Created attachment 111289 [details] test case Steps to reproduce: 1. Open the attached document 2. Launch the attached accessible-event listener in a terminal 3. Arrow amongst the characters at the top of the document 4. Arrow amongst the characters in the table Expected results: Accessible caret-moved events would be seen each time the caret moves. Actual results: Accessible caret-moved events are seen when the caret moves outside the table; but no caret-moved events are seen when the caret moves inside the table.
Created attachment 111290 [details] accessible-event listener
The bug is still here in 4.3.5, e.g. the latest Evolution available on the TDF website. Regards,
Comment on attachment 111289 [details] test case fix mimetype
TESTING with Ubuntu 14.04 + LO 4.3.5.2 and 4.4.0.1 (In reply to Joanmarie Diggs from comment #0) > Steps to reproduce: > 1. Open attachment 111289 [details] > 2. Launch the attached accessible-event listener in a terminal > 3. Arrow amongst the characters at the top of the document > 4. Arrow amongst the characters in the table > > Expected results: Accessible caret-moved events would be seen each time the > caret moves. Yes > Actual results: Accessible caret-moved events are seen when the caret moves > outside the table; but no caret-moved events are seen when the caret moves > inside the table. NOREPRO: I'm seeing caret-moved events whenever I arrow up/down/left/right, whether inside or outside the table. Perhaps this is fixed in 4.3/4.4 now? Jean-P says he can repro in 4.3.5, but I'm not seeing it. I'll Resolve as WFM. Set back to UNCONFIRMED if you can still see the problem with 4.4.0.1 or later.
I can confirm that this is working with the nightly version of LibreOffice Writer. Orca is presenting the table cell contents now. Thanks!
Hi, While the things worked fine /ith 4.2.8.2, the bug appears in 4.3.0.1. Something happened in Writer which created a regression. What kind of output could help? Regards,
(In reply to Jean-Philippe MENGUAL from comment #6) > Hi, > > While the things worked fine /ith 4.2.8.2, the bug appears in 4.3.0.1. > Something happened in Writer which created a regression. > > What kind of output could help? Please test with 4.3.6 -- if it's still there, we could theoretically investigate backporting a fix to 4.3
Hi, I've just tested, the bug is still here on latest master, 5.0. Sorry but it's very boring for users to handle tables /ith this problem. Best regards,
I can confirm that there are still problems in this area but not always. If I open the test case file and press down-arrow i do get carret-moved events but if I start by pressing right arrow 8 times and then down arrow then the problem is back. So steps to reproduce: 1. Open the attached document 2. Launch the attached accessible-event listener in a terminal 3. Press right arrow 8 times 4. Press down arrow 2 times
I couldn't reproduce text-caret-moved not being emitted while moving in the table. I did see that moving downwards sometimes emitted duplicate events, which I guess is unexpected That is to say, moving upwards within a table cell made the attached script output e.g. object:text-caret-moved(-1, 0, 0) source: [paragraph | ] host_application: [application | soffice] object:text-caret-moved(7, 0, 0) source: [paragraph | ] host_application: [application | soffice] While moving downward within a table cell sometimes produced output like this: object:text-caret-moved(-1, 0, 0) source: [paragraph | ] host_application: [application | soffice] object:text-caret-moved(7, 0, 0) source: [paragraph | ] host_application: [application | soffice] object:text-caret-moved(7, 0, 0) source: [paragraph | ] host_application: [application | soffice]
In LibreOffice 5.1.1 I'm able to reproduce. Steps : 1) Open the document 2) Launch the script 3) Move down (event should appear) 4) Move down to the second line : no event appears We are some people able to reproduce the bug so It seems it was confirmed.
** Please read this message in its entirety before responding ** To make sure we're focusing on the bugs that affect our users today, LibreOffice QA is asking bug reporters and confirmers to retest open, confirmed bugs which have not been touched for over a year. There have been thousands of bug fixes and commits since anyone checked on this bug report. During that time, it's possible that the bug has been fixed, or the details of the problem have changed. We'd really appreciate your help in getting confirmation that the bug is still present. If you have time, please do the following: Test to see if the bug is still present on a currently supported version of LibreOffice (5.2.7 or 5.3.3 https://www.libreoffice.org/download/ If the bug is present, please leave a comment that includes the version of LibreOffice and your operating system, and any changes you see in the bug behavior If the bug is NOT present, please set the bug's Status field to RESOLVED-WORKSFORME and leave a short comment that includes your version of LibreOffice and Operating System Please DO NOT Update the version field Reply via email (please reply directly on the bug tracker) Set the bug's Status field to RESOLVED - FIXED (this status has a particular meaning that is not appropriate in this case) If you want to do more to help you can test to see if your issue is a REGRESSION. To do so: 1. Download and install oldest version of LibreOffice (usually 3.3 unless your bug pertains to a feature added after 3.3) http://downloadarchive.documentfoundation.org/libreoffice/old/ 2. Test your bug 3. Leave a comment with your results. 4a. If the bug was present with 3.3 - set version to "inherited from OOo"; 4b. If the bug was not present in 3.3 - add "regression" to keyword Feel free to come ask questions or to say hello in our QA chat: http://webchat.freenode.net/?channels=libreoffice-qa Thank you for helping us make LibreOffice even better for everyone! Warm Regards, QA Team MassPing-UntouchedBug-20170522
Dear all, The issue is still present on LibreOfficeDev 5.5 from 2017-05-30. Best regards.
I think there's a race here, and some degree of the window size matters probably too. Under no load this works for me, but if things are slowed down, e.g. by the presence of a screen reader working then when the cursor enters the table the table toolbar appears at the bottom of the window and the viewport of the document changes which causes the a11y cells contexts to be destroyed.
Bisected on Ubuntu 14.04 (w/ Orca) with 43max to: commit 69a440abc1db5fd999eae36b86e5098c67a5ca13 Author: Matthew Francis <mjay.francis@gmail.com> Date: Thu May 28 17:35:24 2015 +0800 source-hash-76c549eb01dcb7b5bf28a271ce00e386f3d388ba commit 76c549eb01dcb7b5bf28a271ce00e386f3d388ba Author: Steve Yin <steve_y@apache.org> AuthorDate: Fri Nov 29 13:03:27 2013 +0000 Commit: Caolán McNamara <caolanm@redhat.com> CommitDate: Mon Dec 2 10:25:33 2013 +0000 Integrate branch of IAccessible2
*** Bug 105487 has been marked as a duplicate of this bug. ***
Dear Joanmarie Diggs, To make sure we're focusing on the bugs that affect our users today, LibreOffice QA is asking bug reporters and confirmers to retest open, confirmed bugs which have not been touched for over a year. There have been thousands of bug fixes and commits since anyone checked on this bug report. During that time, it's possible that the bug has been fixed, or the details of the problem have changed. We'd really appreciate your help in getting confirmation that the bug is still present. If you have time, please do the following: Test to see if the bug is still present with the latest version of LibreOffice from https://www.libreoffice.org/download/ If the bug is present, please leave a comment that includes the information from Help - About LibreOffice. If the bug is NOT present, please set the bug's Status field to RESOLVED-WORKSFORME and leave a comment that includes the information from Help - About LibreOffice. Please DO NOT Update the version field Reply via email (please reply directly on the bug tracker) Set the bug's Status field to RESOLVED - FIXED (this status has a particular meaning that is not appropriate in this case) If you want to do more to help you can test to see if your issue is a REGRESSION. To do so: 1. Download and install oldest version of LibreOffice (usually 3.3 unless your bug pertains to a feature added after 3.3) from http://downloadarchive.documentfoundation.org/libreoffice/old/ 2. Test your bug 3. Leave a comment with your results. 4a. If the bug was present with 3.3 - set version to 'inherited from OOo'; 4b. If the bug was not present in 3.3 - add 'regression' to keyword Feel free to come ask questions or to say hello in our QA chat: https://kiwiirc.com/nextclient/irc.freenode.net/#libreoffice-qa Thank you for helping us make LibreOffice even better for everyone! Warm Regards, QA Team MassPing-UntouchedBug
I've started looking at this in the context of bug 147084, with the document attached there, but the issue appears to be the same regardless of the document. (However, the NVDA on Windows issue handled in bug 147084 has a different root cause from the Orca on Linux one covered here.) From my observations with current LO master so far, the race only occurs when using Orca, not e.g. just SAL_ACCESSIBILITY_ENABLED=1, and the issue seems to be reliably reproducible e.g. with the steps described in comment 9 if Orca is enabled. What seems to happen is that a11y objects for paragraphs are created as response to queries from Orca via AT-SPI, and if that happens, no accessible event listeners are registered on (some of) the involved a11y objects, so no events get sent when the cursor/caret moves into the corresponding paragraphs. I have an idea for potential approaches to fix that, plan to look further into that as I find time.
Unassigning for now, but I might come back to that at some point in time unless anybody else finds time earlier.
I can still reproduce with LODev 26.8 alpha0. But Caolán in comment #14 seems to be right, this is a bit fiddly and seems to depend on other factors, and window size seems to be one of them. At first I couldn't reproduce, but given I've got a fairly fast machine I though it might have not entered the supposed race, but then shrinking the window and starting over did it. It looks like if the cell overflows the view when the toolbar appears it renders the overflowing cells silent -- or something. But that's still speculation, and Michael might have more details if he dug a bit deeper.
(In reply to cwendling from comment #20) > I can still reproduce with LODev 26.8 alpha0. > > But Caolán in comment #14 seems to be right, this is a bit fiddly and seems > to depend on other factors, and window size seems to be one of them. At > first I couldn't reproduce, but given I've got a fairly fast machine I > though it might have not entered the supposed race, but then shrinking the > window and starting over did it. It looks like if the cell overflows the > view when the toolbar appears it renders the overflowing cells silent -- or > something. > But that's still speculation, and Michael might have more details if he dug > a bit deeper. As far as I can remember, whether or not the problem appears depends on the exact interaction (e.g. how the table is entered) and it was (quite) reliable when using the cursor keys as described in comment 9, but starting Orca in step 2 instead of just the script. According to my earlier investigation, I think that Orca's way of querying information in response to events (e.g. trying to get something using the flows-from/flows-to a11y relations) breaks the way that LO creates a11y listeners for the objects. IIRC, I couldn't reproduce the issue when registering listeners more "aggressively" in the gtk3 a11y bridge (recursively attaching listeners to all children in AtkListener::AtkListener, IIRC) or somehow registering a listener on demand, see demo change at https://gerrit.libreoffice.org/c/core/+/184954 That might have unwanted side effects (e.g. on performance, but I didn't investigate too much.) The possibly better (at least mid-/long-term) approach I had in mind to investigate further is to avoid this problems (and various others) by reworking the a11y event handling more fundamentally to not require a11y listeners to be registered for each a11y object in the first place. (One of the steps in that direction was to deduplicate 20+ a11y implementations by moving common logic including the emission of a11y events of most of them to the OAccessible base class to be able to change the logic in one place at some point. But more work is still needed.)