WebKit Bugzilla
New
Browse
Log In
×
Sign in with GitHub
or
Remember my login
Create Account
·
Forgot Password
Forgotten password account recovery
RESOLVED FIXED
56111
Web Inspector: Optimize backend-frontend data transfer volume for CSS styles
https://bugs.webkit.org/show_bug.cgi?id=56111
Summary
Web Inspector: Optimize backend-frontend data transfer volume for CSS styles
Alexander Pavlov (apavlov)
Reported
2011-03-10 09:37:16 PST
Currently every CSS property is described by a number of key:value pairs in the Web Inspector protocol. This results in computed style transfers of about 30kB for ~220 properties and should certainly be reworked (this is often slow even without remote debugging).
Attachments
[PATCH] Suggested solution
(8.33 KB, patch)
2011-03-10 10:58 PST
,
Alexander Pavlov (apavlov)
pfeldman
: review+
Details
Formatted Diff
Diff
View All
Add attachment
proposed patch, testcase, etc.
Alexander Pavlov (apavlov)
Comment 1
2011-03-10 10:58:50 PST
Created
attachment 85350
[details]
[PATCH] Suggested solution
Patrick Mueller
Comment 2
2011-03-10 13:43:04 PST
Makes sense. I'm curious how much this saves. I've been noticing for weinre that computed styles are coming over far too often, as near as I can tell. For instance, I see them come over even if the computed style disclosure is set to "hide". Might well be something I'm doing wrong, or something that's changed since I last caught up to WI, of maybe they need to come over anyway for some other calculation.
Pavel Feldman
Comment 3
2011-03-10 13:52:07 PST
(In reply to
comment #2
)
> Makes sense. I'm curious how much this saves. > > I've been noticing for weinre that computed styles are coming over far too often, as near as I can tell. For instance, I see them come over even if the computed style disclosure is set to "hide". Might well be something I'm doing wrong, or something that's changed since I last caught up to WI, of maybe they need to come over anyway for some other calculation.
Good point - I think we are sending computed style information regardless of the sidebar section state.
Ilya Tikhonovsky
Comment 4
2011-03-11 11:54:22 PST
was landed as
r80845
Ilya Tikhonovsky
Comment 5
2011-03-11 11:54:58 PST
was landed as
r80845
(In reply to
comment #4
)
> was landed as
r80845
wrong bug
Pavel Feldman
Comment 6
2011-03-14 02:39:36 PDT
Comment on
attachment 85350
[details]
[PATCH] Suggested solution View in context:
https://bugs.webkit.org/attachment.cgi?id=85350&action=review
> LayoutTests/inspector/styles/styles-new-API.html:244 > + InspectorTest.addResult("['" + property.name + "':'" + property.value + "'" + (property.priority === "important" ? " is-important" : "") + ((typeof property.parsedOk === "undefined") ? "" : " non-parsed") +"] @[" + property.startOffset + "-" + property.endOffset + "] " + (property.status || "style"));
("parsedOk" in property)
> Source/WebCore/inspector/InspectorStyleSheet.cpp:453 > + String status = it->disabled ? "disabled" : "active";
I'd suggest that: - active properties have no additional flags, - disabled properties have "disabled" set to true, - style (longhand) properties are placed under source code (shorthand) properties in case of regular styles + computed styles are returned as arrays of longhands.
Alexander Pavlov (apavlov)
Comment 7
2011-03-14 05:54:29 PDT
Committed
r81019
: <
http://trac.webkit.org/changeset/81019
>
Note
You need to
log in
before you can comment on or make changes to this bug.
Top of Page
Format For Printing
XML
Clone This Bug