event.c freezes on XFilterEvent

[Vik]

Hi,

I'm rather new to wine developement, and Xlib coding, and this is my first 
post to the list so bare with me please.... ;-)

dlls/x11drv/event.c seems to freeze at XFilterEvent and brings the app to 
seizure ever since all the TSX functions have been gotten rid off...

This is what happens:
I run notepad, and start typing. Everything works fine until I hit shift+space 
which is the toggle for XIM (kinput in my case). 
Notepad freezes here.

The trace on keyboard tells me that there is no character for the key 
combination and freezes right after;

trace:keyboard:KEYBOARD_MapDeadKeysym no character for dead keysym 0x0000ffe1
> From my investigation, it seems that XFilterEvent (event.c, function

process_events)
<code>
ignore = XFilterEvent( &event, None );
</code>
is causing this. If I get rid of the whole check, and pass all the events to 
EVENT_ProcessEvent() , notepad would not freeze.

XFilterEvent is supposed to return TRUE when the event is filtered by the XIM, 
but chances are that since the XIM functionalities are currently not really 
there, XFilterEvent cannot talk to the XIM context.

As far as I remember when the TSX functions were there, the XIM window did pop 
up(despite XIM not working properly). I'm trying to trace the changes to see 
what might be causing this, but without much luck so far...

I'm more than willing to try and fix this, as XIM support is critical for 
people from my part of the world (Japan, and most likely for Korea & China). 
We've been coming up with rather crude patches to have XIM work till now(upto 
20031118), but can't get it to work because of this strange freezure we are 
experiencing....

Any pointers or thoughts will be greatly appreciated!

Cheers
Vik

[Tom]

Vik wrote:
> Hi,
>
> I'm more than willing to try and fix this, as XIM support is critical for 
> people from my part of the world (Japan, and most likely for Korea & China). 
> We've been coming up with rather crude patches to have XIM work till now(upto 
> 20031118), but can't get it to work because of this strange freezure we are 
> experiencing....
>
> Any pointers or thoughts will be greatly appreciated!

If I were you I would contect Mike McCormack & Jon Griffiths......

I'm sure -debug info would be apprishiated.. (sorry for the spelling i'm 
of english decent)
only the russians and french can spell correct in english :-) (we speak 
it, they fix spelling mistakes = 50/50)

nasda,

Tom
> Cheers
> Vik

[Vik]

Hi,
Happy New Year!
> If I were you I would contect Mike McCormack & Jon Griffiths......

Thanks for the lead, I believe most of the current XIM patch code we use came 
from Mike's patch in 2002, and a mix of many other people's patches.
I'll see if I can get in touch with him...
> I'm sure -debug info would be apprishiated..

I'm sure it is... although, I'm not too sure on how to go about doing this. 
I'm thinking the wine --debugmsg output is not going to help much, so when 
debugging something like this, would I have to go into gdb or something of 
that sorts? (sorry I'm still rather new to all this :) )

I suppose since the whole CJK support issue is on the TODO, it has to be dealt 
with some how sooner or later. I'm more than willing to invest some time on 
this, although I'm not sure how much I can do.

I'll see if I can "fix" it first and submit a patch here for review... :)
XIM and all can come later... It's more of an annoyance issue right now 
anyway, as I often hit shift+space by accident (while typing in English) and 
brings the application to a halt. One thing I can think of right now is to 
get rid of the XIM related calls and think about implementing it better. 
Although that might not be a good way to go about it...
> only the russians and french can spell correct in english :-) (we speak
> it, they fix spelling mistakes = 50/50)

Hehe. The important thing I guess is to get the point across :-)

Cheers
Vik

[Joshua Walker]

<snip>
> I often hit shift+space by accident
> (while typing in English) and 
> brings the application to a halt.

<snip>

In my .xinitrc...

------------------
export LANG="ja_JP.UTF-8"
export XMODIFIERS="@im=kinput2"
xmodmap -e 'keycode 115 = Kanji'

kinput2 -xim -kinput -canna &
----------------------------

The xmodmap command will map the kanji key to the
Windows key, then you have to dump the shift-space
combo

edit this file...
/usr/X11R6/lib/X11/app-defaults/Kinput2

and comment out the following lines with bangs...
-------------------------
*ConversionStartKeys: \
   Shift<Key>space \n\
   Ctrl<Key>Kanji
--------------------------

That would solve your shift-space problem.

[Aric Stewart]

Hello,

I have done extensive XIM work in the crossover tree and have it 
working very well with a wide range of input methods including half and 
full ime input methods used in applications such as office-jp.

I have been quite busy recently but will try hard to get a patch put 
together to get that all into wine. If you want to work on it before i 
get a chance you can hit the crossover source tree which i believe in 
pointed to somewhere on our web site.

-aric

Vik wrote:
> Hi,
> Happy New Year!
> > If I were you I would contect Mike McCormack & Jon Griffiths......
>
> Thanks for the lead, I believe most of the current XIM patch code we use came 
> from Mike's patch in 2002, and a mix of many other people's patches.
> I'll see if I can get in touch with him...
> > I'm sure -debug info would be apprishiated..
>
> I'm sure it is... although, I'm not too sure on how to go about doing this. 
> I'm thinking the wine --debugmsg output is not going to help much, so when 
> debugging something like this, would I have to go into gdb or something of 
> that sorts? (sorry I'm still rather new to all this :) )
>
> I suppose since the whole CJK support issue is on the TODO, it has to be dealt 
> with some how sooner or later. I'm more than willing to invest some time on 
> this, although I'm not sure how much I can do.
>
> I'll see if I can "fix" it first and submit a patch here for review... :)
> XIM and all can come later... It's more of an annoyance issue right now 
> anyway, as I often hit shift+space by accident (while typing in English) and 
> brings the application to a halt. One thing I can think of right now is to 
> get rid of the XIM related calls and think about implementing it better. 
> Although that might not be a good way to go about it...
> > only the russians and french can spell correct in english :-) (we speak
> > it, they fix spelling mistakes = 50/50)
>
> Hehe. The important thing I guess is to get the point across :-)
>
> Cheers
> Vik

[Vik]

Hi,
> I have been quite busy recently but will try hard to get a patch put
> together to get that all into wine.

Wow! That will be fantastic!
> If you want to work on it before i
> get a chance you can hit the crossover source tree which i believe in
> pointed to somewhere on our web site.

I'll make sure I get my paws on the source tree, and see what I can do.

Cheers
Vik

On Tuesday 06 January 2004 23:56, Aric Stewart wrote:
> Hello,
>
> I have done extensive XIM work in the crossover tree and have it
> working very well with a wide range of input methods including half and
> full ime input methods used in applications such as office-jp.
>
> I have been quite busy recently but will try hard to get a patch put
> together to get that all into wine. If you want to work on it before i
> get a chance you can hit the crossover source tree which i believe in
> pointed to somewhere on our web site.
>
> -aric
>
> Vik wrote:
> > Hi,
> > Happy New Year!
> > > If I were you I would contect Mike McCormack & Jon Griffiths......
> >
> > Thanks for the lead, I believe most of the current XIM patch code we use
> > came from Mike's patch in 2002, and a mix of many other people's patches.
> > I'll see if I can get in touch with him...
> > > I'm sure -debug info would be apprishiated..
> >
> > I'm sure it is... although, I'm not too sure on how to go about doing
> > this. I'm thinking the wine --debugmsg output is not going to help much,
> > so when debugging something like this, would I have to go into gdb or
> > something of that sorts? (sorry I'm still rather new to all this :) )
> >
> > I suppose since the whole CJK support issue is on the TODO, it has to be
> > dealt with some how sooner or later. I'm more than willing to invest some
> > time on this, although I'm not sure how much I can do.
> >
> > I'll see if I can "fix" it first and submit a patch here for review... :)
> > XIM and all can come later... It's more of an annoyance issue right now
> > anyway, as I often hit shift+space by accident (while typing in English)
> > and brings the application to a halt. One thing I can think of right now
> > is to get rid of the XIM related calls and think about implementing it
> > better. Although that might not be a good way to go about it...
> > > only the russians and french can spell correct in english :-) (we speak
> > > it, they fix spelling mistakes = 50/50)
> >
> > Hehe. The important thing I guess is to get the point across :-)
> >
> > Cheers
> > Vik