--===============4330556504894942534== Content-Type: multipart/alternative; boundary="===============6292060165227955166==" --===============6292060165227955166== Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: quoted-printable > On Dec. 30, 2012, 7:42 p.m., Ivan =C4=8Cuki=C4=87 wrote: > > May I suggest a bit different approach - instead of handling the null a= ctivity like it is something special, why not just do this - when an activi= ty that has windows only on it is deleted, those are moved to the current a= ctivity. This would be more generic. > = > Thomas L=C3=BCbking wrote: > /** > * updates clients when an activity is destroyed. > * this ensures that a client does not get 'lost' if the only activit= y it's on is removed. > */ > void Workspace::slotActivityRemoved(const QString &activity) > ... > = > That's not the problem, also see bug #309203 > I ran into it later on as well - but don't ask me for what reason (an= d there was actually no "legal" way to get back to the client, but "xprop -= remove _KDE_NET_WM_ACTIVITIES -id ") > = > Apparently whether getting lost in dbus or whatever, it can under som= e circumstances happen that windows are assigned to the null uuid and, mayb= e it is due to its special nature, never leave it, resp. not leave it autom= atically. > = > Whenever the activitymanagerd crashes, or stops responding or dbus is= flooded or ppl. say "i don't need no st.. what was aarons blogpost ;-)" or= for whatever reason the library says "nulluuid" the result should not be i= naccessible windows, never. > Thus the sane solution is to treat this as invalid, thus absent of sp= ecific, thus "ALL". > = > We could also remove the special "ALL" property and use the nulluuid = as it. > = > Ivan =C4=8Cuki=C4=87 wrote: > > We could also remove the special "ALL" property and use the nulluui= d as it. > = > Like! Ok, patch in KWin is here, but there's at least a reference in "service/ui/= plugins/declarativeui/DeclarativeUi.cpp" you need to get rid of as well, si= nce we're not gonna support this any more. - Thomas ----------------------------------------------------------- This is an automatically generated e-mail. To reply, visit: http://git.reviewboard.kde.org/r/107855/#review24287 ----------------------------------------------------------- On Dec. 30, 2012, 3:47 p.m., Thomas L=C3=BCbking wrote: > = > ----------------------------------------------------------- > This is an automatically generated e-mail. To reply, visit: > http://git.reviewboard.kde.org/r/107855/ > ----------------------------------------------------------- > = > (Updated Dec. 30, 2012, 3:47 p.m.) > = > = > Review request for kwin, Ivan =C4=8Cuki=C4=87 and Martin Gr=C3=A4=C3=9Fli= n. > = > = > Description > ------- > = > This is replied as activity by the lib cache in case the server is (yet) = not up, but no window should ever be assigned to it - so we should rather r= emove it (and in doubt treat this as "all activities" condition) > = > = > Diffs > ----- > = > kwin/client.cpp b555c11 = > = > Diff: http://git.reviewboard.kde.org/r/107855/diff/ > = > = > Testing > ------- > = > semi - i once ran into one or more window/s being on that id, not with th= e patch since after (surprise) > = > = > Thanks, > = > Thomas L=C3=BCbking > = > --===============6292060165227955166== Content-Type: text/html; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: quoted-printable
This is an automatically generated e-mail. To reply, visit: http://git.revie= wboard.kde.org/r/107855/

On December 30th, 2012, 7:42 p.m., Ivan =C4= =8Cuki=C4=87 wrote:

May I sug=
gest a bit different approach - instead of handling the null activity like =
it is something special, why not just do this - when an activity that has w=
indows only on it is deleted, those are moved to the current activity. This=
 would be more generic.

On December 30th, 2012, 8:21 p.m., Thomas L=C3=BCbking wrote:

/**
 * updates clients when an activity is destroyed.
 * this ensures that a client does not get 'lost' if the only activ=
ity it's on is removed.
 */
void Workspace::slotActivityRemoved(const QString &activity)
...

That's not the problem, also see bug #309203
I ran into it later on as well - but don't ask me for what reason (and =
there was actually no "legal" way to get back to the client, but =
"xprop -remove _KDE_NET_WM_ACTIVITIES -id <wid>")

Apparently whether getting lost in dbus or whatever, it can under some circ=
umstances happen that windows are assigned to the null uuid and, maybe it i=
s due to its special nature, never leave it, resp. not leave it automatical=
ly.

Whenever the activitymanagerd crashes, or stops responding or dbus is flood=
ed or ppl. say "i don't need no st.. what was aarons blogpost ;-)&=
quot; or for whatever reason the library says "nulluuid" the resu=
lt should not be inaccessible windows, never.
Thus the sane solution is to treat this as invalid, thus absent of specific=
, thus "ALL".

We could also remove the special "ALL" property and use the nullu=
uid as it.

On December 30th, 2012, 8:32 p.m., Ivan =C4=8Cuki=C4=87 wrote: 

> We c=
ould also remove the special "ALL" property and use the nulluuid =
as it.

Like!
Ok, patch i=
n KWin is here, but there's at least a reference in "service/ui/pl=
ugins/declarativeui/DeclarativeUi.cpp" you need to get rid of as well,=
 since we're not gonna support this any more.

- Thomas


On December 30th, 2012, 3:47 p.m., Thomas L=C3=BCbking wrote:

Review request for kwin, Ivan =C4=8Cuki=C4=87 and Martin Gr=C3=A4=C3= =9Flin.
By Thomas L=C3=BCbking.

Updated Dec. 30, 2012, 3:47 p.m.

Descripti= on 

This is replied as activity by the lib cache in case the ser=
ver is (yet) not up, but no window should ever be assigned to it - so we sh=
ould rather remove it (and in doubt treat this as "all activities"=
; condition)

Testing <= /h1> 
semi - i once ran into one or more window/s being on that id=
, not with the patch since after (surprise)

Diffs=

  • kwin/client.cpp (b555c11)

View Diff
--===============6292060165227955166==-- --===============4330556504894942534== Content-Type: text/plain; charset="us-ascii" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit Content-Disposition: inline _______________________________________________ kwin mailing list kwin@kde.org https://mail.kde.org/mailman/listinfo/kwin --===============4330556504894942534==--