GNOME Bugzilla – Bug 626032
substitutions from the pronunciation dictionary are not being applied when performing say all
Last modified: 2012-08-07 14:51:12 UTC
From Paul on the mailing list: http://mail.gnome.org/archives/orca-list/2010-August/msg00021.html -- I noticed that Orca isn't substituting words that I have alternate pronunciations for in my pronunciation dictionary. The words are being substituted if I read by word or line but not during say-all. It doesn't seem to matter if say-all is set to line or sentance and I've verified the problem in both gedit and OO Writer. ---
*** Bug 626035 has been marked as a duplicate of this bug. ***
Created attachment 167129 [details] [review] possible patch for bgo#626032 - substitutions from the pronunciation dictionary are not being applied when performing say all Hi Joanie, This seems to solve the issue for me, for sd and otts. I think it should work for gnomespeech, could you please try it? Does it look ok otherwise? Pylinted to 10. Thanks.
Hi Jon, I just tried the patch with both speech-dispatcher and gnome-speech and it seems to be working correctly.
Works fine using OTTS. Thanks.
Created attachment 167145 [details] [review] patch for bgo#626032 - substitutions from the pronunciation dictionary are not being applied when performing say all Thanks Paul and Jose for your testing. Improvements to previous patch: created speechserver.preprocess() which does both the pronunciation dictionary corrections, and speak multi case as words substitutions for us. Now its only a functioncall in those places which require us to check for pronunciation substitutions. Future preprocessing happens in speechserver.preprocess, and it is done for both normal speaking and for when we are performing sayall, which is cleaner. Pylinted to 10. Thank you
Thanks for doing this Jon. From a quick glance, it seems okay. However given where the change is being made, it should be regression tested. In addition, I'm trying to get us back into the habit of providing a new regression test for each feature/fix where it makes sense to do so, namely cases where: * Writing such a test is feasible. * The existing regression tests do not cover the situation being fixed/implemented. In terms of this bug, the second instance is definitely the case. And I believe that the first is as well. Therefore, would you mind writing at least one regression test for this fix? Thanks!
*** Bug 680010 has been marked as a duplicate of this bug. ***
Joanie, If i brought this up to date would you mind doing the test harness bit? I am sorry I havent got around to setting up a test harness environment/got round to writing some test cases. If so, what branches are active these days/should i provide patch against? Thanks.
I had forgotten about this. Thanks for the reminder. For now, going with the simple/safe fix. master: http://git.gnome.org/browse/orca/commit/?id=bf73251dba974529d4badef7ef13c2d972420a2e gnome-3-6: http://git.gnome.org/browse/orca/commit/?h=gnome-3-6&id=80ea68c387961a4960b6e9a3587ded4b21c2baa5