De-ExceptionState-ify SVGLengthContext

De-ExceptionState-ify SVGLengthContext

The 1.1 specification does not define if exceptions should be thrown
when font-size/x-height/viewport cannot be determined (and hence
relative units not properly evaluated).
UAs don't really agree here and throw/don't throw in different cases.
The gist of this CL is that most of the uses of SVGLengthContext and
SVGLength::value() will not be throwing exceptions at all, so move that
part of the functionality to the callers that might care
(ie SVGLengthTearOff). This in essence checks for the primary failure
modes up front and then throws without calling into SVGLengthContext
at all. Some fidelity in exception messages is sacrificed, and cases
which shouldn't be considered errors (like zero font-size) no longer
cause exceptions to be thrown.

Reduces binary-size by ~15k (only a smaller part of that in
SVGLengthContext itself).

Committed: https://src.chromium.org/viewvc/blink?view=rev&revision=189659

[fs, 2015-02-05 17:10:22]

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

[fs, 2015-02-05 17:15:45]

Was reminded that I had this lying around...

Played a bit [1] with the API to see what different UAs did, only to discover
it's a bit of a mess...

Initially I figured I'd just drop the exceptions altogether, but ended up
landing on this version - which should try to be somewhat deterministic [2].
Decided on plugging those holes separately though...

[1] http://jsfiddle.net/uLByx41r/
[2] There's still randomness involved depending on if a layout has happened or
not etc.

[pdr., 2015-02-05 22:12:32]

LGTM! I tried this many moons ago and gave up when I hit the exception changes.

https://codereview.chromium.org/901193002/diff/1/Source/core/svg/SVGLength.h
File Source/core/svg/SVGLength.h (right):

https://codereview.chromium.org/901193002/diff/1/Source/core/svg/SVGLength.h#...
Source/core/svg/SVGLength.h:84: return unitType == LengthTypePercentage
(for another patch)
Does this need to be updated? http://www.w3.org/TR/css3-values/#relative-lengths

https://codereview.chromium.org/901193002/diff/1/Source/core/svg/SVGLengthTea...
File Source/core/svg/SVGLengthTearOff.cpp (right):

https://codereview.chromium.org/901193002/diff/1/Source/core/svg/SVGLengthTea...
Source/core/svg/SVGLengthTearOff.cpp:155:
exceptionState.throwDOMException(NotSupportedError, "No context could be
found.");
Bikeshed: is the concept of a length context something users would know about?
We may just want to say "Could not resolve relative length".

[fs, 2015-02-06 12:25:13]

New patchsets have been uploaded after l-g-t-m from pdr@chromium.org

[fs, 2015-02-06 12:26:04]

https://codereview.chromium.org/901193002/diff/1/Source/core/svg/SVGLength.h
File Source/core/svg/SVGLength.h (right):

https://codereview.chromium.org/901193002/diff/1/Source/core/svg/SVGLength.h#...
Source/core/svg/SVGLength.h:84: return unitType == LengthTypePercentage
On 2015/02/05 22:12:32, pdr wrote:
> (for another patch)
> Does this need to be updated?
http://www.w3.org/TR/css3-values/#relative-lengths

I will need an update when SVGLength actually buys the new fancy pants that CSS3
values and units offer. For now it's still dull, gray and old old though...

https://codereview.chromium.org/901193002/diff/1/Source/core/svg/SVGLengthTea...
File Source/core/svg/SVGLengthTearOff.cpp (right):

https://codereview.chromium.org/901193002/diff/1/Source/core/svg/SVGLengthTea...
Source/core/svg/SVGLengthTearOff.cpp:155:
exceptionState.throwDOMException(NotSupportedError, "No context could be
found.");
On 2015/02/05 22:12:32, pdr wrote:
> Bikeshed: is the concept of a length context something users would know about?
> We may just want to say "Could not resolve relative length".

Funny you should say that, because that's pretty much what I intended to change
it too before I decided to be lazy... =P

[fs, 2015-02-06 13:43:55]

The CQ bit was checked by fs@opera.com

[commit-bot: I haz the power, 2015-02-06 13:44:11]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/901193002/40001

[commit-bot: I haz the power, 2015-02-06 18:02:29]

Committed patchset #3 (id:40001) as
https://src.chromium.org/viewvc/blink?view=rev&revision=189659