New issue
Advanced search Search tips
Note: Color blocks (like or ) mean that a user may not be available. Tooltip shows the reason.

Issue 765006 link

Starred by 10 users

Issue metadata

Status: Available
Owner: ----
Cc:
Components:
EstimatedDays: ----
NextAction: ----
OS: ----
Pri: 3
Type: Task



Sign in to add a comment

Reduce incorrect language identification in Chrome

Project Member Reported by napper@chromium.org, Sep 14 2017

Issue description

Cc: yyushkina@chromium.org
Project Member

Comment 2 by bugdroid1@chromium.org, Oct 16 2017

The following revision refers to this bug:
  https://chromium.googlesource.com/chromium/src.git/+/bbf38447b31ce7e81bd06ff8c9ae84c5f2426ead

commit bbf38447b31ce7e81bd06ff8c9ae84c5f2426ead
Author: Jon Napper <napper@chromium.org>
Date: Mon Oct 16 04:37:53 2017

Added missing well-known conflicting language codes

Added missing well-known conflicting language codes to whitelist of
codes that are often incorrect (see
go/reducing-language-detection-conflicts)

Bug: 765006
Change-Id: I159be446d6521c356461187f94ed372a7494cafc
Reviewed-on: https://chromium-review.googlesource.com/720578
Reviewed-by: Jia Meng <jiameng@chromium.org>
Commit-Queue: Jon Napper <napper@chromium.org>
Cr-Commit-Position: refs/heads/master@{#508983}
[modify] https://crrev.com/bbf38447b31ce7e81bd06ff8c9ae84c5f2426ead/components/translate/core/language_detection/language_detection_util.cc

Project Member

Comment 3 by bugdroid1@chromium.org, Oct 18 2017

The following revision refers to this bug:
  https://chromium.googlesource.com/chromium/src.git/+/be3a712c6c8f56b9195f36c642c04d2c6f94f97d

commit be3a712c6c8f56b9195f36c642c04d2c6f94f97d
Author: Jon Napper <napper@chromium.org>
Date: Wed Oct 18 23:24:02 2017

Revert "Added missing well-known conflicting language codes"

This reverts commit bbf38447b31ce7e81bd06ff8c9ae84c5f2426ead.

Reason for revert: Possible performance regression, https://bugs.chromium.org/p/chromium/issues/detail?id=775828#c3

Original change's description:
> Added missing well-known conflicting language codes
> 
> Added missing well-known conflicting language codes to whitelist of
> codes that are often incorrect (see
> go/reducing-language-detection-conflicts)
> 
> Bug: 765006
> Change-Id: I159be446d6521c356461187f94ed372a7494cafc
> Reviewed-on: https://chromium-review.googlesource.com/720578
> Reviewed-by: Jia Meng <jiameng@chromium.org>
> Commit-Queue: Jon Napper <napper@chromium.org>
> Cr-Commit-Position: refs/heads/master@{#508983}

TBR=napper@chromium.org,jiameng@chromium.org

# Not skipping CQ checks because original CL landed > 1 day ago.

Bug: 765006
Change-Id: Id466d355e1468bc9784f613af5b94212549c1d18
Reviewed-on: https://chromium-review.googlesource.com/726680
Reviewed-by: Jia Meng <jiameng@chromium.org>
Commit-Queue: Jon Napper <napper@chromium.org>
Cr-Commit-Position: refs/heads/master@{#509921}
[modify] https://crrev.com/be3a712c6c8f56b9195f36c642c04d2c6f94f97d/components/translate/core/language_detection/language_detection_util.cc

Project Member

Comment 4 by bugdroid1@chromium.org, Oct 30 2017

The following revision refers to this bug:
  https://chromium.googlesource.com/chromium/src.git/+/1ed2ebd683e0f3cd16c5b9e0fe182eafa577b75c

commit 1ed2ebd683e0f3cd16c5b9e0fe182eafa577b75c
Author: Jon Napper <napper@chromium.org>
Date: Mon Oct 30 00:51:36 2017

Added missing well-known conflicting language codes.

This patch adds missing well-known conflicting language codes. It also
changes the lookup from a linear search to a binary search to improve
performance (a previous version caused a performance regression).

Bug: 765006
Change-Id: I0271eb620d0f412f168b4767de584e4896f86693
Reviewed-on: https://chromium-review.googlesource.com/727779
Commit-Queue: Jon Napper <napper@chromium.org>
Reviewed-by: Jia Meng <jiameng@chromium.org>
Cr-Commit-Position: refs/heads/master@{#512427}
[modify] https://crrev.com/1ed2ebd683e0f3cd16c5b9e0fe182eafa577b75c/chrome/browser/translate/translate_manager_browsertest.cc
[add] https://crrev.com/1ed2ebd683e0f3cd16c5b9e0fe182eafa577b75c/chrome/test/data/french_page_lang_conflict.html
[add] https://crrev.com/1ed2ebd683e0f3cd16c5b9e0fe182eafa577b75c/chrome/test/data/korean_page_lang_conflict.html
[modify] https://crrev.com/1ed2ebd683e0f3cd16c5b9e0fe182eafa577b75c/components/translate/core/language_detection/language_detection_util.cc
[modify] https://crrev.com/1ed2ebd683e0f3cd16c5b9e0fe182eafa577b75c/components/translate/core/language_detection/language_detection_util.h
[modify] https://crrev.com/1ed2ebd683e0f3cd16c5b9e0fe182eafa577b75c/components/translate/core/language_detection/language_detection_util_unittest.cc

Cc: groby@chromium.org durga.behera@chromium.org napper@chromium.org djweiss@chromium.org riesa@chromium.org talo@chromium.org kavvaru@chromium.org abakalov@chromium.org hdodda@chromium.org
 Issue 678398  has been merged into this issue.
Status: Fixed (was: Assigned)
Cc: mdw@chromium.org
Issue 771861 has been merged into this issue.
Project Member

Comment 8 by bugdroid1@chromium.org, Mar 20 2018

The following revision refers to this bug:
  https://chromium.googlesource.com/chromium/src.git/+/1babd8a2883a189e33ad9892842825c73aa56a25

commit 1babd8a2883a189e33ad9892842825c73aa56a25
Author: Michael Martis <martis@chromium.org>
Date: Tue Mar 20 01:10:47 2018

Reverted additions to the "well-known conflicting language codes" list.

Even our optimized version of these changes caused multiple performance
regression warnings:  crbug.com/780176 ,  crbug.com/780174 .

Bug: 765006
Change-Id: I59924ff464a6161cf0afee332f2ceb6c40c7ee0b
Reviewed-on: https://chromium-review.googlesource.com/969681
Reviewed-by: Andrew Moylan <amoylan@chromium.org>
Commit-Queue: Michael Martis <martis@chromium.org>
Cr-Commit-Position: refs/heads/master@{#544244}
[modify] https://crrev.com/1babd8a2883a189e33ad9892842825c73aa56a25/components/translate/core/language_detection/language_detection_util.cc
[modify] https://crrev.com/1babd8a2883a189e33ad9892842825c73aa56a25/components/translate/core/language_detection/language_detection_util_unittest.cc

Comment 9 by martis@chromium.org, Mar 20 2018

Owner: yyushkina@chromium.org
Status: Available (was: Fixed)
I've reverted our recent additions to the "known conflicting language codes" list, since both approaches we've tried have caused performance regression warnings.

Yana: up to you how you'd like to proceed.
Cc: anthonyvd@chromium.org
Labels: -Pri-2 Pri-3
Owner: ----
Any additional detail on what went wrong? What kinds of performance regressions did we see?

Sign in to add a comment