| === rob_w_away is now known as rob_w | ||
| Satoris | Anyone have an idea on what gesture detection backend Chromium uses: http://www.chromium.org/user-experience/multitouch | 08:37 |
|---|---|---|
| === MacSlow is now known as MacSlow|lunch | ||
| === MacSlow|lunch is now known as MacSlow | ||
| === rob_w is now known as rob_w_away | ||
| === rob_w_away is now known as rob_w | ||
| cnd | bregma, I'm working on the libgrip fix now | 15:27 |
| cnd | it needs to be an sru now, so I'll go through that process | 15:27 |
| bregma | cnd, I pushed up my patch to lp:libgrip/oneiric, it should just need a once-over and an SRU | 15:35 |
| cnd | bregma, the patch was an older version, so I'm updating it to the latest version that was merged | 15:38 |
| cnd | but other than that, it looks ready | 15:38 |
| bregma | it was an older version on purpose | 15:39 |
| cnd | oh? | 15:39 |
| bregma | but it makes no difference | 15:39 |
| cnd | well, I haven't pushed | 15:40 |
| cnd | I just assumed it was stale | 15:40 |
| bregma | the latest change in themerge proposal was just a global search-and-replace rename | 15:40 |
| cnd | should I leave it as is? | 15:40 |
| cnd | ahh | 15:40 |
| bregma | I've had SRUs rejected because of trivial stuff like that | 15:40 |
| bregma | but functionally, it makes no difference | 15:40 |
| bregma | it may just work better as a cherry-pick from the upstream | 15:41 |
| bregma | so to speak | 15:41 |
| cnd | yeah | 15:41 |
| cnd | I would rather have a clean cherry-pick of the merge | 15:41 |
| bregma | OK | 15:41 |
| cnd | but we can always remove the rename if they balk | 15:41 |
| bregma | yep | 15:42 |
| cnd | bregma, can you give me a one sentence description of the fix? | 15:45 |
| bregma | let me call it up from my long-term memory.... | 15:46 |
| bregma | ah, yes: "Does not incorrectly register all widgets for gesture recognition when a top-level widget gets mapped to a window." | 15:47 |
| cnd | is that double negative a mistake? | 15:48 |
| bregma | no, but if it sounds confising you should have seen the code | 15:49 |
| cnd | huh? It's supposed to "incorrectly" register all widgets for gesture recognition? | 15:49 |
| bregma | it was previously trying to register all widgets, not it only registers widgets with the just-mapped toplevel widget | 15:49 |
| bregma | *now* it only... | 15:50 |
| cnd | ahh | 15:50 |
| cnd | and the crash was occurring because? | 15:51 |
| cnd | were some non-top-level widgets disappearing, and then the gesture would fire on them and segfault? | 15:52 |
| bregma | what happened was that widgets get created before their containing window is mapped to a native window | 15:53 |
| bregma | libgrip would queue registration up for when the mapping was complete | 15:53 |
| bregma | because it needed the native window | 15:53 |
| bregma | what happened was when _any_ window mapping was complete, libgrip tried to map _all_ pending widgets to their native windows | 15:54 |
| cnd | ahhh | 15:54 |
| cnd | I got it | 15:54 |
| bregma | but not all widgets had had their toplevel mapped to a native window yet | 15:54 |
| bregma | *boom* | 15:54 |
| cnd | ok, sru done | 16:15 |
| cnd | bregma, let me know when you have a few minutes to mumble | 16:15 |
| bregma | how about now? | 16:17 |
| cnd | ok | 16:17 |
| === rob_w is now known as rob_w_away | ||
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!