Re: PATCH: zle input changes for Unicode

[Bart Schaefer <schaefer@brasslantern.com>]

On Feb 18,  1:50pm, Peter Stephenson wrote:
}
} Please tell me if you think I'm doing something wrong

My only wish is that you might have tagged the repository before the
commit (e.g. "cvs tag pre-unicode-input" or some such).  I'm not sure
I have an uncorrupted sandbox from which I can do so.  Maybe a checkout
by date ...

} I moved the definitions from system.h to the more appropriate zle.h.

I heartily approve.

} Note also that lastchar, and its new companion lastchar_wide, are now
} set by the routines that actually read the character.  This is obviously
} neater and it was hard to see from the code if there was a good reason
} why it wasn't done before, but I strongly suspect there wasn't.

The obvious question is whether there is any case where getkey() was
called but lastchar was not updated.  Scanning through the patch I see
two places where that might have been true -- in deltochar() and in
getkeybuf() -- and one where it appears really to be true, that is, in
getzlequery().  But maybe the callers fix up something thereafter, I
didn't look carefully at the whole context.

} Note that I've renamed some functions with "getkey" to "getbyte" to
} reflect the function better.

I generally think this sort of thing is a good idea, if for no other
reason than that it means the compiler will point a lot of the places
that you might need to do other special handling.

Misc. remarks:

Is last_widechar_valid set to 1 too soon in getrestchar()?

Sometimes you assign 1 to last_widechar_valid, and sometimes TRUE.

Nothing else jumps out at me.