A new, extensible speech dictionary type for nvda

[aleksey_s]

sometimes one dictionary must be applied to more than one voices or even synthesizers. it is possible when user wants to handle voices of one manufacturer or one language.
So nvda must have functionality to make such dictionaries.
i offer to implement a new type of speech dicts: smart dictionary. this one besides list of entries will have regular expression pattern, which must match with current voice and synthesizer for this dictionary to be active. for users, who are unfamiiliar with regular expressions, nvda will offer dialog with all voices of all available synthesizers, where user can check items and regular expression will be constructed.
i have allready implemented smart dictionaries, for now i am working on gui for manipulate its.

here is a log of my local bzr branch:

2263 Алексей Садовой 2008-10-31

*core.py: fixed some inconsistend logging.

2264 Алексей Садовой 2008-11-01

• speech dicts: implemented a new type of dictionary, smart dictionary. this one can be applied to all synthesizers and voices, if some regexpr matches synthesizer-voice string. we can have many such dictionaries at a time loaded and processed.
• some fixes to existing speech dictionaries implementation

2265 Алексей Садовой 2008-11-02

speechDictHandler.py:

• now dictionary files can have native characters in names.
• Smart dictionaries now have name. file name will be constructed on this name.

keyboardHandler.py:

• fixed one typo

[aleksey_s]

(In #212) functionality, requested in this ticket is implemented in #219

[aleksey_s]

initial GUI support added. now only dialog with multiselect list of all voices and synthesizers is required. diff.txt - bzr merge directive.

[aleksey_s]

maked changes to reflect last revision

[aleksey_s]

bzr merge-directive

[aleksey_s]

updated feature to match main branch. Core developers, are you going to review it?..

[ragb]

Another related idea is to have application specific dictionaries. This was discuced in NVDA's portuguese list, it seems some users use it extensively in JAWS and request this feature in NVDA.
Should I fill another ticket for this suggestion?

[aleksey_s]

for further progress, see bzr branch at http://bzr.nvaccess.org/nvda/smartDictionaries