Updates to Github #github

espeak-ng@groups.io Integration <espeak-ng@...>

[espeak-ng:master] New Comment on Issue #761 Initial support for Turkmen language.
By boracasli98:

espeak-ng-turkmen.zip This is the updated version of the Turkmen synthesizer, but I've got a problem: When I want espeak-ng to pronounce "köçelerden", it pronounces it "köçölördän" instead of "köçölördön". I still haven't found where the problem is after minutes of trying to figure it out.

[espeak-ng:master] New Comment on Issue #784 Error processing file 'tests/phontab': No such file or directory.
By rhdunn:

  TEST      tests/api.test
Error processing file 'tests/phontab': No such file or directory.
Error processing file 'tests/phontab': No such file or directory.
Cannot set intonation: language not set, or is invalid.

These are expected messages. That file is testing the espeak API calls with various parameters to check handling with nulls, etc. In the above case, it is testing the behaviour when passed a directory that does not contain espeak voices. If it was an issue, it would have failed at that point.

testing lv
< 060b6f94cd505e5c3e649d94c41856d801c0c695
> e209267d9c35a612f24906da39b2927b4e8d2a87

This is the reason why make check is failing. These tests are testing the phoneme to audio logic for the different languages, then checking that the hash of the generated version matches the expected/reference version. I'm not sure why it is generating different output here -- the languages prior to this are generating the correct output, so it could be a code path that is performed by the lv phoneme tests but not the others.

First, I would investigate why your environment doesn't support unicode characters. There shouldn't be � in the output:

That happens e.g. if the font does not support the characters. Those are only for some emoji characters and some languages -- other emoji and languages (e.g. Hindi) are displaying correctly.

Join espeak-ng@groups.io to automatically receive all group messages.