Martijn Wargers (dead) Assignee
	Description • 18 years ago

See testcase, the computed style of the various -moz-border-radius and -moz-outline-radius properties return a negative value.
According to the border-radius working draft, negative values are not allowed.

With new border rewrite, borders/outlines with negative radius values, now look a bit strange.

	David Baron :dbaron:
	Comment 2 • 18 years ago

Comment on attachment 267047 [details] [diff] [review]
patch

r+sr=dbaron

Could you add an invalid_values item to the test items in layout/style/test/property_database.js as well?

	Martijn Wargers (dead) Assignee
	Comment 3 • 18 years ago

Thanks, I was already wondering about how to add a testcase for this.

	David Baron :dbaron:
	Comment 4 • 18 years ago

Comment on attachment 267058 [details] [diff] [review]
patch

r+sr=dbaron

	Martijn Wargers (dead) Assignee
	Comment 5 • 18 years ago

database.js
Checking in nsCSSParser.cpp;
/cvsroot/mozilla/layout/style/nsCSSParser.cpp,v  <--  nsCSSParser.cpp
new revision: 3.358; previous revision: 3.357
done
Checking in test/property_database.js;
/cvsroot/mozilla/layout/style/test/property_database.js,v  <--  property_databas
e.js
new revision: 1.15; previous revision: 1.14
done

Checked into trunk.

	pal-moz
	Comment 6 • 18 years ago

Mozilla/5.0 (Windows; U; Windows NT 5.1; en-US; rv:1.9a5) Gecko/20070606 GranParadiso/3.0a5 ID:2007060601

no check-in to GP a5?

	Martijn Wargers (dead) Assignee
	Comment 7 • 18 years ago

Why?
This isn't critical or a widely occuring bug.

	pal-moz
	Comment 8 • 18 years ago

screen-shot:

after check-in
no.1 [Trunk/2007060504] http://img267.imageshack.us/img267/6307/a1fe9.jpg
no.2 [Gran Paradiso/2007060601] http://img509.imageshack.us/img509/3927/a2vx9.jpg

before check-in
no.3 [Trunk/2007060304] http://img267.imageshack.us/img267/9494/a3fd6.jpg

no.2 and no.3 is same result.
i.e.)GP-Build ID is newest(build after check-in), but result is same as one(build before check-in).

so I asked it.

	Martijn Wargers (dead) Assignee
	Comment 9 • 18 years ago

Yeah, that means the patch isn't in GP a5, apparently.

	pal-moz
	Comment 10 • 18 years ago

OK,
If I am not wrong, GP a4(2007042705)/GP a3(2007032219)... have all check-ins before its Build ID.
so why a5 says "2007060601", though there are not some check-ins like this bug (this check-in was 20070603**) ?

tell me which check-ins are not in a5(from when), if you know ?

	Martijn Wargers (dead) Assignee
	Comment 11 • 18 years ago

Sorry, I don't know.
Apparently, the GP a5 tree is somewhat older, although I know there were a few respins.

	:Gavin Sharp [email: gavin@gavinsharp.com]
	Comment 12 • 18 years ago

A5 RC1 was pulled and built, and then it was later decided to respin for the patch in bug 382482. The tag for the original A5 was used, and the patch for 382482 was applied to it. The build ID is the build time, not source-pull time, so these kinds of things happen.

	pal-moz
	Comment 13 • 18 years ago

(In reply to comment #11)
(In reply to comment #12)

Thank you for a detailed explanation.