On Fri, Apr 12, 2002 at 05:18:23PM +0200, Jan Martinek wrote:
> Hello,
>
> my friend ( [EMAIL PROTECTED] ) made a patch for fixing the bad
> behaviour of czech keyboard in fvwm. He cannot write to you himself,
> because he must have gone to military service :-(((
> This is what ho wrote about the patch:
>
>
> The patch
> + solves our problem
> + shouldn't break other things more than they are now
> - is not particularly clean
> - doesn't solve the modifier problem generally
>
> * Czech (Slovak...) keyboard has traditionally two modes: Czech and English,
> with LShift-RShift or NumLock or something like that switching between
> them. This introduces another modifier, ``keyoard mode'', which actually
> gets its own bit in the `state' field assigned (on our machines it happens
> to be 0x2000, but I don't know how deterministic is it). This modifier
> doesn't really modify anything and should be completely ignored.
>
> * Generally, it doesn't make sense to keep any modifiers which are not in
> ALL_MODIFIERS (fvwm-2.2 did it approximately this way), however, fvwm-2.4
> keeps them all and ignores only explicitely specified (known) modifiers
> (why?).
Because accoridung to the spec there are no other modifiers than
Lock, Shift, Meta and 1 to 5.
> * Originally I thought mere adding `|~ALL_MODIFIERS' to `mods_unused' (i.e.
> setting all bits corresponding to should-be-ignored modifiers) would
> solve the problem, but it generated a HUGE number of following errors:
>
> [FVWM][<<ERROR>>]: FvwmErrorHandler *** internal error ***
> [FVWM][<<ERROR>>]: FvwmErrorHandler Request 33, Error 2, EventType: 20
>
> I don't understand why, but I didn't tried very hard to.
Don't fiddle with the unused modifiers. Fvwm tries to grab each
combination of unused modifiers. E.g. if Caps-Lock and Num-Lock
are ignored and you issue
key f1 a n bla
fvwm grabs (no modifiers), (caps-lock), (num-lock) and (num-lock +
caps-lock). If you add another 24 bits to the unused modifiers,
fvwm tries to grab 2^26 combinations of modifiers. Most of them
do not exist, hence the error messages.
> * At last, I modified CheckBinding() because it works, though it's probably
> quite dirty.
No, that is a good solution, but it may be incomplete. It is
never wrong to 'and' with ALL_MODIFIERS anywhere. There may be
more places that should have a similar patch applied.
> The fact I can't just modify `mods_unused' means either I'm
> completely wrong, or I don't understand all its uses, or there something
> broken in fvwm.
see above.
> I kept MatchBindingExactly() intact because it's never used (and maybe
> could
> be completely removed).
>
> * I can't check whether it breaks something with other strange national
> keyboards because I don't use them.
>
>
> --- fvwm-2.4.5.orig/libs/Bindings.c Wed Feb 20 20:19:35 2002
> +++ fvwm-2.4.5/libs/Bindings.c Wed Feb 20 20:19:38 2002
> @@ -550,7 +550,7 @@
> Binding *b;
> unsigned int used_modifiers = ~dead_modifiers;
>
> - modifier &= used_modifiers;
> + modifier &= used_modifiers & ALL_MODIFIERS;
>
> for (b = blist; b != NULL; b = b->NextBinding)
> {
Bye
Dominik ^_^ ^_^
--
Dominik Vogt, email: [EMAIL PROTECTED]
LifeBits Aktiengesellschaft, Albrechtstr. 9, D-72072 Tuebingen
fon: ++49 (0) 7071/7965-0, fax: ++49 (0) 7071/7965-20
--
Visit the official FVWM web page at <URL:http://www.fvwm.org/>.
To unsubscribe from the list, send "unsubscribe fvwm-workers" in the
body of a message to [EMAIL PROTECTED]
To report problems, send mail to [EMAIL PROTECTED]
