Bug 191003 – [GTK] Layout test imported/w3c/web-platform-tests/mimesniff/mime-types/charset-parameter.window.html is failing

• Component: WebKitGTK
• Status: NEW

[Michael Catanzaro]

Layout test imported/w3c/web-platform-tests/mimesniff/mime-types/charset-parameter.window.html has been failing since it was added in r236787 "Import WPT mimesniff resources". Text diff:

--- /home/slave/webkitgtk/gtk-linux-64-release-tests/build/layout-test-results/imported/w3c/web-platform-tests/mimesniff/mime-types/charset-parameter.window-expected.txt
+++ /home/slave/webkitgtk/gtk-linux-64-release-tests/build/layout-test-results/imported/w3c/web-platform-tests/mimesniff/mime-types/charset-parameter.window-actual.txt
@@ -6,14 +6,14 @@
 PASS text/html;x=(;charset=gbk 
 PASS text/html;charset=gbk;charset=windows-1255 
 PASS text/html;charset=();charset=GBK 
-PASS text/html;charset =gbk 
+FAIL text/html;charset =gbk assert_equals: expected "UTF-8" but got "GBK"
 PASS text/html ;charset=gbk 
 PASS text/html; charset=gbk 
 PASS text/html;charset= gbk 
 FAIL text/html;charset= "gbk" assert_equals: expected "UTF-8" but got "GBK"
-PASS text/html;charset='gbk' 
-PASS text/html;charset='gbk 
-PASS text/html;charset=gbk' 
+FAIL text/html;charset='gbk' assert_equals: expected "UTF-8" but got "GBK"
+FAIL text/html;charset='gbk assert_equals: expected "UTF-8" but got "GBK"
+FAIL text/html;charset=gbk' assert_equals: expected "UTF-8" but got "GBK"
 PASS text/html;charset=';charset=GBK 
 PASS text/html;test;charset=gbk 
 PASS text/html;test=;charset=gbk 
@@ -24,13 +24,13 @@
 FAIL text/html;charset= ";charset=GBK assert_equals: expected "GBK" but got "UTF-8"
 FAIL text/html;charset=";charset=foo";charset=GBK assert_equals: expected "GBK" but got "UTF-8"
 PASS text/html;charset="gbk" 
-FAIL text/html;charset="gbk assert_equals: expected "GBK" but got "UTF-8"
+PASS text/html;charset="gbk 
 PASS text/html;charset=gbk" 
-FAIL text/html;charset=" gbk" assert_equals: expected "GBK" but got "UTF-8"
-FAIL text/html;charset="gbk " assert_equals: expected "GBK" but got "UTF-8"
-FAIL text/html;charset="\ gbk" assert_equals: expected "GBK" but got "UTF-8"
-FAIL text/html;charset="\g\b\k" assert_equals: expected "GBK" but got "UTF-8"
-FAIL text/html;charset="gbk"x assert_equals: expected "GBK" but got "UTF-8"
+PASS text/html;charset=" gbk" 
+PASS text/html;charset="gbk " 
+PASS text/html;charset="\ gbk" 
+PASS text/html;charset="\g\b\k" 
+PASS text/html;charset="gbk"x 
 FAIL text/html;charset="";charset=GBK assert_equals: expected "GBK" but got "UTF-8"
 PASS text/html;charset=";charset=GBK 
 PASS text/html;charset={gbk} 


I'm not sure what's up, but I'll caution that encoding detection is probabilistic and it's expected that different implementations (e.g. ICU vs. uchardet) will return different results for the same input.