I'm home, and now have a big-enough screen to see both what I'm typing and
the message to which I reply.
Kevin J. McCarthy writes:
- Added the unicode logic to the imap_unmunge_mbox_name() call. (I'm
assuming we shouldn't be trying to utf7 decode strings from the server
after we've enabled utf8. If this is incorrect, please let me know.)
It shouldn't matter, or rather, it should matter for mailbox names that
shouldn't occur. But I like your way better. It's cleaner.
- I changed the parameter to pass in idata instead of the unicode flag.
(This was just personal preference: it seemed better to encapsulate the
decision of what to look at inside the function, in case we need to look
at other flags or capabilities in the future).
Yes, personal preference, and IMO the maintainer's personal preference
counts more than the submitter's.
- The imap_utf7_encode/decode were also performing conversion from
the local Charset to UTF-8 before doing the UTF-7 conversion. This
conversion needs to be performed in either case.
Therefore I renamed them to imap_utf_encode/decode, threaded through
idata, and put the "if" for utf-7 inside of them. That way the
Charset conversion will always occur.
Bit uncertain about the wisdom of trying to support unicode mail on a
terminal that doesn't use unicode... but I guess your change is either
neutral or an improvement, so fine in either case.
- Changed cmd_parse_enabled() to use ascii_strncasecmp() instead of
strcasecmp(). (See "BEWARE" in the source tree.)
I noticed the decode is taking place twice after a LIST command while
browsing. I haven't tracked down why yet.
I noticed that too, but I didn't want to digress in the patch... and anyway
tracking it looked like real work.
Arnt
