WebKit Bugzilla
New
Browse
Log In
×
Sign in with GitHub
or
Remember my login
Create Account
·
Forgot Password
Forgotten password account recovery
RESOLVED FIXED
Bug 134153
[EFL] [GTK] Install Latin Modern Math on the bots and update the MathML test references
https://bugs.webkit.org/show_bug.cgi?id=134153
Summary
[EFL] [GTK] Install Latin Modern Math on the bots and update the MathML test ...
Frédéric Wang (:fredw)
Reported
2014-06-20 23:46:15 PDT
Follow-up of
bug 133604
.
https://github.com/mrobinson/webkitgtk-test-fonts/pull/5
Attachments
Patch
(97.57 KB, patch)
2014-06-22 01:31 PDT
,
Frédéric Wang (:fredw)
no flags
Details
Formatted Diff
Diff
Patch
(253.09 KB, patch)
2014-06-24 08:33 PDT
,
Frédéric Wang (:fredw)
no flags
Details
Formatted Diff
Diff
Patch
(256.70 KB, patch)
2014-06-24 10:21 PDT
,
Frédéric Wang (:fredw)
no flags
Details
Formatted Diff
Diff
Patch
(1.53 KB, patch)
2014-06-24 23:28 PDT
,
Frédéric Wang (:fredw)
gyuyoung.kim
: review+
Details
Formatted Diff
Diff
Show Obsolete
(3)
View All
Add attachment
proposed patch, testcase, etc.
Frédéric Wang (:fredw)
Comment 1
2014-06-22 01:31:16 PDT
Created
attachment 233560
[details]
Patch Just a quick run with the Latin Modern Math font: - The mathml/opentype/ tests pass again ;-) - mathml/presentation/menclose-notation-default-longdiv-expected.txt just has a 1px size diff in the text reference - the pixel tests mathml/presentation/roots-expected.txt should be updated to use the new style. - same for mathml/presentation/mo-stretch.html. It seems that Latin Modern Math is not able to stretch the simple integral at arbitrary size, so this is different from what is tested here. I don't think that's too important for MathML though, since by default integrals are just "largeop" not "stretchy". - mathml/presentation/
bug95015
.html fails. I think the test is a bit fragile and does not pass on all platforms. - same for mathml/presentation/mo-stretch.html. Here the difference is even not visible at the naked eye.
Frédéric Wang (:fredw)
Comment 2
2014-06-24 08:33:03 PDT
Created
attachment 233707
[details]
Patch
Frédéric Wang (:fredw)
Comment 3
2014-06-24 08:41:52 PDT
@Martin: can you review the patch and take care of the GTK bots? @Gyuyoung Kim: I 've finally been able to build the EFL port again, but not to execute the tests. However, I've verified that the MathML pages look ok with Latin Modern Math. I'll let you rebaseline the tests and I expect it will be more or less the same as what is done for GTK. The version with Latin Modern Math is
https://github.com/mrobinson/webkitgtk-test-fonts/archive/master.zip
(perhaps Martin can suggest a better link). (In reply to
comment #1
)
> Created an attachment (id=233560) [details] > - same for mathml/presentation/mo-stretch.html. Here the difference is even not visible at the naked eye.
I meant mo-stacked-glyphs.html. This is already covered by mo-stretch, though.
Martin Robinson
Comment 4
2014-06-24 09:06:13 PDT
Comment on
attachment 233707
[details]
Patch You can actually take care of the bots in this patch. I pushed the 0.0.5 tag to the webkitgtk-test-fonts repository on GitHub. You simply need to update the tag reference in Tools/gtk/jhbuild.modules.
Frédéric Wang (:fredw)
Comment 5
2014-06-24 10:21:44 PDT
Created
attachment 233716
[details]
Patch
Martin Robinson
Comment 6
2014-06-24 10:28:04 PDT
Comment on
attachment 233716
[details]
Patch Is the appropriate font now selected in mathml.css?
Frédéric Wang (:fredw)
Comment 7
2014-06-24 10:41:03 PDT
(In reply to
comment #6
)
> (From update of
attachment 233716
[details]
) > Is the appropriate font now selected in mathml.css?
This was done in
bug 133604
. It's just that the jhbuild envs didn't have Latin Modern Math installed.
Frédéric Wang (:fredw)
Comment 8
2014-06-24 23:28:28 PDT
Created
attachment 233792
[details]
Patch
Gyuyoung Kim
Comment 9
2014-06-24 23:30:32 PDT
Comment on
attachment 233792
[details]
Patch LGTM.
Frédéric Wang (:fredw)
Comment 10
2014-06-24 23:31:52 PDT
Committed
r170419
: <
http://trac.webkit.org/changeset/170419
>
Carlos Alberto Lopez Perez
Comment 11
2014-06-27 10:17:30 PDT
This caused failures on the following tests: fast/dynamic/text-combine.html fast/regions/repaint/japanese-rl-selection-repaint-in-regions.html fast/repaint/japanese-rl-selection-clear.html fast/repaint/japanese-rl-selection-repaint.html fast/text/decorations-with-text-combine.html fast/text/emphasis-vertical.html fast/text/international/002.html fast/text/international/text-combine-image-test.html fast/text/international/text-spliced-font.html fast/text/international/wrap-CJK-001.html fast/writing-mode/border-vertical-lr.html fast/writing-mode/japanese-lr-selection.html fast/writing-mode/japanese-lr-text.html fast/writing-mode/japanese-rl-selection.html fast/writing-mode/japanese-rl-text.html I have rebaselined the tests on
https://bugs.webkit.org/show_bug.cgi?id=134400
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