[17:02] tjaalton, bryce: how come IgnoreABI is no longer in hw/xfree86/parser/Flags.c ? [17:07] tseliot: isn't it a driver option? [17:08] nope. server option [17:08] works too... [17:08] ok [17:09] I have noticed that if I start X with startx -- -ignoreABI the driver starts [17:09] while if I set Option "IgnoreABI" "True" in the xorg.conf [17:09] it doesn't work [17:09] Mine works [17:10] In section server flags [17:10] yes, I put it there [17:11] hmm. interesting. [18:00] Alexia_Death: how is it going with your work on (sane) tablet hotplugging with dbus? [18:01] tseliot: it somewhat works [18:01] tseliot: but you cant test it with current x... [18:02] not withiout some serious patching [18:02] Alexia_Death: why? [18:02] like rebuilding it with dbus support? [18:02] 1.6 input is baadly broken [18:02] no [18:02] oh, right, you said "serious" patching [18:03] does it segfault or what? [18:03] There are at least 4 separate patches to X(one of hem is a hack because fix isnt made for 1.6 branch yet) and one patch to wacom driver [18:03] but you can test it with intrepid and x 1.5 [18:04] I even have the packages in my ppa [18:04] Theres one crash in X and one crash maker in wacom driver [18:04] to get buttons to work right on tablet theres the hack [18:05] and to fix event history theres 2 distinct patches or GIMP wil act weird [18:05] does it affect GTK at large or only GIMP? [18:05] it affects anything that uses XGetHistorybufer [18:06] err... or something like that. [18:06] Gimp does. [18:06] I havent seen anything else that does. [18:06] xournal does not [18:06] ok [18:08] I traced them and posted bugs at sorg bugzilla. history buffer fixes have patches and Ive tested them. Buttons issue has patc but does not apply to 1.6 obly to master, wacom issue is reported and hopefully will be fixed... [18:10] ok, I have to postpone my configuration app anyway. Hopefully the missing bits will be ready for Jaunty+1 [18:10] tseliot: dont [18:10] tseliot: configuration app has a place in the hoplug trick [18:10] currently it runs wacomcp [18:10] l [18:11] Alexia_Death: I have to because there's not enough time until feature freeze [18:11] tseliot: make it a standalone using xsetwacom? [18:12] when is feature freeze anyway? [18:12] yes, it will use xsetwacom too [18:12] I have settings dunping/loading implemented in my daemon [18:12] February 19th [18:13] which may not be enough for me to design and implement something that doesn't suck ;) [18:14] tseliot: think positive [18:14] IT jsut has to do better than wacomcpl [18:14] hehe [18:15] tseliot: I can give you my X packages if you want [18:15] 32 bit tho [18:15] the ones in your PPA? https://launchpad.net/~alexiade/+archive [18:15] those a re for intrepid [18:16] I dont have jauntys there because I dont have signing set up here. [18:16] but I can just give you a deb set [18:16] can you give me a link to the source? [18:17] better plan [18:17] or to the patches [18:17] Im going to give you the patch set;) [18:17] good idea [18:20] tseliot: http://a.death.pri.ee/series.tar.gz [18:20] but you need to isntall patched wacom driver too [18:20] or the X will die in unplug [18:21] that pach is a hack too because I could not get the fix to work [18:21] Alexia_Death: any links to the patch for the wacom driver? [18:22] tseliot: I dont have that in patch form, but i can pack and post an already patched version [18:22] * tseliot would rather not see X die when tablets are unplugged [18:23] Alexia_Death: if you give me the source, I'll make a diff [18:23] Ok. [18:23] I can give you just one file then [18:23] good [18:23] its ftom 0.8.2 [18:24] http://a.death.pri.ee/wcmConfig.c [18:25] tseliot: it has a bit of noise, but the two lines commented out are the catch [18:26] x tries to free the priv struct even tho wacom has freed it already and aborts. Not freeing it lets it slip [18:26] * tseliot has a look at the code [18:26] theres a null check in X side but it does not work if I set priv=NULL [18:27] hmm [18:29] what is priv set to? Or where does it point to? [18:29] anyway [18:30] * tseliot > dinner [18:30] thanks for your patches [22:09] tjaalton: bug #316136 - offscreen pixmaps problem is baack [22:09] Launchpad bug 316136 in xorg-server "Xaa Offscreen Pixmaps regression" [High,Triaged] https://launchpad.net/bugs/316136 [22:42] bryce_: yes, tormod already commented that it's fixed upstream [22:43] it wasn't merged to master the last time, because there was supposed to be a more correct fix on the way, but wasn't [22:43] tjaalton: ahh [22:43] is that something we should backport, or does it come automatically with the xserver now in our git tree? [22:43] I just posted what Timo said to the bug report :) [22:44] heh, so it seems [22:44] bryce: it's in the 1.6 branch, one commit after the last jaunty merge [22:45] ok kewl [22:46] bryce, have you considered maintaining -ati on git.d.o? [22:46] tjaalton: btw, I've posted a few little xserver patches to lp [22:46] they're the ones marked In Progress [22:48] tormod: I've had some fairly mixed experiences using git to maintain X stuff so far [22:50] tormod: it's probably my own just lack of git-specific know-how, but it seems like it always takes me more time to do something when git's involved [22:50] bryce_: ok cool [22:50] bryce_: we never went through it at UDS :) [22:51] but I will say I've had good experiences maintaining xorg in git so far [22:51] bryce_: I see. there's a certain overhead, true. but sometimes it can be very efficient. [22:51] xorg-server I've had such frustrating problems with that I am scared to commit anything to it now ;-) [22:52] we could also have other branches there.. [22:52] if we wanted to backport some stuff [22:52] tormod: mostly I post my patches back upstream. We've pulled new -ati git snapshots regularly enough that it's worked adequately [22:53] do you think they will take your "exa by default" soon? [22:54] when I proposed it to alex it sounded like he was thinking that it was about time to switch [22:54] one thing I'm thinking is we may need to build in some structure for quirking certain cards back to XAA; I've seen at least one bug report on that so far [22:55] asac's bug I guess :) [23:02] yep