Tiger-speech needs Apple's speech engine, which this add-on does
not ship.

Put the contents of a Mac OS X 10.4 (Tiger) install here, so that this folder
contains:

    Speech\Synthesizers\MacinTalk.SpeechSynthesizer\
    Speech\Voices\<name>.SpeechVoice\
    SpeechDictionary.framework\Versions\A\

Dropping the extracted folder in whole, one level down, works too.

The easiest way to produce it is the extractor in the project repository:

    py -3 tools\extract_tiger.py "Mac OS X 10.4 Tiger.iso"

It needs 7-Zip installed, reads your own installer image, and writes straight
into this folder. You need an Intel Tiger image; the PowerPC discs carry the
same voices but a PowerPC engine, which cannot run here.

The extractor is a single Python file. It is not bundled with the add-on
because reading an ISO or DMG needs 7-Zip, which cannot be shipped inside an
NVDA add-on -- so you download and run it yourself:

    https://github.com/tgeczy/panthera-speech/blob/main/tiger/tools/extract_tiger.py

It needs Python 3.8 or newer installed (tested on 3.13), and 7-Zip.

If you would rather keep the engine on another drive, put its full path into a
file called tigerspeech-data.txt in the configuration folder instead.

If that file appeared on its own after an upgrade, it is a note left behind
when the engine folder moved into macintalk, so that going back to an older
version of this add-on still finds it. Deleting it is safe unless you do that.

Delete the file called "do-not-ask" here if you told NVDA to stop reminding
you and would like the reminder back.
