The implications of this are large for using eSpeak-ng as a phonemizer. It means that it's now possible to get accurate mappings between the phonemized word and the original word (up till now it was pretty much impossible). This is huge for many applications that use it as a side-tool.

Attempt to fix issue #1732 and add more sane symbols.

I did some major improvements to the words balance and garage and a
minor improvement to the word behilflich.
This pull request supersedes the pull request
https://github.com/espeak-ng/espeak-ng/pull/1875 as my pronunciation of
the word garage sounds more natural.

This PR adds support for **Shavian**, a phonemic alphabet for English:
https://en.wikipedia.org/wiki/Shavian_alphabet (Unicode
[U+10450–U+1047F](https://www.unicode.org/charts/PDF/U10450.pdf); ISO
15924 - Shaw)
It correctly pronounces most English words, except sometimes it
misplaces the word stress. All checks are passing.
There are many Shavian communities online; I wanted to join in without
excluding people using screen readers.
If needed, this website has more resources: https://shavian.info/
![Screenshot of GNOME terminal, showing the difference between phonemes
in "This is an example sentence in the Shavian alphabet", and the same
sentence but for
Latin](https://github.com/espeak-ng/espeak-ng/assets/113068485/75d43247-1b70-4c16-8e55-6a9ffab7784d)

Continuation of #1818 due to an issue in a rebase

- Uncomment and translate untranslated Czech emojis

Signed-off-by: Michal Bida <[email protected]>

This PR adds, in addition to the existing windows CMake pipeline, a
Github Action to build release and debug installers for Windows. It
includes the espeak-ng-data path fix from
https://github.com/espeak-ng/espeak-ng/issues/1705. ~~Unfortunately, in
order to build successfully, it removes Arabic and Faroese from the
Windows installer for the reasons listed in
https://github.com/espeak-ng/espeak-ng/issues/1774~~ (which
incidentally, I think, demonstrates the value of having this as a
separate CI pipeline)

This reverts commit 56a4c82d99.

`#if PLATFORM_WINDOWS` doesn't work for some reason.

Fix suggested by @beqabeqa473

`#if PLATFORM_WINDOWS` doesn't work for some reason.

Fix suggested by @beqabeqa473

Added many Hebrew rules. tested by playing the audio and hear the sound.

https://github.com/espeak-ng/espeak-ng/issues/1980 |
https://github.com/espeak-ng/espeak-ng/issues/1979 |
https://github.com/espeak-ng/espeak-ng/issues/1897 |
https://github.com/espeak-ng/espeak-ng/issues/1143 |
https://github.com/espeak-ng/espeak-ng/issues/1009


Few issues with hebrew:

----
I can't match a word contains quotes. eg.

ארה״ב
ארהב
it should be aRtsot habRit

----
I can't match a word while ignoring the diacritics.
eg. imagine that ארהב came as with diacritics
אִרְהָב
I want to ignore the diacritics. if there's a match it should be like
without.

---

English words doesn't sound good in Hebrew at all. how should I improve
it?

he rules

he rules

he rules

more rules

add rules

This fixes #1011 by integrating mbrola voices details in ssml processing

So that SelectVoice can select an mbrola language.

Fixes #1011

Otherwise we won't ever be able to select an mbrola voice in ssml
processing.

We should initialize current_voice_id so that ssml processing knows what
voice we currently have, and not try to change it unless really needed.

When e.g. using mbrola voices, the sample rate may change. When the
application plays audio itself, it needs to know it, and
espeak_ng_GetSampleRate does not report that.

When e.g. using mbrola voices, the sample rate may change. When the
application plays audio itself, it needs to know it, and
espeak_ng_GetSampleRate does not report that.

1. Add a list of names with the correct pronunciations in Portuguese:
5b725ddc94.

The names in the list are not pronounced correctly by espeak-ng when
they are without accents/diacritics. For example, Mario is pronounced
[maɾˈiʊ] instead of [mˈaɾjʊ].


2. Add a list of foreign names adapted to Portuguese:
88513a5564.

The pronunciations of the added names are simple adaptations, i.e. the
foreign names were "aportuguesados" (adapted to Portuguese).

3. Remove uncommon names from Portuguese dict source:
e11ad2f11f

A few days ago, more than 800 names were added, but some of them are not
as common as I thought. Names that had a frequency of less than 1,000 in
the 2010 Brazilian Census were removed.