| commit | 47cea9af5548b8c9f6e98f4be0be7d452b6bb306 | [log] [tgz] |
|---|---|---|
| author | Victor Tan <[email protected]> | Mon Apr 04 19:55:48 2022 |
| committer | Chromium LUCI CQ <[email protected]> | Mon Apr 04 19:55:48 2022 |
| tree | 37042e7ae684e0d0a5ea9924ee7e3c691e2dc598 | |
| parent | 71487f2db27c7a58aeedf252331181a3743ea6d9 [diff] [blame] |
Adding metrics comparing Accept-Language with xml:lang and HTML lang values. A follow-up Cl for https://crrev.com/c/3513984, as the majority of sites do not send the Content-Language in the response header. Instead, sites try to use the html `lang` tag to indicate the document language. We are trying to record the values of the language tags to the Accept-Language header to determine: * How many page loads have xml:lang or html lang match one of the accepted languages? * How many page loads have xml:lang or html lang match the top-most preferred language? Why WebLanguageDetectionDetails: WebLanguageDetectionDetails can easily access html lang and `xml:lang` attribute value. Also, we can read the Accept-Language from Prefs efficiently. Proposal doc: [Limited to Google employees] https://docs.google.com/document/d/1bDaCMJP9w6VJjlM_5nRO1KxyB2fkV3JRtKkfUS0GJ_Y Other doc: https://developer.mozilla.org/en-US/docs/Web/HTML/Global_attributes/lang Bug: 1304818 Change-Id: I7d6c7c03e0f44405939273d944aa6a21ac515e40 Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/3553383 Reviewed-by: Nico Weber <[email protected]> Reviewed-by: Max Curran <[email protected]> Reviewed-by: Trevor Perrier <[email protected]> Reviewed-by: Ali Beyad <[email protected]> Commit-Queue: Victor Tan <[email protected]> Cr-Commit-Position: refs/heads/main@{#988628}
diff --git a/components/translate/content/renderer/translate_agent.cc b/components/translate/content/renderer/translate_agent.cc index d6c0b9b..ef6beef 100644 --- a/components/translate/content/renderer/translate_agent.cc +++ b/components/translate/content/renderer/translate_agent.cc
@@ -196,6 +196,9 @@ WebLanguageDetectionDetails web_detection_details = WebLanguageDetectionDetails::CollectLanguageDetectionDetails(document); + WebLanguageDetectionDetails::RecordAcceptLanguageAndXmlHtmlLangMetric( + document); + std::string content_language = web_detection_details.content_language.Utf8(); std::string html_lang = web_detection_details.html_language.Utf8(); std::string model_detected_language;