[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: OS X Server, Or Something Else?



Bart,


No problem ;), I'm just not sure this is "the fix".


As for debugging, try to come up with the minimal set of protocol commands from ./mac that exhibits the latency (this goes for anyone who wants to help out).


Dig in further from there.


I'll upgrade tomorrow when I grab my Mac laptop.


Thanks all,

David



On Sun, Oct 27, 2013 at 4:04 PM, Bart Bunting <bart@bunting.net.au> wrote:
David,

Your version is much cleaner.  Sorry was just needing a quick fix and
thought everyone else might also.  I'll confirm with you next time before
pushing.

I think this patch still doesn't solve the issue entirely though.

Although everything is being spoken as expected as far as I can tell at
the moment, there is still intermittent sluggishness.

Not sure where it's comeing from or how to debug.  It's not consistently
reproducible.  Some times when arrowing over lines there is a pause of
at least half a second before speech starts.  The same line then works
fine.



Kind regards

David Tseng <davidtseng@gmail.com> writes:

> Bart,
>
> Please ping me for patches of that type. Checked in some cleanup.
>
> Thanks,
> David
>
>
> On Sun, Oct 27, 2013 at 2:43 PM, Bart Bunting <bart@bunting.net.au> wrote:
>
>> Morning everyone,
>>
>> Sorry was away from the computer this weekend.
>>
>> I have implemented David's suggestion and made the speechlock part of
>> the code dependent on OSX version.
>>
>> It is only executed if OSX version is less than 10.9.
>>
>> I've checked this in to the e-mac-speak svn repo.
>>
>> As I've only just made the changes it hasn't had much testing.  Having
>> said that it certainly appears to help things here.
>>
>> I haven't noticed the voicelock issues that Haden mentions as yet.  They
>> may still be there though.
>>
>> If people could report success or otherwise on both 10.9 and earlier
>> that would be appreciated.
>>
>>
>> Kind regards
>> Bart
>>
>> David Tseng <davidtseng@gmail.com> writes:
>>
>> > To reiterate...anyone doing serious work on a Mac should probably refrain
>> > from upgrading from 10.8 to 10.9 at least until 10.9.1.
>> >
>> > - David
>> >
>> >
>> > On Sun, Oct 27, 2013 at 10:54 AM, Haden Pike <haden.pike@gmail.com>
>> wrote:
>> >
>> >> Commenting out everything from speechLock.acquire() to the final
>> >> speechLock.release() at the top of the processSpeechQueue() function
>> >> partially solves the problem.  Emacspeak reads everything it should, but
>> >> the server still has some strange interactions with voice-lock.  If
>> >> switching buffers fast, for instance, the server won’t say the name, of
>> the
>> >> buffer.
>> >>
>> >> Nevertheless, I’ve got the most important functionality of Emacspeak
>> back.
>> >> Haden
>> >>
>> >> On Oct 27, 2013, at 1:16 PM, David Tseng <davidtseng@gmail.com> wrote:
>> >>
>> >> While implementing the server a while ago, I discovered a ref counting
>> bug
>> >> in the Mac tts implementation. If I recall correctly, end callbacks were
>> >> not getting sent properly as a result (i.e. end callbacks of the
>> previous
>> >> utterance were sent *after* start callbacks of the current utterance).
>> My
>> >> guess is that 10.9 has corrected the issue.
>> >>
>> >> I've not upgraded to Mavericks since I use my Mac for production work,
>> but
>> >> I can take a look next weekend on a machine with Mavericks. The fix will
>> >> likely change/remove the extra acquire/release on the lock controlling
>> >> access to the speech queue for Mavericks only since the bug obviously
>> still
>> >> exists pre-Mavericks.
>> >>
>> >>
>> >> - David
>> >>
>> >>
>> >> On Sun, Oct 27, 2013 at 10:04 AM, Haden Pike <haden.pike@gmail.com>
>> wrote:
>> >>
>> >>> > > Maybe starting the application that is stored in environment
>> variable
>> >>> DTK_PROGRAM is enough to run the server.
>> >>> >
>> >>> > It is.
>> >>>
>> >>> > This may help in determining the cause of the problem.
>> >>>
>> >>> I tried that and these results might help diagnose the problem.
>> >>>
>> >>> I ran ./mac from the Emacspeak servers directory.  After receiving the
>> >>> “Server ready” message, I typed the following:
>> >>>
>> >>> q this is a test
>> >>> q of the mac speech server.
>> >>>
>> >>> Typing d gave me “this is a test”.  Typing d again gave me the second
>> bit
>> >>> of queued text: “of the Mac speech server.
>> >>>
>> >>> Based on this, it looks like the speech queue is not being processed
>> >>> correctly.  I don’t know the Cocoa technologies well, so am not sure
>> where
>> >>> to start fixing this.
>> >>> Haden
>> >>>
>> >>>
>> >>>
>> -----------------------------------------------------------------------------
>> >>> To unsubscribe from the emacspeak list or change your address on the
>> >>> emacspeak list send mail to "emacspeak-request@cs.vassar.edu" with a
>> >>> subject of "unsubscribe" or "help".
>> >>>
>> >>>
>> >>
>> >>
>> Bart
>> --
>>
>>
>> Kind regards
>>
>> Bart
>>
Bart
--


Kind regards

Bart

-----------------------------------------------------------------------------
To unsubscribe from the emacspeak list or change your address on the
emacspeak list send mail to "emacspeak-request@cs.vassar.edu" with a
subject of "unsubscribe" or "help".




If you have questions about this archive or had problems using it, please send mail to:

priestdo@cs.vassar.edu No Soliciting!

Emacspeak List Archive | 2010 | 2009 | 2008 | 2007 | 2006 | 2005 | 2004 | 2003 | 2002 | 2001 | 2000 | 1999 | 1998 | Pre 1998

Emacspeak Files | Emacspeak Blog | Search the archive