[PATCH libX11 0/2] Mend some XIM design issues.
These two patches attempt to fix some issues with the design of XIM - the X Input Method. Both address races and missing synchronizations in the protocol. Due to the rather complicated matter and the conditions these errors occur each explanation for the fix is longer than the fix itself. Therefore please refer to the respective log entry as well as the comments in the code for explanations. Since only few applications use XIM today the issue might not be so pressing. Noneteless the patches have been 'field-tested' on SUSE enterprise products for quite some time now - while those fixes have eliminated the error that has been seen so far we are not aware of any reports of regressions caused by those fixes. Egbert Eich (2): XIM: Fix race on focus change: set 'FABRICATED' only when keyev filters in place. XIM: Fix sync problem on focus change. modules/im/ximcp/imDefIc.c | 27 +++ modules/im/ximcp/imDefLkup.c |4 +++- 2 files changed, 30 insertions(+), 1 deletions(-) -- 1.7.7 ___ xorg-devel@lists.x.org: X.Org development Archives: http://lists.x.org/archives/xorg-devel Info: http://lists.x.org/mailman/listinfo/xorg-devel
Re: [PATCH libX11 0/2] Mend some XIM design issues.
Happy New Year! On Fri, Dec 07, 2012 at 05:57:50AM -0500, Egbert Eich wrote: > These two patches attempt to fix some issues with the design > of XIM - the X Input Method. [...] > Noneteless the patches have been 'field-tested' on SUSE enterprise > products for quite some time now - while those fixes have eliminated > the error that has been seen so far we are not aware of any reports > of regressions caused by those fixes. Any takers for these - anyone willing to ack? I know, XIM is painful and I understand if noone wants to bother: The underlying issues are really messy. On the other hand these patches have resolved the bugs described and no regression has been seen so far although these patches have been used on SUSE Enterprise products for quite a while now. XIM the underlying input method doesn't play an important role any more today, it is still used by a lot of legacy applications and software that is (or at least used to be) based on legacy X11 libraries. The bugs addressed here have been found in Java applications. Therefore I believe it would still be beneficial to include these fixes in XIM. > > Egbert Eich (2): > XIM: Fix race on focus change: set 'FABRICATED' only when keyev > filters in place. > XIM: Fix sync problem on focus change. > > modules/im/ximcp/imDefIc.c | 27 +++ > modules/im/ximcp/imDefLkup.c |4 +++- > 2 files changed, 30 insertions(+), 1 deletions(-) > > -- > 1.7.7 Cheers, Egbert. ___ xorg-devel@lists.x.org: X.Org development Archives: http://lists.x.org/archives/xorg-devel Info: http://lists.x.org/mailman/listinfo/xorg-devel
Re: [PATCH libX11 0/2] Mend some XIM design issues.
Hi Egbert, I'm using ibus instead of scim so I can't really test. (trying to switch between ibus and scim in the past has caused me massive headaches, so I'm not real gung-ho to do that again) I tried switching between two xterms and I don't see the problem (again, with ibus). However, if the fixes work, you don't see any regressions and have been 'field-tested', I say do it. Just my two cents, Matt On 01/03/2013 09:58 AM, Egbert Eich wrote: Happy New Year! On Fri, Dec 07, 2012 at 05:57:50AM -0500, Egbert Eich wrote: These two patches attempt to fix some issues with the design of XIM - the X Input Method. [...] Noneteless the patches have been 'field-tested' on SUSE enterprise products for quite some time now - while those fixes have eliminated the error that has been seen so far we are not aware of any reports of regressions caused by those fixes. Any takers for these - anyone willing to ack? I know, XIM is painful and I understand if noone wants to bother: The underlying issues are really messy. On the other hand these patches have resolved the bugs described and no regression has been seen so far although these patches have been used on SUSE Enterprise products for quite a while now. XIM the underlying input method doesn't play an important role any more today, it is still used by a lot of legacy applications and software that is (or at least used to be) based on legacy X11 libraries. The bugs addressed here have been found in Java applications. Therefore I believe it would still be beneficial to include these fixes in XIM. Egbert Eich (2): XIM: Fix race on focus change: set 'FABRICATED' only when keyev filters in place. XIM: Fix sync problem on focus change. modules/im/ximcp/imDefIc.c | 27 +++ modules/im/ximcp/imDefLkup.c |4 +++- 2 files changed, 30 insertions(+), 1 deletions(-) -- 1.7.7 Cheers, Egbert. ___ xorg-devel@lists.x.org: X.Org development Archives: http://lists.x.org/archives/xorg-devel Info: http://lists.x.org/mailman/listinfo/xorg-devel ___ xorg-devel@lists.x.org: X.Org development Archives: http://lists.x.org/archives/xorg-devel Info: http://lists.x.org/mailman/listinfo/xorg-devel
Re: [PATCH libX11 0/2] Mend some XIM design issues.
Hi! Quite some time ago I've sent a set of patches: http://lists.x.org/archives/xorg-devel/2012-December/034699.html http://lists.x.org/archives/xorg-devel/2012-December/034698.html http://lists.x.org/archives/xorg-devel/2012-December/034700.html I haven't received any ack, but Matt suggested to push them: http://lists.x.org/archives/xorg-devel/2013-January/035116.html I haven't done so, yet, but - unless I receive a nack within the next few days I'm going to push an updated version of these. Cheers, Egbert. ___ xorg-devel@lists.x.org: X.Org development Archives: http://lists.x.org/archives/xorg-devel Info: http://lists.x.org/mailman/listinfo/xorg-devel
Re: [PATCH libX11 0/2] Mend some XIM design issues.
On 07/29/13 02:21 AM, Egbert Eich wrote: Hi! Quite some time ago I've sent a set of patches: http://lists.x.org/archives/xorg-devel/2012-December/034699.html http://lists.x.org/archives/xorg-devel/2012-December/034698.html http://lists.x.org/archives/xorg-devel/2012-December/034700.html I haven't received any ack, but Matt suggested to push them: http://lists.x.org/archives/xorg-devel/2013-January/035116.html I haven't done so, yet, but - unless I receive a nack within the next few days I'm going to push an updated version of these. Sounds good to me, I really have no idea how the XIM code is supposed to work. -- -Alan Coopersmith- alan.coopersm...@oracle.com Oracle Solaris Engineering - http://blogs.oracle.com/alanc ___ xorg-devel@lists.x.org: X.Org development Archives: http://lists.x.org/archives/xorg-devel Info: http://lists.x.org/mailman/listinfo/xorg-devel