Description of problem: to type sinhala letter mahaapraana dayanna, ධ in SCIM-SINHALA , the key combination are as follows d+H in the above key sequence. the latenccy in typing the second letter, after pressing the first key results in producing a different output. the unicode value of the letter, which should appear correctly is 0DB0 Version-Release number of selected component (if applicable): scim-sinhala-0.2.0-1.fc6 How reproducible: always Steps to Reproduce: 1.activate scim-sinhala 2.open kedit 3.type d 4.wait for 3-5 seconds 5.type H Actual results: it displays two letters: දහ U+0DAF U+0DC4 (given the unicode values of the two letters) Expected results: ධ the key combination, d+H should produce the charector 0DB0 (unicode value of the charector) Additional info:
Is this with scim qtimm or XIM qtimm?
And is this related to bug 206244?
(In reply to comment #1) > Is this with scim qtimm or XIM qtimm? this is with scim-qtimm-0.9.4.5
(In reply to comment #2) > And is this related to bug 206244? No, this is not related to bug id 206244. i have used Kedit to verify this functionality for Gedit suffers with some bugs and you can't test this issue on Gedit properly. for the bugs exists in Gedit will hamper the tests. this is a commom issue for Gedit/Kedit/ooffice for this is a problem with the input method. Kedit is the best environment to be used for testing for that doesn't have any other bugs ontop of it. so you can use Kedit for testing.
My findings... Gnome Desktop 1) Works with scim-1.4.4-34.fc6 scim-sinhala-0.2.0-1.fc6 pango-1.14.3-1.fc6 2) Not works with scim-1.4.4-34.fc6 scim-sinhala-0.2.0-1.fc6 pango-1.14.3-2.fc6 Remarks - typing "H" in (2) does nothing to "d" KDE 1) Does not works in kdeutils-3.5.4-2.fc6 2) Works in kdeutils-3.5.4-3.fc6 Remarks - the delay, if acused when typing the second char, outputs two characters, rather than modifying the first char. A better way to reproduce this bug would be to type some junk chars "abcd[space]abcd" after typing "d". This would clear any buffers (if present) & would be better than to wait for a random time. Thanks, Mayank
Thanks Mayank, reassigning to pango based on your comments.
It does not smell like a rendering engine problem. Does it sounds like d and H get committed seperately if it is typed slowly? Keep in mind that it may not be 100% reproducible, so it looks like if different minor version of the packages can make a difference? Does cut and paste work? If yes, then it should not be a rendering problem.
Hi Leon, Copy paste from kedit to gedit does works. >Does cut and paste work? If yes, then it should not be a rendering problem. Umm... if it works in Kedit, it should have also worked (or not worked) in the same way with gedit. Behdad, what do you say on this? Thanks, Mayank
(In reply to comment #7) > Does it sounds like d and H get committed seperately if it is typed slowly? yes, this is exactly where the problem is. if you type slowly those two letters in scim-sinhala, the letters d and H , it displays two letters. but it must render just one letter, unicode value of which is 0DB0. after pressing "d", the second letter "H" changes the shape of the first letter. but this is not happenning if you're slow in typing.
Re: comment #8, remember there is a difference in input toolchains as well, namely on the immodule: scim-bridge-gtk vs scim-qtimm; and the toolkit IM API.
(In reply to comment #10) > Re: comment #8, remember there is a difference in input toolchains as well, > namely on the immodule: scim-bridge-gtk vs scim-qtimm; and the toolkit IM API. Aah.. yes, i completely forgot that. I agree with llch.
Latency issue also found in... pango-1.14.4-1.fc6 evolution-2.8.0-5.fc6 scim-1.4.4-34.fc6 scim-sinhala-0.2.0-1.fc6 Test case: Fire evolution, new mail, start scim-sinhala & type "kRoo" After typing "k" from "kRoo", either wait for some time or type random number of scpaces " " so as to clear any buffers/create-latency. Then also try with "kRoo" in one go, ie. without any delay.
Ok the reason for this is very simple: as I suspected scim-sinhala-trans is indeed using a timer between key events - if more than 1s passes then the previous event is forgotten. Tyronne, it is trivial to increase the timeout or even remove it, but I don't know how that we will feel for users?
I added a patch to remove the timeout in scim-sinhala-0.2.0-2.fc6.
Verified in latest FC6: scim-sinhala-0.2.0-2.fc6 scim-1.4.4-34.fc6 It is the expect input in kedit