RESOLVED WONTFIX 75116
[chromium] Add support for layer tree debugging
https://bugs.webkit.org/show_bug.cgi?id=75116
Summary [chromium] Add support for layer tree debugging
vollick
Reported 2011-12-22 13:12:41 PST
[chromium] Add support for creating dotty-friendly layer tree dumps
Attachments
Patch (26.69 KB, patch)
2011-12-22 13:19 PST, vollick
no flags
Patch (29.32 KB, patch)
2012-01-30 08:39 PST, vollick
no flags
Patch (18.34 KB, patch)
2012-03-06 18:31 PST, Peter Kotwicz
jamesr: review-
vollick
Comment 1 2011-12-22 13:19:22 PST
WebKit Review Bot
Comment 2 2011-12-22 14:37:38 PST
Please wait for approval from fishd@chromium.org before submitting because this patch contains changes to the Chromium public API.
James Robinson
Comment 3 2011-12-22 18:37:40 PST
Comment on attachment 120371 [details] Patch View in context: https://bugs.webkit.org/attachment.cgi?id=120371&action=review What is dotty? > Source/WebCore/ChangeLog:8 > + No new tests. (OOPS!) this won't blend
vollick
Comment 4 2012-01-04 05:25:30 PST
(In reply to comment #3) > (From update of attachment 120371 [details]) > View in context: https://bugs.webkit.org/attachment.cgi?id=120371&action=review > > What is dotty? It's a graph viz tool. The hope was that we could generate a graph of the webkit layers for debugging purposes. Also, I should have said 'in the dot language' rather than 'dotty-friendly' http://en.wikipedia.org/wiki/DOT_language > > > Source/WebCore/ChangeLog:8 > > + No new tests. (OOPS!) > > this won't blend Does this mean a test is required or that there were merge conflicts? I'm pretty sure you meant the former, but I thought I'd double-check, just in case.
Peter Kotwicz
Comment 5 2012-01-27 19:35:25 PST
Ping
James Robinson
Comment 6 2012-01-27 19:37:15 PST
I'm not sure this is really worth having, to be honest.
vollick
Comment 7 2012-01-30 08:39:07 PST
Peter Kotwicz
Comment 8 2012-03-06 18:31:46 PST
WebKit Review Bot
Comment 9 2012-03-06 18:35:28 PST
Please wait for approval from fishd@chromium.org, abarth@webkit.org or jamesr@chromium.org before submitting because this patch contains changes to the Chromium platform API.
James Robinson
Comment 10 2012-03-06 18:38:46 PST
Comment on attachment 130513 [details] Patch It seems like you attached this patch to the wrong bug, no? I'm not at all excited about adding a synchronous cross-thread commit flow unless it's super super valuable. There are lots of subtle threading/ordering bugs to flush out to make this work well. What is this needed for?
Peter Kotwicz
Comment 11 2012-03-09 15:11:15 PST
Sorry for taking so long to reply. I want to use this for displaying the impl layertree in WebUI. On the chromium side, I have created a WebUI page which will display the layer trees for both the browser and the webkit layertree. This is similar to the flags --show-composited-layer-tree and --ui-show-layer-tree The advantage of doing this in WebUI are: 1) The layertree is scrollable (the browser layer tree is large) 2) We can explore with interesting ways of showing the layertree data if useful I believe that something like this could be useful for people starting out with WebKit. I chose to display the impl layer tree because we were already generating the Webkit layer tree. One option would be to generate the non impl layer tree in LayerChromium and use that. wjmaclean would this still be useful?
W. James MacLean
Comment 12 2012-03-13 11:25:33 PDT
(In reply to comment #11) > Sorry for taking so long to reply. > I want to use this for displaying the impl layertree in WebUI. On the chromium side, I have created a WebUI page which will display the layer trees for both the browser and the webkit layertree. > This is similar to the flags --show-composited-layer-tree and --ui-show-layer-tree > The advantage of doing this in WebUI are: > 1) The layertree is scrollable (the browser layer tree is large) > 2) We can explore with interesting ways of showing the layertree data if useful > > I believe that something like this could be useful for people starting out with WebKit. > > I chose to display the impl layer tree because we were already generating the Webkit layer tree. > One option would be to generate the non impl layer tree in LayerChromium and use that. wjmaclean would this still be useful? Having seen this in action, I think it will be useful for large layer trees, and esp. in Aura development it's nice being able to have the layer tree displayed in a separate window from the page being debugged. The benefit of being able to make display more informative by showing additional information seems pretty useful. Ultimately it would be nice to see both main- and impl-thread trees (although these need not be in the same window). In the short term I'm not sure which tree is more important to display (or if there's even an answer to that question).
Note You need to log in before you can comment on or make changes to this bug.