WIP: GHCJS support

[bgamari]

This is a first attempt at adding support for GHCJS, addressing #5. Of the two approaches I opted to reimplement text-short's interface on top of Data.JSString.

There currently seem to be some inconsistencies between Javascript's treatment of surrogates and that expected by text-short.

[hvr]

There currently seem to be some inconsistencies between Javascript's treatment of surrogates

Can you elaborate?

The other obvious difference is probably that there won't be a O(0) conversion from ShortText to ShortByteString :-)

And btw, isn't JSString UTF16-backed internally?

[bgamari]

There currently seem to be some inconsistencies between Javascript's treatment of surrogates

Can you elaborate?

Admittedly I can't add much more detail. I implemented this in a hurry since we needed to quickly port a project using text-short to GHCJS. I ran the testsuite and saw a number of failures, all apparently involving high codepoints, but didn't investigate any further since the patch is working fine for our needs.

I thought it would be best that I open an MR here so that others might benefit from it.

The other obvious difference is probably that there won't be a O(0) conversion from ShortText to ShortByteString :-)

And btw, isn't JSString UTF16-backed internally?

JSString is backed by a JavaScript string and AFAIK the ECMAScript specification doesn't specify an encoding. I did a quick search which suggests that at very least V8 uses UTF-8. On the other hand, it looks like SpiderMonkey uses UTF-16.

[phadej]

@bgamari what is left here? Will continue to work on this?