This reverts commit 02783150cd.

This reverts commit aa765c7680.

Avoid converting input data to Float32Array

It is perhaps more autoconfy to have AM_CONDITIONAL evaluate a variable
rather than call it with what are essentially constants representing
either true or false. It might be even more autoconfy to use
AC_CHECK_PROG here instead; I don't know if there's a good reason why
that wasn't already done here.

Previously, if configure determined that gradle could not be found, the
summary at the end of the configure output would nevertheless print the
name or path of the nonexistent gradle executable that it tried to find.
Now it clears that variable.

Previously, the configure script would only successfully determine that
gradle exists if the user specified an absolute path to the executable.
It would fail if the user specified only an executable name or if the
default value of "gradle" were used. Now it should succeed in either
situation.

This will make it easier to set a default voice other than
English. This is important for cases when a language will fall back to
the default voice.

Some references to L('e', 'n') still need to be changed.

Found by parsing the results from this bash script:
for define in $(grep define *.h |  awk'{print $2}');
do if ! grep -q $define *.c
then
        echo "$define not found";
fi
done

This is unnecessary since the string is already null terminated in the
V_LANGUAGE switch case. sscanf creates a null terminated string in
language_name, which is then copied to voice_language by strcpy (keeping
the \0).

he: Add ךְ and ךָ

It's probably unnecessary code and could be removed. However, it would
break tests. This is a temporary solution.

The solution is to do it once in VoiceReset() and then if the "formant"
keyword changes voice->width.

Language option "stressOpt" already handles separate flags instead of a
single integer. There's no reason to have two ways of setting the same
option.

This is to make sure "stressopt" language option works. Current tests
would fail if it didn't.

language option "numbers" could be changed to use the same function, but
it needs tests.

ReadNumbers8) parses an unknown number of numbers from language files.
Used mostly for setting flags.
Used mostly for setting flags.

It is called multiple times in LoadVoice().

SetSpeed is needed when "speed" keyword changes voice->speed_percent.

1: Added missing voice variants in the NVDA category including Iven 4, Steph, Steph 2 and Steph 3.
2: Completed the Persian translation.

Substitute using existing defined done of ReadableStreamDefaultReader.read() for value.length < this.channelDataLength &&
(await this.inputReader.read()).done to verify all input stream bytes are written

- Verify last bytes of input stream are written to audio output stream before disconnecting audio nodes
- Start MediaRecorder before first AudioData is written to audio output stream
- Include error handling in disconnect() and abortHandler for abort() called immediately following start() for case of close() called on a closed stream
- Check if data set in recorder dataavailable event is defined before calling data.arrayBuffer()

182aba4cc started calling Norwegian nb instead of no. The result is that
code in tr_languages.c was never run.

Added tests are copied from existing tests from the same language family
or English. The tested phonemes should be adapted for better test
coverage.