Shrink canonicalizedTitle a bit.

Shrink canonicalizedTitle a bit.

Noticed that it was (in content_shell x64 for Linux) 421 and 425 bytes
which seemed excessive. Just rewrote it a bit to do less work for the
same result.

gcc (not even going to guess why createQualifiedName changed):
---------------------------------------------------------------------------------------------------------------------
 -635 - Source: /home/bratell/src/chromium/src/third_party/WebKit/Source/core/dom/Document.cpp - (gained 8, lost 643)
---------------------------------------------------------------------------------------------------------------------
  Grown symbols:
         +8: WebCore::Document::didChangeVisibilityState() type=t, (was 299 bytes, now 307 bytes)
  Shrunk symbols:
        -64: WebCore::createQualifiedName(WTF::AtomicString const&, WTF::AtomicString const&, WebCore::ExceptionState&) type=t, (was 1112 bytes, now 1048 bytes)
       -579: WebCore::Document::updateTitle(WTF::String const&) type=t, (was 1541 bytes, now 962 bytes)

clang:
  Shrunk symbols:
       -241: WebCore::Document::updateTitle(WTF::String const&) type=t, (was 1124 bytes, now 883 bytes)

Committed: https://src.chromium.org/viewvc/blink?view=rev&revision=181746

[Daniel Bratell, 2014-07-21 12:12:47]

esprehn, can you take a look at this micro cleanup patch? Basically a followup
of your title cleanup last year.

(This is really a random change, I was studying bloat tables and just got
curious what the function did)

[Daniel Bratell, 2014-07-21 12:14:10]

Ah, I changed StringBuffer as well (removed an unused method and
un-force-inlined shrink) so an owner of WTF need to review. Adding eseidel.

[esprehn, 2014-08-01 14:13:25]

Can we do this in two changes? First change WTF and then change caller? Also do
you need to add WTF unit tests?

[Daniel Bratell, 2014-08-01 14:24:24]

On 2014/08/01 14:13:25, esprehn wrote:
> Can we do this in two changes? First change WTF and then change caller? Also
do
> you need to add WTF unit tests?

Sure can, though I'm leaving for vacation in minus 23 minutes so it will take a
while. I don't know if there already are unit tests for shrink or not. Possibly
not.

[esprehn, 2014-08-19 02:01:07]

How goes this?

[Daniel Bratell, 2014-08-29 13:39:32]

On 2014/08/19 02:01:07, esprehn wrote:
> How goes this?

Looked at it now. Splitting into two commits is hard since the cleanup in
StringBuffer depends partly on that canonicalizedTitle stops using
StringBuffer::resize(). It can be split into three commits though.

I haven't found any existing unit tests at all so I will produce some.

[Daniel Bratell, 2014-08-29 14:40:21]

On 2014/08/29 13:39:32, Daniel Bratell wrote:
> On 2014/08/19 02:01:07, esprehn wrote:
> > How goes this?
> 
> Looked at it now. Splitting into two commits is hard since the cleanup in
> StringBuffer depends partly on that canonicalizedTitle stops using
> StringBuffer::resize(). It can be split into three commits though.
> 
> I haven't found any existing unit tests at all so I will produce some.

That part is in https://codereview.chromium.org/520723002/
As soon as that change has landed I will rebase this one.

[Daniel Bratell, 2014-09-01 14:59:39]

bratell@opera.com changed reviewers:
+ jochen@chromium.org

[Daniel Bratell, 2014-09-01 14:59:40]

The outlining of shrink has now landed.

jochen, can you please take a look at the wtf change? (removal of an unused
method).

esprehn, does the Document change look fine to you?

[jochen (gone - plz use gerrit), 2014-09-02 08:21:58]

wtf/ lgtm

[Daniel Bratell, 2014-09-10 15:47:58]

The CQ bit was checked by bratell@opera.com

[commit-bot: I haz the power, 2014-09-10 15:48:40]

CQ is trying da patch. Follow status at
 https://chromium-status.appspot.com/cq/bratell@opera.com/390003003/100001

[commit-bot: I haz the power, 2014-09-10 16:51:22]

Committed patchset #4 (id:100001) as 181746