Support scientific notation in CSS numbers/percentages/dimensions

Support scientific notation in CSS numbers/percentages/dimensions

Previously, scientific notation (scinot) was only supported for SVG
presentation attributes through a special tokenizer quirk. Since scinot
has now been added to the CSS syntax spec [1][2][3] and is already
supported by Gecko [4] and IE(11).
To harmonize, the call to parseSVGNumber(...) is replaced with a call to
charactersToDouble(...). This results in a small behavioral change, since
the former would reject numbers that were too large. These will now parse
as "Infinity" [5].
Since this removes the last caller of isSVGNumberParsingEnabledForMode()
that function is removed as well.

[1] http://dev.w3.org/csswg/css-syntax/#number-token-diagram
[2] http://lists.w3.org/Archives/Public/www-style/2012Aug/0900.html
[3] http://lists.w3.org/Archives/Public/www-style/2013Apr/0428.html
[4] https://bugzilla.mozilla.org/show_bug.cgi?id=964529
[5] Tested browsers were not entirely consistent in their treatment of
    infinities (Gecko: Infinity, IE: FLT_MAX).

BUG=388161
BUG=344645
Committed: https://src.chromium.org/viewvc/blink?view=rev&revision=181560

[fs, 2014-09-05 15:33:43]

fs@opera.com changed reviewers:
+ pdr@chromium.org, timloh@chromium.org

[fs, 2014-09-08 08:28:39]

@timloh: I noticed your comment on issue 344645, but at that point I was pretty
much "done" with this fix already... =) Seeing that both Gecko and IE seem to
support this, and people apparently hitting this (I'm thinking more of the issue
w/ Illustrator in issue 388161 than the one in issue 344645 TBH) I couldn't
really motivate sitting on it.

[Timothy Loh, 2014-09-08 09:24:06]

On 2014/09/08 08:28:39, fs wrote:
> @timloh: I noticed your comment on issue 344645, but at that point I was
pretty
> much "done" with this fix already... =) Seeing that both Gecko and IE seem to
> support this, and people apparently hitting this (I'm thinking more of the
issue
> w/ Illustrator in issue 388161 than the one in issue 344645 TBH) I couldn't
> really motivate sitting on it.

Can we add some tests with CSS properties? Maybe something similar to
css-parser/counter-parsing.html (I'd add an optional parameter to
assert_valid_value for what we expect the value to be when read back).

Also it looks like parseSVGNumber is unused now.

[fs, 2014-09-08 11:10:07]

On 2014/09/08 09:24:06, Timothy Loh wrote:
> On 2014/09/08 08:28:39, fs wrote:
> > @timloh: I noticed your comment on issue 344645, but at that point I was
> pretty
> > much "done" with this fix already... =) Seeing that both Gecko and IE seem
to
> > support this, and people apparently hitting this (I'm thinking more of the
> issue
> > w/ Illustrator in issue 388161 than the one in issue 344645 TBH) I couldn't
> > really motivate sitting on it.
> 
> Can we add some tests with CSS properties? Maybe something similar to
> css-parser/counter-parsing.html (I'd add an optional parameter to
> assert_valid_value for what we expect the value to be when read back).

Added css-parser/scientific-notation.html. Added the optional parameter - wasn't
sure "how optional" it should be though, so the form I chose required a
modification to counter-parsing.html. If you meant the slightly "more optional"
version (only check the read back value if optional param. is passed) I can
change it to that.
 
> Also it looks like parseSVGNumber is unused now.

Good catch. Removed.

[Timothy Loh, 2014-09-08 11:29:26]

lgtm

[fs, 2014-09-08 13:54:13]

The CQ bit was checked by fs@opera.com

[commit-bot: I haz the power, 2014-09-08 13:55:19]

CQ is trying da patch. Follow status at
 https://chromium-status.appspot.com/cq/fs@opera.com/545173003/20001

[commit-bot: I haz the power, 2014-09-08 13:59:16]

Committed patchset #2 (id:20001) as 181560