Use StringView for String and AtomicString operator==.

Use StringView for String and AtomicString operator==.

This allows the compiler to compute the length of literal strings that
are compared and then we can early abort comparing if the length is
different.

This does introduce a strlen call for non-literal strings, which we
have very few of today.

This patch fixes the web idl enum checking loop to use WTF::equal to
avoid the strlen since that seemed like an easy way to remove lots of strlen
calls. We can make other callers use a similar approach in the future if
if we see strlen on profiles.

By code auditing (with code search method finding) I noticed a few other
places will now call strlen (ex. iframe sandbox and permissions attribute
parsing) but this patch opts not to convert all places we now call strlen and 
instead go for the simple path of fixing those if we run into problems. This
is a similar trade off to code elsewhere in Chromium that uses
base::StringPiece which also puts a strlen call for non-literal strings.

I also took this as an opportunity to remove the operator== methods which aren't
used for things like LChar* (strangely UChar* was missing) and Vector<char>.

BUG=678266
Review-Url: https://codereview.chromium.org/2614123002
Cr-Commit-Position: refs/heads/master@{#441914}
Committed: https://chromium.googlesource.com/chromium/src/+/e827c1a786ea596461b2c76ba6dafffd40f1bec8

[esprehn, 2017-01-06 04:39:25]

Description was changed from

==========
Use StringView for String and AtomicString operator==.

This allows the compiler to compute the length of literal strings that
are compared and then we can early abort comparing if the length is
different.

This does introduce a strlen call for non-literal strings, which we
have very few of today.

This patch fixes the web idl enum checking loop to use WTF::equal to
avoid the strlen. We can make other callers use that in the future if
we see strlen on profiles. A few other places will now call strlen (ex.
iframe sandbox and permissions attribute parsing) but this patch opts
not to convert all places we now call strlen and instead go for the simple
path of fixing those if we run into problems. This is a similar trade to
code elsewhere in Chromium that uses base::StringPiece which also puts
a strlen call for non-literal strings.

BUG=678266
==========

to

==========
Use StringView for String and AtomicString operator==.

This allows the compiler to compute the length of literal strings that
are compared and then we can early abort comparing if the length is
different.

This does introduce a strlen call for non-literal strings, which we
have very few of today.

This patch fixes the web idl enum checking loop to use WTF::equal to
avoid the strlen since that seemed like an easy way to remove lots of strlen
calls. We can make other callers use a similar approach in the future if
if we see strlen on profiles.

By code auditing (with code search method finding) I noticed a few other
places will now call strlen (ex. iframe sandbox and permissions attribute
parsing) but this patch opts not to convert all places we now call strlen and 
instead go for the simple path of fixing those if we run into problems. This
is a similar trade off to code elsewhere in Chromium that uses
base::StringPiece which also puts a strlen call for non-literal strings.

BUG=678266
==========

[esprehn, 2017-01-06 04:42:56]

esprehn@chromium.org changed reviewers:
+ csharrison@chromium.org

[esprehn, 2017-01-06 04:44:37]

Description was changed from

==========
Use StringView for String and AtomicString operator==.

This allows the compiler to compute the length of literal strings that
are compared and then we can early abort comparing if the length is
different.

This does introduce a strlen call for non-literal strings, which we
have very few of today.

This patch fixes the web idl enum checking loop to use WTF::equal to
avoid the strlen since that seemed like an easy way to remove lots of strlen
calls. We can make other callers use a similar approach in the future if
if we see strlen on profiles.

By code auditing (with code search method finding) I noticed a few other
places will now call strlen (ex. iframe sandbox and permissions attribute
parsing) but this patch opts not to convert all places we now call strlen and 
instead go for the simple path of fixing those if we run into problems. This
is a similar trade off to code elsewhere in Chromium that uses
base::StringPiece which also puts a strlen call for non-literal strings.

BUG=678266
==========

to

==========
Use StringView for String and AtomicString operator==.

This allows the compiler to compute the length of literal strings that
are compared and then we can early abort comparing if the length is
different.

This does introduce a strlen call for non-literal strings, which we
have very few of today.

This patch fixes the web idl enum checking loop to use WTF::equal to
avoid the strlen since that seemed like an easy way to remove lots of strlen
calls. We can make other callers use a similar approach in the future if
if we see strlen on profiles.

By code auditing (with code search method finding) I noticed a few other
places will now call strlen (ex. iframe sandbox and permissions attribute
parsing) but this patch opts not to convert all places we now call strlen and 
instead go for the simple path of fixing those if we run into problems. This
is a similar trade off to code elsewhere in Chromium that uses
base::StringPiece which also puts a strlen call for non-literal strings.

I also took this as an opportunity to remove the operator== methods which aren't
used for things like LChar* (strangely UChar* was missing) and Vector<char>.

BUG=678266
==========

[haraken, 2017-01-06 05:13:33]

LGTM

[Charlie Harrison, 2017-01-06 05:22:20]

Yeah I think this is better than having the StringView as an argument to the
operator==. As long as the function is inlined we should get the strlen
optimization.

Do we need ALWAYS_INLINE to ensure this? I'm not fully sure if any platforms
fail to inline on official builds without it.

In any case, LGTM

[esprehn, 2017-01-06 07:05:22]

On 2017/01/06 at 05:22:20, csharrison wrote:
> Yeah I think this is better than having the StringView as an argument to the
operator==. As long as the function is inlined we should get the strlen
optimization.
> 
> Do we need ALWAYS_INLINE to ensure this? I'm not fully sure if any platforms
fail to inline on official builds without it.
> 
> In any case, LGTM

Those functions are so small I've never seen a build without it being inlined.
If we find one we can add an always inline macro.

[esprehn, 2017-01-06 07:05:30]

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

[commit-bot: I haz the power, 2017-01-06 07:05:39]

Dry run: CQ is trying da patch. Follow status at
 
https://chromium-cq-status.appspot.com/v2/patch-status/codereview.chromium.or...

[commit-bot: I haz the power, 2017-01-06 08:40:11]

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

[commit-bot: I haz the power, 2017-01-06 08:40:12]

Dry run: This issue passed the CQ dry run.

[esprehn, 2017-01-06 10:42:34]

The CQ bit was checked by esprehn@chromium.org

[commit-bot: I haz the power, 2017-01-06 10:42:46]

CQ is trying da patch. Follow status at
 
https://chromium-cq-status.appspot.com/v2/patch-status/codereview.chromium.or...

[commit-bot: I haz the power, 2017-01-06 10:45:51]

CQ is committing da patch.

Bot data: {"patchset_id": 1, "attempt_start_ts": 1483699354454630, "parent_rev":
"555f939f4c962e9e942b1f9b9f02ee98e554ead3", "commit_rev":
"e827c1a786ea596461b2c76ba6dafffd40f1bec8"}

[commit-bot: I haz the power, 2017-01-06 10:46:21]

Description was changed from

==========
Use StringView for String and AtomicString operator==.

This allows the compiler to compute the length of literal strings that
are compared and then we can early abort comparing if the length is
different.

This does introduce a strlen call for non-literal strings, which we
have very few of today.

This patch fixes the web idl enum checking loop to use WTF::equal to
avoid the strlen since that seemed like an easy way to remove lots of strlen
calls. We can make other callers use a similar approach in the future if
if we see strlen on profiles.

By code auditing (with code search method finding) I noticed a few other
places will now call strlen (ex. iframe sandbox and permissions attribute
parsing) but this patch opts not to convert all places we now call strlen and 
instead go for the simple path of fixing those if we run into problems. This
is a similar trade off to code elsewhere in Chromium that uses
base::StringPiece which also puts a strlen call for non-literal strings.

I also took this as an opportunity to remove the operator== methods which aren't
used for things like LChar* (strangely UChar* was missing) and Vector<char>.

BUG=678266
==========

to

==========
Use StringView for String and AtomicString operator==.

This allows the compiler to compute the length of literal strings that
are compared and then we can early abort comparing if the length is
different.

This does introduce a strlen call for non-literal strings, which we
have very few of today.

This patch fixes the web idl enum checking loop to use WTF::equal to
avoid the strlen since that seemed like an easy way to remove lots of strlen
calls. We can make other callers use a similar approach in the future if
if we see strlen on profiles.

By code auditing (with code search method finding) I noticed a few other
places will now call strlen (ex. iframe sandbox and permissions attribute
parsing) but this patch opts not to convert all places we now call strlen and 
instead go for the simple path of fixing those if we run into problems. This
is a similar trade off to code elsewhere in Chromium that uses
base::StringPiece which also puts a strlen call for non-literal strings.

I also took this as an opportunity to remove the operator== methods which aren't
used for things like LChar* (strangely UChar* was missing) and Vector<char>.

BUG=678266

Review-Url: https://codereview.chromium.org/2614123002
Cr-Commit-Position: refs/heads/master@{#441914}
Committed:
https://chromium.googlesource.com/chromium/src/+/e827c1a786ea596461b2c76ba6da...
==========

[commit-bot: I haz the power, 2017-01-06 10:46:21]

Committed patchset #1 (id:1) as
https://chromium.googlesource.com/chromium/src/+/e827c1a786ea596461b2c76ba6da...