Summary: | Incoming source should be preload scanned when the parser is blocked | ||||||
---|---|---|---|---|---|---|---|
Product: | WebKit | Reporter: | Antti Koivisto <koivisto> | ||||
Component: | Page Loading | Assignee: | Nobody <webkit-unassigned> | ||||
Status: | RESOLVED FIXED | ||||||
Severity: | Normal | CC: | abarth, eric, simonjam, tonyg, webkit.review.bot | ||||
Priority: | P2 | ||||||
Version: | 528+ (Nightly build) | ||||||
Hardware: | PC | ||||||
OS: | OS X 10.5 | ||||||
Bug Depends on: | 58352 | ||||||
Bug Blocks: | |||||||
Attachments: |
|
Description
Antti Koivisto
2011-04-07 23:18:43 PDT
Created attachment 88781 [details]
patch
Comment on attachment 88781 [details] patch View in context: https://bugs.webkit.org/attachment.cgi?id=88781&action=review > LayoutTests/ChangeLog:-9650 > ->>>>>>> .r82178 This change looks suspicious. :) > LayoutTests/http/tests/loading/preload-append-scan.php:18 > +usleep(200000); Does this test really take 20 seconds to run? tonyg, would you be willing to review this patch? (In reply to comment #2) > > +usleep(200000); > > Does this test really take 20 seconds to run? 200ms > 200ms
I'm really too tired to be commenting in bugs. :)
Comment on attachment 88781 [details] patch View in context: https://bugs.webkit.org/attachment.cgi?id=88781&action=review This looks good, but we are starting to get a lot of entry points for scanning. I'm wondering if we shouldn't fix bug 57376, then always kick off the preload scanner first thing in insert() and append() (and nowhere else). > LayoutTests/http/tests/resources/slow-script.pl:1 > +#!/usr/bin/perl -w Is it possible to avoid adding a new resource and instead reuse misc/resources/script-slow1.pl? http://trac.webkit.org/changeset/83321 might have broken Qt Linux Release The following tests are not passing: http/tests/loading/preload-append-scan.php |