Make some unicode symbols be rendered by Segoe UI Symbol on Windows

Make some unicode symbols be rendered by Segoe UI Symbol on Windows

This CL makes unicode blocks categorized as Runic, Dingbats, general punctuation, miscellaneous symbols, and supplemental mathematical operators be rendered by Segoe UI Symbol on Windows.

BUG=108591, 225862, 264195, 426365
Committed: https://src.chromium.org/viewvc/blink?view=rev&revision=193642

[hichris123, 2015-03-12 23:45:54]

hichris123@gmail.com changed reviewers:
+ esprehn@chromium.org

[hichris123, 2015-03-12 23:48:19]

esprehn@: PTAL. This CL does have some Layout Tests failing (due to the change
in rendering font), but I made UBLOCK_SUPPLEMENTAL_MATHEMATICAL_OPERATORS and
UBLOCK_DINGBATS render with Segoe UI Symbol so more of the symbols render.

Also, should I create a new Layout Test for this? Thanks!

[pdr., 2015-03-13 02:25:26]

pdr@chromium.org changed reviewers:
+ eae@chromium.org

[pdr., 2015-03-13 02:25:39]

pdr@chromium.org changed reviewers:
+ pdr@chromium.org

[pdr., 2015-03-13 02:25:39]

+cc eae

[esprehn, 2015-03-13 02:27:54]

eae@ is a better reviewer for this, I don't know much about font code on
windows.

[esprehn, 2015-03-13 02:28:08]

esprehn@chromium.org changed reviewers:
- esprehn@chromium.org

[eae, 2015-03-13 15:19:04]

Thanks for working on this,

UBLOCK_GENERAL_PUNCTUATION seems overly broad for this, what makes you think
Segoe UI Symbol would be better for punctuation? 

As for bug 225862 ZWJ should be handled by the complex path normalization stage
and not be dependent on font fallback.

Finally, all functional changes to blink require test coverage. Please add a set
of layout tests covering each of the new code blocks.

[hichris123, 2015-03-13 19:48:48]

On 2015/03/13 15:19:04, eae wrote:
> Thanks for working on this,
> 
> UBLOCK_GENERAL_PUNCTUATION seems overly broad for this, what makes you think
> Segoe UI Symbol would be better for punctuation? 
> 
> As for bug 225862 ZWJ should be handled by the complex path normalization
stage
> and not be dependent on font fallback.

I was thinking it would be better mostly because of bug 225862, and also because
of the boxes on http://www.charbase.com/block/general-punctuation. Segoe UI
Symbol renders quite a few more of those unicode characters than Lucida Sans
does.

Since you say 225862 should not be handled by font fallback, do you think I
should remove UBLOCK_GENERAL_PUNCTUATION from this CL?
> Finally, all functional changes to blink require test coverage. Please add a
set
> of layout tests covering each of the new code blocks.
Okay. I'll work on that. :)

[eae, 2015-03-13 20:23:14]

On 2015/03/13 19:48:48, hichris123 wrote:
> On 2015/03/13 15:19:04, eae wrote:
> > Thanks for working on this,
> > 
> > UBLOCK_GENERAL_PUNCTUATION seems overly broad for this, what makes you think
> > Segoe UI Symbol would be better for punctuation? 
> > 
> > As for bug 225862 ZWJ should be handled by the complex path normalization
> stage
> > and not be dependent on font fallback.
> 
> I was thinking it would be better mostly because of bug 225862, and also
because
> of the boxes on http://www.charbase.com/block/general-punctuation. Segoe UI
> Symbol renders quite a few more of those unicode characters than Lucida Sans
> does.

Thanks for the description, given the expanded coverage keeping it makes sense
to me. I just wanted to make sure you didn't include that to add support for ZWJ
and the like.

> Okay. I'll work on that. :)

Thanks!

[hichris123, 2015-03-14 03:11:08]

On 2015/03/13 20:23:14, eae wrote:
> On 2015/03/13 19:48:48, hichris123 wrote:
> > On 2015/03/13 15:19:04, eae wrote:
> > > Thanks for working on this,
> > > 
> > > UBLOCK_GENERAL_PUNCTUATION seems overly broad for this, what makes you
think
> > > Segoe UI Symbol would be better for punctuation? 
> > > 
> > > As for bug 225862 ZWJ should be handled by the complex path normalization
> > stage
> > > and not be dependent on font fallback.
> > 
> > I was thinking it would be better mostly because of bug 225862, and also
> because
> > of the boxes on http://www.charbase.com/block/general-punctuation. Segoe UI
> > Symbol renders quite a few more of those unicode characters than Lucida Sans
> > does.
> 
> Thanks for the description, given the expanded coverage keeping it makes sense
> to me. I just wanted to make sure you didn't include that to add support for
ZWJ
> and the like.
Ah, gotcha.

> > Okay. I'll work on that. :)
> 
> Thanks!

I finished the test,
usinghttps://src.chromium.org/viewvc/blink?revision=190372&view=revision

[hichris123, 2015-03-14 03:11:09]

On 2015/03/13 20:23:14, eae wrote:
> On 2015/03/13 19:48:48, hichris123 wrote:
> > On 2015/03/13 15:19:04, eae wrote:
> > > Thanks for working on this,
> > > 
> > > UBLOCK_GENERAL_PUNCTUATION seems overly broad for this, what makes you
think
> > > Segoe UI Symbol would be better for punctuation? 
> > > 
> > > As for bug 225862 ZWJ should be handled by the complex path normalization
> > stage
> > > and not be dependent on font fallback.
> > 
> > I was thinking it would be better mostly because of bug 225862, and also
> because
> > of the boxes on http://www.charbase.com/block/general-punctuation. Segoe UI
> > Symbol renders quite a few more of those unicode characters than Lucida Sans
> > does.
> 
> Thanks for the description, given the expanded coverage keeping it makes sense
> to me. I just wanted to make sure you didn't include that to add support for
ZWJ
> and the like.
Ah, gotcha.

> > Okay. I'll work on that. :)
> 
> Thanks!

I finished the test,
usinghttps://src.chromium.org/viewvc/blink?revision=190372&view=revision

[hichris123, 2015-03-14 03:14:42]

On 2015/03/14 03:11:09, hichris123 wrote:
> On 2015/03/13 20:23:14, eae wrote:
> > On 2015/03/13 19:48:48, hichris123 wrote:
> > > On 2015/03/13 15:19:04, eae wrote:
> > > > Thanks for working on this,
> > > > 
> > > > UBLOCK_GENERAL_PUNCTUATION seems overly broad for this, what makes you
> think
> > > > Segoe UI Symbol would be better for punctuation? 
> > > > 
> > > > As for bug 225862 ZWJ should be handled by the complex path
normalization
> > > stage
> > > > and not be dependent on font fallback.
> > > 
> > > I was thinking it would be better mostly because of bug 225862, and also
> > because
> > > of the boxes on http://www.charbase.com/block/general-punctuation. Segoe
UI
> > > Symbol renders quite a few more of those unicode characters than Lucida
Sans
> > > does.
> > 
> > Thanks for the description, given the expanded coverage keeping it makes
sense
> > to me. I just wanted to make sure you didn't include that to add support for
> ZWJ
> > and the like.
> Ah, gotcha.
> 
> > > Okay. I'll work on that. :)
> > 
> > Thanks!
> 
> I finished the test,
> usinghttps://src.chromium.org/viewvc/blink?revision=190372&view=revision

Ugh, it's so easy to make mistakes on mobile. I used r190372's test as a guide,
sticking to symbols wherever possible, and mainly using HTML codes for invisible
characters. All of these symbols rendered with my local build, and I don't
believe I left any symbols out.

I think I need to edit TestExpectations for a rebaseline of this new test & the
older tests that are failing due to the change in font, correct? Thanks!

[hichris123, 2015-03-15 20:42:47]

I also added UBLOCK_MISCELLANEOUS_SYMBOLS to this CL, as ~1/2 are currently
rendered, and all of them render with Segoe UI Symbol.

[hichris123, 2015-03-17 18:22:57]

On 2015/03/15 20:42:47, hichris123 wrote:
> I also added UBLOCK_MISCELLANEOUS_SYMBOLS to this CL, as ~1/2 are currently
> rendered, and all of them render with Segoe UI Symbol.

Running the trybots to see if that will make any other tests need
re-rebaseline'd.

[hichris123, 2015-03-25 01:53:55]

Hi all, any thoughts on the CL (should I have all the failing tests
auto-rebaselined?).

[ebraminio, 2015-03-26 18:16:29]

Hi. I suggest moving MISCELLANEOUS_SYMBOLS and DINGBATS to emojiFont which also
will fallback to symbol. Mozilla send any non BMP character to Emoji font first:
https://bug889401.bugzilla.mozilla.org/attachment.cgi?id=8424616

Per
https://code.google.com/p/chromium/codesearch#chromium/src/third_party/icu/so...
and http://commons.wikimedia.org/wiki/Emoji it would be nice to add and move
UBLOCK_PLAYING_CARDS, UBLOCK_ENCLOSED_ALPHANUMERICS and
UBLOCK_TRANSPORT_AND_MAP_SYMBOLS to the emoji font. Thank you

[hichris123, 2015-03-28 02:00:54]

On 2015/03/26 18:16:29, ebraminio wrote:
> Hi. I suggest moving MISCELLANEOUS_SYMBOLS and DINGBATS to emojiFont which
also
> will fallback to symbol. Mozilla send any non BMP character to Emoji font
first:
> https://bug889401.bugzilla.mozilla.org/attachment.cgi?id=8424616
> 
> Per
>
https://code.google.com/p/chromium/codesearch#chromium/src/third_party/icu/so...
> and http://commons.wikimedia.org/wiki/Emoji it would be nice to add and move
> UBLOCK_PLAYING_CARDS, UBLOCK_ENCLOSED_ALPHANUMERICS and
> UBLOCK_TRANSPORT_AND_MAP_SYMBOLS to the emoji font. Thank you

Interesting. Do you happen to know if there's a specific reason for sending
those to Emoji first? Or just "Firefox did it so we should probably be in line
with them"?

I wonder... having all non-BMP characters fallback to Segoe UI Emoji & Symbol
might be easier... eae@ / pdr@ thoughts? Seems more long term than this CL, but
an interesting thought.

[pdr., 2015-03-28 03:54:00]

On 2015/03/28 at 02:00:54, hichris123 wrote:
> On 2015/03/26 18:16:29, ebraminio wrote:
> > Hi. I suggest moving MISCELLANEOUS_SYMBOLS and DINGBATS to emojiFont which
also
> > will fallback to symbol. Mozilla send any non BMP character to Emoji font
first:
> > https://bug889401.bugzilla.mozilla.org/attachment.cgi?id=8424616
> > 
> > Per
> >
https://code.google.com/p/chromium/codesearch#chromium/src/third_party/icu/so...
> > and http://commons.wikimedia.org/wiki/Emoji it would be nice to add and move
> > UBLOCK_PLAYING_CARDS, UBLOCK_ENCLOSED_ALPHANUMERICS and
> > UBLOCK_TRANSPORT_AND_MAP_SYMBOLS to the emoji font. Thank you
> 
> Interesting. Do you happen to know if there's a specific reason for sending
those to Emoji first? Or just "Firefox did it so we should probably be in line
with them"?
> 
> I wonder... having all non-BMP characters fallback to Segoe UI Emoji & Symbol
might be easier... eae@ / pdr@ thoughts? Seems more long term than this CL, but
an interesting thought.

I'm afraid I don't know much about this area. I'll defer to @eae, Lord High
Admiral of the emoji seas.

[ebraminio, 2015-03-28 13:43:48]

> Interesting. Do you happen to know if there's a specific reason for sending
> those to Emoji first? Or just "Firefox did it so we should probably be in line
> with them"?
Sending them to emoji first will cause they rendered as a colored glyph when it
is supported http://crbug.com/333011 and everyone loves colored emojis :) As far
as I know, this behavior will match with Chrome on Android which currently can
show colored emojis and I guess also with Safari/Chrome/Firefox on other
platforms.

[hichris123, 2015-04-12 23:32:25]

eae@: Any thoughts?

[eae, 2015-04-12 23:58:37]

Change looks good but you'll need to update test expectations.

LGTM

[hichris123, 2015-04-13 00:29:21]

On 2015/04/12 23:58:37, eae wrote:
> Change looks good but you'll need to update test expectations.
> 
> LGTM

I updated the test expectations. If there's no objections, I'll CQ it if the try
bots pass.

[hichris123, 2015-04-13 15:10:27]

The CQ bit was checked by hichris123@gmail.com

[hichris123, 2015-04-13 15:10:28]

The patchset sent to the CQ was uploaded after l-g-t-m from eae@chromium.org
Link to the patchset: https://codereview.chromium.org/1002613002/#ps120001
(title: "Fix TestExpectations")

[commit-bot: I haz the power, 2015-04-13 15:10:53]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1002613002/120001

[commit-bot: I haz the power, 2015-04-13 15:25:48]

Committed patchset #7 (id:120001) as
https://src.chromium.org/viewvc/blink?view=rev&revision=193642