Hi michael,
Do we have any compilation errors in the espeak NG itself? i.e. some phonemedata cannot be generated properly?
if not, i suggest merging it to master for further testing of all slavic language branches...

To @zstanecic, @nidza07 and k-kolev1985:
Please test the following try build of NVDA with latest espeak-ng master: https://ci.appveyor.com/api/buildjobs/e3c2cb43cjtd26pf/artifacts/output%2Fnvda_snapshot_try-latestEspeak-15686%2C0f0916a7.exe
Let me know if there are still regressions in eSpeak-ng from what is in NVDA 2018.2.

Hi michael,
I can confirm, that all bugs are gone in this build of espeak NG.
I have tested bulgarian from the sample text of stoyan stoyanov, Gramatika na balgarskiot ezik
for macedonian i used a sample text from the website of macedonian radio television. i tested the simple characters, too!
the macedonian, as well as bulgarian, serbian and bosnian are now fixed.all palatalisations which were present are now gone.
@michaelDCurran
best,
Zvonimir

btw, on new thing for macedonian, is added transliterated text reading
so all text romanized in the british manner, but written in macedonian will be read

@michaelDCurran I've tested a little bit and I can confirm that the wrong pronunciations in bulgarian are fixed in the test build.