[Grml] making hardware speech work

[John Heim]

I was wondering if the grml team would be interested in making hardware
speech work in grml. I have a kernel patch that fixes a problem in the linux
screen reader, speakup, to make it work with certain hardware speech
synthesizers. I don't know how to get it into the actual kernel code and I
haven't even tried yet. But the patch file can be downloaded here:

http://www.math.wisc.edu/~jheim/downloads/patch-2012-03-06.patch

Then you cd to your linux source dir and do this:
patch -i patch-2012-03-06.patch drivers/staging/speakup/serialio.c"

If you do that, you could also include a udev rules file that starts speech
automatically if certain hardware synths are connected at boot time. That
file is here:

http://www.math.wisc.edu/~jheim/downloads/10-speakup.rules

[Michael Prokop]

Hi John,

Post by John Heim
I was wondering if the grml team would be interested in making
hardware speech work in grml.

If we get feedback and patches like yours: sure :)

Post by John Heim
I have a kernel patch that fixes a problem in the linux screen
reader, speakup, to make it work with certain hardware speech
synthesizers. I don't know how to get it into the actual kernel
code and I haven't even tried yet. But the patch file can be
http://www.math.wisc.edu/~jheim/downloads/patch-2012-03-06.patch

Thanks for the patch, could you by any chance try to get it
applied upstream? We try to stay as close to upstream and Debian
kernel as possible and we'd like to see improvements taking place
there so it's not just the Grml users who can benefit from it. :)

Post by John Heim
patch -i patch-2012-03-06.patch drivers/staging/speakup/serialio.c"
If you do that, you could also include a udev rules file that starts
speech automatically if certain hardware synths are connected at boot
http://www.math.wisc.edu/~jheim/downloads/10-speakup.rules

Hm, this might be interesting also to Debian in general?
Maybe it would fit into the speakup-tools package. John, are you
familiar with reporting a bug within Debian using reportbug?

Thanks, John.

regards,
-mika-

[John Heim]

Post by Michael Prokop

Post by John Heim
reader, speakup, to make it work with certain hardware speech
synthesizers. I don't know how to get it into the actual kernel
code and I haven't even tried yet. But the patch file can be
http://www.math.wisc.edu/~jheim/downloads/patch-2012-03-06.patch

Thanks for the patch, could you by any chance try to get it
applied upstream? We try to stay as close to upstream and Debian
kernel as possible and we'd like to see improvements taking place
there so it's not just the Grml users who can benefit from it. :)

The problem is that the kernel developers and the speakup developers cannot
agree on how to fix the bug. Of course, I am neither a kernel developer or
one of the speakup developers. The kernel developers want speakup rewritten
but the speakup developers say that if they rewrote it the way the kernel
people tell them to, it would lose critical functionality. For example, the
very first thing the kernel developers said when I brought this up on the
kernel list was that speakup should be a run-time system. Well, that showed
an incredible lack of understanding of the purpose of speakup itself since
it needs to be loaded at boot time in order to get boot messages. So some
guys from the kernel development team and some from speakup got into a
little debate, which, BTW, I say the speakup people won. On the other hand,
I have to admit the speakup code is, well, it has problems.

I am still hoping to get the patch, perhaps a modified version of it,
accepted into the kernel code. I got a run around just asking how it could
be done though. If you examine my patch, all it does is ignore the return
code from a kernel function call. But there is absolutely no way to get a
successful return code from the function call. When I pointed this out, I
got no response. Very frustrating.

Post by Michael Prokop
Maybe it would fit into the speakup-tools package. John, are you
familiar with reporting a bug within Debian using reportbug?

Yeah, But I already emailed the developer of that package and he said maybe
it could go into the udev package. So I emailed the udev developer and he
suggested I try to get it into speakup-tools. Well, actually, he didn't
know the first thing about speakup but that's what he meant. Honestly, I
think you guys on the grml team are victims of your own reasonableness. I
come here with my suggestions because you folks seem to be the only
reasonable people on the internet.

Alright... While thinking about this over the weekend, I had decided to just
give up. But after composing this message, I think I will indeed give it
another try with the kernel people. And you needn't bother with the udev
stuff since that's useless w/o the kernel patch.