Micro-optimise Length::initFromLength

Micro-optimise Length::initFromLength

This appears to improve html5-full-render by 1-2% on my system (gcc 4.6.3) by using memcpy instead of copying members (and branching to copy the union).

R=eseidel@chromium.org
BUG=

Committed: https://src.chromium.org/viewvc/blink?view=rev&revision=148621

[Eric Willigers, 2013-04-18 03:01:50]

lgtm

On Intel with g++ -O3, this change replaces the byte reads and writes, and the
branch to read and write the union contents, with a movq read and a movq write.

[eseidel, 2013-04-18 05:17:32]

This is great, but I'm confused why.  Doesn't one normally define either the
copy constructor in terms of the assignment operator or vice versa?  Why do we
have this 3rd helper function to do so?  I guess the whole point of this is to
do the incrementCalculatedRef?

This stuff is incredibly hot code and important that we get right, so I'm glad
you looked at this.  lgtm.

[commit-bot: I haz the power, 2013-04-18 05:22:04]

CQ is trying da patch. Follow status at
https://chromium-status.appspot.com/cq/timloh@chromium.org/14283003/1

[Timothy Loh, 2013-04-18 05:27:37]

On 2013/04/18 05:17:32, Eric Seidel (Google) wrote:
> This is great, but I'm confused why.  Doesn't one normally define either the
> copy constructor in terms of the assignment operator or vice versa?  Why do we
> have this 3rd helper function to do so?  I guess the whole point of this is to
> do the incrementCalculatedRef?

Yep, it was added when calculated values were added to Length objects. Actually
looking at it again it seems that Length::operator= should also see if it needs
to decrement, so there might be a leak there -- I'll look into it.

[commit-bot: I haz the power, 2013-04-18 05:55:59]

Retried try job too often on mac_layout_rel for step(s) webkit_tests
http://build.chromium.org/p/tryserver.chromium/buildstatus?builder=mac_layout...

[commit-bot: I haz the power, 2013-04-18 06:01:03]

CQ is trying da patch. Follow status at
https://chromium-status.appspot.com/cq/timloh@chromium.org/14283003/1

[commit-bot: I haz the power, 2013-04-18 06:22:43]

Change committed as 148621