Description of problem: When visit the URL http://focus.bne.redhat.com/I18N/IndicTC/lang/te_IN/font/indicFonctTestCaseGSUB-Telugu.html via the firefox browser, and test the #285 font there were the cursoring error for that I must knock three times the left /right cursor to make the cursor move to the left /right with the current font , and when use the BackSpace key to delete the word ,it could not match the number of the unicode.More over,we must knock the Delete key three times to delete the word. What's more ,in the konqueror browser, the #285 word seems not the same with the image. The Unicode of the word is U+0C37 U+0C4D U+0C1F U+0C4D U+0C30 U+0C40 Version-Release number of selected component (if applicable): fonts-telugu-2.3.1-1.el5 firefox-3.0-0.beta5.6.el5 konqueror in rhel 5.2 pango-1.14.9-3.el5 How reproducible: Always Steps to Reproduce: 1.visit the website given above and turn to the te_IN GSUB #285 ,and just knock the Backspace key according to the unicode (knock the Backspace for time s the same as the number of the unicode of the word) 2.Knock the Delete key on the word just for once 3.Knock the left/right key on the word just for once Actual results: The cursor can't move as the requirement below: Delete:it should delete the entire composed char at a time Backspace:need to several times to remove the char, it depends on each char unicode cursor(left,right): must not be positioned at the middle of the char anyhow,so cursor just for one time Expected results: All the cursor s can work as the requirement. Additional info:
Created attachment 317455 [details] This is the page displayed in the firefox
Created attachment 317457 [details] This is the page displayed in the konqueror
I think the correct link to the test cases is, http://focus.bne.redhat.com/I18N/IndicTC/lang/te_IN/font/IndicFontTestCaseGSUB-Telugu.html
This should be rendering problem? If so, the bugs should be submitted against the rendering engine of both browser.
I've just separated this bug into two different bugs,#463891 and #463892 because they are really two different bugs either in firefox or in konqueror.So just check them.
Then what happen to this one?
This is not really a bug with pango, firefox or konqueror. Its actually a bug in test case data and also a bug in font. The reasoning as follows: When the character sequence given in the test case (U+0C37 U+0C4D U+0C1F U+0C4D U+0C30 U+0C40) is typed into gedit or kedit, using either Inscript layout or raw code, the bug is not reproduced. When verified the sequence that is actually entered in the test case, it is found that the test case actually has a wrong sequence entered. The sequence of characters present in the test case is (U+0C37 U+0C1F U+0C30 U+0C40), i.e. without U+0C4D as opposed to what it is supposed to be. Still the wrong sequence being rendered as a (fake) correct one, is also a bug in font. But we need to correct the character sequence used in the test case as well. I will keep this bug open until the wrong sequence is removed from the font. And close the other two that are reported for this one.
Thanks for the analysis Rahul. Will have the Test Case updated.