[sword-devel] Anger on mailing list (was Re: AmTract Encoding)

[Chris Little]

peter wrote:
> On a practical note, would it be useful to convert the remaining non
> UTF-8 modules into UTF-8? If so, I am happy to do the actual work over
> the next few weeks. I would probably need some advise, but could do the
> actual foot work. I am not at all suggesting this should be done, I have
> no clue, whatsoever. I presume though that users of very old modules and
> very old software will have a large degree of overlap. People who want
> to write new front ends and do not want to support old modules could
> then probably simply add a test ?UTF-8 and advise the user to update the
> module. People who can try out new software will have access to new
> modules update too.

There's no real purpose to converting the existing modules. We have to 
support the old modules anyway because we don't want to force upgrades.

Supporting all module encodings is pretty trivial. We're consistent 
about UTF-8 modules being UTF-8 throughout (in the content, the keys, 
and the .conf). And likewise CP1252 modules are consistently so. (I 
would classify the key encoding of GerLutherpredigten as a bug that 
should be fixed by re-releasing the module, not in code.) And we provide 
a method in the library (in addition to the one provided by ICU) to 
convert CP1252 to UTF-8. And if one wishes to work in UTF-16 or UTF-32, 
I think we provide routes to those from UTF-8 in the library (in 
addition to what ICU provides).

--Chris