Dear Maintainer,

Translating a short test file using the language pair eng-hbs;

         Mary had a little lamb.
         It's fleece was white as snow.


gave the result

         #Marica #ima @a little @lamb.
         #Free #biti *fleece #biti @white @as #snijeg#.


Problems shown on every word except 'little' was is not translated anyway!


Translating with eng-spa & eng-cat works well.


The command 'apertium -l' shows many Eastern European language pairs with
Serbian, Croatian, Bosnian, Bulgarian, Slovenian, Macedonian.
But the only way into this group from English, is via eng-hbs,
hence my interest.


Regards,
Peter

P.S.
Upstream there is an outstanding three year old pull request.
https://github.com/apertium/apertium-hbs-eng/pulls

I tried patching this in, but it did not fix the problem.

Agreed, direction eng-hbs doesn't look good. Direction hbs-eng looks much
better.

It is clear from
https://github.com/apertium/apertium-hbs-eng/tree/master/texts that this
pair was developed with hbs-eng in mind, and eng-hbs probably shouldn't
have been published. And
https://github.com/apertium/apertium-hbs-eng/blob/master/modes.xml
corroborates this with the comment that eng-hbs was only installed for
debugging.

If you wish to work on the eng-hbs performance, Apertium is always looking
for new contributors (OFTC #apertium). As for this bug, I think best I can
do in the short term is remove the eng-hbs direction.

(The PR (now merged) was just comments, so it couldn't affect quality.)

Thanks for your prompt response. I was wondering myself if the eng-hbs direction was problematic.

Would it be possible to leave it in for devs debugging, but hide it from the 'apertium -l' output?
If not, dropping it seems the best option to me too.

Either would be less confusing to new users.

Thanks again,
Peter