGBUILD file that allowed 
> > me to
> > Ky> upgrade to the latest git master, and doesn't depend on a package 
> > that for
> > Ky> whatever reason is no longer in AUR. I also installed 
> > automake-1.12 from AUR
> > Ky> to satisfy the build dependency on the previous version of 
> > automake. All is
> > Ky> running fine now, except the RSGames client, which I can live 
> > without until
> > Ky> everything is synchronized in the official repos or until RSGames 
> > ships a
> > Ky> Python 3.x version. This is the price I pay for living on the 
> > bleeding edge
> > Ky> <smiles>. I also received PKGBUILDs for the latest at-spi2 and 
> > atk packages,
> > Ky> and will be upgrading those today as well. I look forward to the 
> > great
> > Ky> things in the latest versions of the accessibility stack. I do, 
> > however,
> > Ky> agree with others who have suggested downgrading 
> > speech-dispatcher for now,
> > Ky> as that is certainly the easiest way to keep a speaking system, 
> > especially
> > Ky> if you play RSGames, if you have the older speech-dispatcher package
> > Ky> available. Apparently, I already cleaned mine to save disk space 
> > on the /
> > Ky> partition, so I just upgraded Orca. Thanks for all the help. Oh, 
> > and yes, I
> > Ky> do plan on purchasing a UPS some time in the near future <grin>.
> > Ky> ~Kyle
> > Ky> http://kyle.tk/