Bug 42990 - REGRESSION (r63854): Safari RSS layout is broken due to changes in vendor prefix handling
Summary: REGRESSION (r63854): Safari RSS layout is broken due to changes in vendor pre...
Alias: None
Product: WebKit
Classification: Unclassified
Component: CSS (show other bugs)
Version: 528+ (Nightly build)
Hardware: Mac (Intel) OS X 10.6
: P1 Major
Assignee: Mark Rowe (bdash)
Keywords: InRadar, Regression
: 42896 42898 (view as bug list)
Depends on:
Reported: 2010-07-26 12:09 PDT by Steven Kolins
Modified: 2010-07-26 15:54 PDT (History)
15 users (show)

See Also:

Patch (4.84 KB, patch)
2010-07-26 13:46 PDT, Mark Rowe (bdash)
mitz: review+
Details | Formatted Diff | Diff

Note You need to log in before you can comment on or make changes to this bug.
Description Steven Kolins 2010-07-26 12:09:44 PDT
sometimes rss feeds don't show anything but the first entry and sometimes the trail to the right but cut off, and sometimes there is a navigation step sideways provided. The traditional vertical list is by far the best but doesn't work at all.

See <feed:https://service.govdelivery.com/service/rss/updates.rss?code=USLOC> (entries go sideways to the right but get chopped off)

<feed://www.physorg.com/rss-feed/space-news/> where no entries except the first are visible at all

<feed://www.sciencecentric.com/feeds/> where sideways steps are provided but of far less utility than the original vertical default.

The end result is that reading rss feeds is almost useless.
Comment 1 Mark Rowe (bdash) 2010-07-26 12:43:52 PDT
This broke with <http://trac.webkit.org/changeset/63854>.
Comment 2 Mark Rowe (bdash) 2010-07-26 12:45:02 PDT
Sadly r63854 doesn’t mention which bug it is related to.
Comment 3 Mark Rowe (bdash) 2010-07-26 12:45:59 PDT
Found it.  It seems to be related to bug 42093.
Comment 4 Mark Rowe (bdash) 2010-07-26 12:50:07 PDT
Comment 5 Eric Seidel (no email) 2010-07-26 12:53:14 PDT
cd "/System/Library/Frameworks/PubSub.framework/Versions/A/Resources/PubSubAgent.app/Contents/Resources/"

grep -e "-apple-" -r .

Main.css:    -apple-text-size-adjust: none;
Main.css:    -apple-line-clamp: 100%;
Main.css:        -apple-text-size-adjust: none;

grep -e "-khtml-" -r .

Main.css:    -khtml-user-select: none;
Main.css:    display: -khtml-box;
Main.css:    -khtml-box-orient: vertical;
Main.css:    display: -khtml-box;
Main.css:    -khtml-box-orient: horizontal;
Main.css:    -khtml-box-align: baseline;
Main.css:    -khtml-user-select: none;
Main.css:    -khtml-margin-top-collapse:discard;
Main.css:    -khtml-margin-top-collapse:discard;
Main.css:        -khtml-user-select: none;
Main.css:.apple-rss-multi-line-mode  .apple-rss-subject { -khtml-box-flex: 60;                                           }
Main.css:.apple-rss-multi-line-mode  .apple-rss-author  { -khtml-box-flex: 40;                                           }
Main.css:.apple-rss-single-line-mode .apple-rss-subject { -khtml-box-flex: 1; -khtml-box-flex-group: 1;                  }
Main.css:.apple-rss-single-line-mode .apple-rss-source  { -khtml-box-flex: 1; -khtml-box-flex-group: 2; font-size: 12px; }
Main.css:.apple-rss-single-line-mode .apple-rss-summary { -khtml-box-flex: 1; -khtml-box-flex-group: 4;                  }    
Main.css:.apple-rss-single-line-mode .apple-rss-date    { -khtml-box-flex: 1; -khtml-box-flex-group: 100; font-size: 12px; text-align: right; margin-right: 0; }

Are the culprits
Comment 6 Mark Rowe (bdash) 2010-07-26 13:03:29 PDT
I find it incredibly unlikely that Safari RSS is the only thing that uses -khtml or -apple vendor prefixes.  Dashboard widgets, Dashcode templates, and third-party applications (Yahoo! Messenger) all appear to make use of these prefixes in some form.
Comment 7 Mark Rowe (bdash) 2010-07-26 13:09:46 PDT
I would like to propose rolling out r63854 and having that issue revisited in a manner that takes backwards-compatibility in to consideration.
Comment 8 Mark Rowe (bdash) 2010-07-26 13:46:31 PDT
Created attachment 62599 [details]

This reverts only the WebCore portion of r63854.  I see no reason to revert the layout test portion of that revision.
Comment 9 WebKit Review Bot 2010-07-26 13:47:57 PDT
Attachment 62599 [details] did not pass style-queue:

Failed to run "['WebKitTools/Scripts/check-webkit-style']" exit_code: 1
WebCore/css/CSSParser.cpp:5622:  Tests for true/false, null/non-null, and zero/non-zero should all be done without equality comparisons.  [readability/comparison_to_zero] [5]
Total errors found: 1 in 2 files

If any of these errors are false positives, please file a bug against check-webkit-style.
Comment 10 Mark Rowe (bdash) 2010-07-26 13:53:15 PDT
Landed in r64071.
Comment 11 WebKit Review Bot 2010-07-26 14:11:25 PDT
http://trac.webkit.org/changeset/64071 might have broken GTK Linux 32-bit Release
Comment 12 Mark Rowe (bdash) 2010-07-26 15:54:03 PDT
*** Bug 42896 has been marked as a duplicate of this bug. ***
Comment 13 Mark Rowe (bdash) 2010-07-26 15:54:38 PDT
*** Bug 42898 has been marked as a duplicate of this bug. ***