Speed up adding literal chars when the buffer is known to be one-byte

Speed up adding literal chars when the buffer is known to be one-byte

We have fast-paths in identifier and string scanning that know that so far they have only handled one-byte input. In that case, we can add chars to the literal buffer assuming that it's still in one-byte state, avoiding 2 branches. This seemingly reduces identifier scanning from ~5% to ~4% on octane-codeload.

BUG=

Committed: https://crrev.com/e7de678a2c108949856e3005fe62f43d49aad7ec
Cr-Commit-Position: refs/heads/master@{#36868}

[Toon Verwaest, 2016-06-08 11:43:35]

Description was changed from

==========
Speed up adding literal chars when the buffer is known to be one-byte

We have fast-paths in identifier and string scanning that know that so far they
have only handled one-byte input. In that case, we can add chars to the literal
buffer assuming that it's still in one-byte state, avoiding 2 branches. This
seemingly reduces identifier scanning from ~5% to ~4% on octane-codeload; but
lets wait for bot results to verify.

BUG=
==========

to

==========
Speed up adding literal chars when the buffer is known to be one-byte

We have fast-paths in identifier and string scanning that know that so far they
have only handled one-byte input. In that case, we can add chars to the literal
buffer assuming that it's still in one-byte state, avoiding 2 branches. This
seemingly reduces identifier scanning from ~5% to ~4% on octane-codeload.

BUG=
==========

[Toon Verwaest, 2016-06-08 11:43:35]

verwaest@chromium.org changed reviewers:
+ vogelheim@chromium.org

[Toon Verwaest, 2016-06-08 11:43:42]

ptal

[Toon Verwaest, 2016-06-08 13:32:07]

The CQ bit was checked by verwaest@chromium.org to run a CQ dry run

[commit-bot: I haz the power, 2016-06-08 13:32:27]

Dry run: CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/2044233004/1

[commit-bot: I haz the power, 2016-06-08 14:00:16]

The CQ bit was unchecked by commit-bot@chromium.org

[commit-bot: I haz the power, 2016-06-08 14:00:17]

Dry run: This issue passed the CQ dry run.

[vogelheim, 2016-06-09 12:05:15]

lgtm

https://codereview.chromium.org/2044233004/diff/1/src/parsing/scanner.h
File src/parsing/scanner.h (right):

https://codereview.chromium.org/2044233004/diff/1/src/parsing/scanner.h#newco...
src/parsing/scanner.h:163: DCHECK(is_one_byte_);
Maybe also dcheck whether code_unit is actually ascii. (e.g: DCHECK(0 <=
code_unit && code_unit <= kMaxAscii);

[Toon Verwaest, 2016-06-09 14:01:31]

Addressed comment

[Toon Verwaest, 2016-06-09 14:03:21]

verwaest@chromium.org changed reviewers:
+ adamk@chromium.org

[Toon Verwaest, 2016-06-09 14:03:21]

adamk: ptal as owner of parsing

[adamk, 2016-06-09 14:40:44]

lgtm

[Toon Verwaest, 2016-06-09 14:45:42]

The CQ bit was checked by verwaest@chromium.org

[Toon Verwaest, 2016-06-09 14:45:42]

The patchset sent to the CQ was uploaded after l-g-t-m from
vogelheim@chromium.org
Link to the patchset: https://codereview.chromium.org/2044233004/#ps40001
(title: " ")

[commit-bot: I haz the power, 2016-06-09 14:45:46]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/2044233004/40001

[commit-bot: I haz the power, 2016-06-09 15:12:46]

Description was changed from

==========
Speed up adding literal chars when the buffer is known to be one-byte

We have fast-paths in identifier and string scanning that know that so far they
have only handled one-byte input. In that case, we can add chars to the literal
buffer assuming that it's still in one-byte state, avoiding 2 branches. This
seemingly reduces identifier scanning from ~5% to ~4% on octane-codeload.

BUG=
==========

to

==========
Speed up adding literal chars when the buffer is known to be one-byte

We have fast-paths in identifier and string scanning that know that so far they
have only handled one-byte input. In that case, we can add chars to the literal
buffer assuming that it's still in one-byte state, avoiding 2 branches. This
seemingly reduces identifier scanning from ~5% to ~4% on octane-codeload.

BUG=
==========

[commit-bot: I haz the power, 2016-06-09 15:12:47]

Committed patchset #3 (id:40001)

[commit-bot: I haz the power, 2016-06-09 15:15:03]

Description was changed from

==========
Speed up adding literal chars when the buffer is known to be one-byte

We have fast-paths in identifier and string scanning that know that so far they
have only handled one-byte input. In that case, we can add chars to the literal
buffer assuming that it's still in one-byte state, avoiding 2 branches. This
seemingly reduces identifier scanning from ~5% to ~4% on octane-codeload.

BUG=
==========

to

==========
Speed up adding literal chars when the buffer is known to be one-byte

We have fast-paths in identifier and string scanning that know that so far they
have only handled one-byte input. In that case, we can add chars to the literal
buffer assuming that it's still in one-byte state, avoiding 2 branches. This
seemingly reduces identifier scanning from ~5% to ~4% on octane-codeload.

BUG=

Committed: https://crrev.com/e7de678a2c108949856e3005fe62f43d49aad7ec
Cr-Commit-Position: refs/heads/master@{#36868}
==========

[commit-bot: I haz the power, 2016-06-09 15:15:04]

Patchset 3 (id:??) landed as
https://crrev.com/e7de678a2c108949856e3005fe62f43d49aad7ec
Cr-Commit-Position: refs/heads/master@{#36868}