Thank you for the work regarding this.
Can I understand as follows?
The planned modification for ESpeak includes the fixes
in SynthDriver class methods in synthDriverHandler.py.
Regarding other synthesizers,
if an item of SupportedSettings just uses the base class methods,
such as SynthDriver.RateSetting(),
the change for the item is not necessary for the driver source itself.
--
Takuya Nishimoto
nishimotz@xxxxxxxxx
2015-06-29 10:38 GMT+09:00 Joseph Lee <joseph.lee22590@xxxxxxxxx>:
Hi all, mostly synthesizer driver developers:----------------------------------------------------------------
A copy of this message was sent to development list, but since we have some
synth devs here, I’m pasting it in case you have some additional comments:
As part of adding translatable string contexts, I’m working on a way for
synth drivers and synth settings ring to output the correct name for a
setting while changing the setting via synth settings ring.
Background: A number of translators informed NVDA core contributors that
some messages doesn’t have Gettext contexts, thereby using the same message
in multiple contexts. One of the major example was usage of translatable
string reserved for voice settings – the message with accelerator will be
heard when stepping through voice settings via synth settings ring, like
this:
· Voice Setting message: Pitch (&P)
· Synth Settings message: Pitch (P) 50
This is the case when a language that uses non-Latin characters (East Asian
languages and others) define an English accelerator, and this is used in
synth settings message after the ampersand (accelerator specifier) is
removed.
To remedy this problem, I’m thinking that a new attribute should be added in
synth setting classes to specify strings to be used for synth settings and
voice settings dialog. The plan is as follows:
· Synth setting class: Replace i18nName with two new parameters: a
mandatory displayNameWithaccelerator, and an optional displayName.
· Display name with accelerator is same as i18nName. I18nName will
be preserved for backwards compatibility but might be removed (don’t know
when, it’s up to the lead developers).
· Display name is a string that will be used to output the “real”
(display) name of the setting when stepping through synth settings ring. For
backwards compatibility, this will be optional (default will be None).
· When looking up which output string to use, voice settings will
use displayNameWithAccelerator and synth settings ring will first try
displayName before resorting to stripping ampersand from
displayNameWithAccelerator.
· To help translators, both displayName and
displayNameWithaccelerator will be translatable with different contexts
applied.
Implementation for specific synthesizers:
· ESpeak: Nothing to change, current interface is compatible with
planned modification.
· Other synths: it’s up to you.
Ticket reference:
http://community.nvda-project.org/ticket/5185
Background ticket:
http://community.nvda-project.org/ticket/5100
Please let me know by July 1st if it is fine with you to proceed with the
above plan (for most synth drivers, no need to change anything).
Cheers,
Joseph