Hi Ben,
Ben Finney wrote:
> On 19-Oct-2010, Axel Beckert wrote:
> > This is probably cause by the code at the end of the patch causes
> > that the order of digraphs are no more interchangeable.
>
> That's necessary, since RFC1345 defines many mnemonics that are only
> distinguished from others by the sequence of characters.
>
> > Without this hunk, at least the ":a" digraph is working again.
>
> Without order dependency, there's no way for the input processor to
> distinguish the following pairs:
>
> {'*', 'X', 0x00d7}, /* MULTIPLICATION SIGN */
> {'X', '*', 0x03a7}, /* GREEK CAPITAL LETTER CHI */
>
> {'.', ':', 0x2234}, /* THEREFORE */
> {':', '.', 0x2235}, /* BECAUSE */
>
> and many more.
I know that the change was made because of such digraphs where order
is important, but as I read the code, the change didn't seem
necessary.
Now with your examples I can reproduce that it doesn't work, although
I still don'tunderstand why.
Basically I'd expect that the logic is the following:
If the typed digraph combination is in the table in the typed order,
it of course should use that table entry. If there is no table entry,
it should reverse the order and look it up in the table again. That
should make both cases (order is important like in the examples above,
and both orders should be working if there's only one table entry)
working as expected.
> Someone is going to need to decide which is more important: to allow
> the full range of two-character RFC1345 mnemonics in Screen, or to
> allow the lax character ordering currently in Screen. They are not
> compatible.
I still believe we can get both cases where they don't interfere. Of
course if there are different table entries for each order, they
should be honoured. But if there's only one entry for a given
combination, both orders should work.
> My preference, obviously, is to have the two-character RFC1345
> mnemonics as the authoritative list of digraphs, overruling the
> existing lax ordering. This keeps it consistent with other systems
> that use the full set of two-character RFC1345 mnemonics, such as
> input modes in Emacs, IBus, and others.
Yeah, exactly. Guess from where I'm used to that :a works the same way
as a: does. :-)
Regards, Axel
--
,''`. | Axel Beckert <a...@debian.org>, http://people.debian.org/~abe/
: :' : | Debian Developer, ftp.ch.debian.org Admin
`. `' | 1024D: F067 EA27 26B9 C3FC 1486 202E C09E 1D89 9593 0EDE
`- | 4096R: 2517 B724 C5F6 CA99 5329 6E61 2FF9 CD59 6126 16B5
--
To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
