[darcs-users] Re: darcs-users Digest, Vol 19, Issue 13

[rmuschall]

Jamie Webb <j at jmawebb.cjb.net> wrote:

> Actually, the same logic should also work for at least Big5,
> Shift-JIS and EUC. UTF-16 is more tricky, though I imagine Darcs
> could check for a BOM if anyone felt the need.

> What's the alternative? Make Darcs encoding-aware? I suspect that

IMHO yes.  I.e.

  darcs add foo.c

tries to check in foo.c as ASCII if it can verify that it is ASCII,
otherwise binary in the old way (in order not to break old behavior).

  darcs add japanese-doc.txt --content-type=text/prose --content-transfer-encoding=UTF-8

does smart things with UTF-8 "prose" (i.e. reformatting is allowed --
text/plain would disallow that, so we can manage poetry collections
or even program sources ;-)).

The first command would be equivalent to

  darcs add foo.c --content-type=text/plain --content-transfer-encoding=ASCII

If somebody has too much time, he could implement a smart handling
for Openoffice documents (zipped directory trees full of XML files
using UTF-8).

Changing the content-type or content-transfer-encoding would need to
be treated like a patch.

Ralf