Consider merging https://chromium.googlesource.com/chromium/blink/+/873a4a1f950d7cd7ebe27baba2c0e7e9491190f4 This pattern saved memory allocation at the expense of bloating code size. Prefer small code size when these aren't bottlenecks. Saves 72 bytes per string, for a total of 2448 bytes for this change. We might want to look for the same code elsewhere in WebCore.
Reverted in https://chromium.googlesource.com/chromium/blink/+/773e141f0deb3abd2492e1f41dcdd65bd68b2662.
Re-landed in https://chromium.googlesource.com/chromium/blink/+/198dcfa2035c8a92971ffc7cd100840eae9e2b36
Do we prefer size over speed or vice versa Ryosuke?
(In reply to comment #3) > Do we prefer size over speed or vice versa Ryosuke? That's a good question. I think the argument here is that this code isn't performance critical.
<rdar://problem/16014102>
(In reply to comment #3) > Do we prefer size over speed or vice versa Ryosuke? Static also have overhead. I doubt these are justified by speed.