GNOME Bugzilla – Bug 430278
The Where_Am_I script does not work in most applications.
Last modified: 2007-05-02 16:08:45 UTC
Please describe the problem: When I open an application and I want to invoke the "Where_Am_I" function nothing happens. This ocurs in almost all apps except Gedit, where it works correctly and Firefox, where it wokrs partially. Selected text is not spoken in all OpenOffice programs including Evolution. In Nautilus selected files/folders were not spoken. Pressing Orca+Numpad-Enter to hear which application has focus results in silencce. In Firefox the latter works, but the "Where_Am_I" function does not speak, but this is known. Steps to reproduce: 1. Open an OO application, Nautilus or Evolution. 2. Select text you've typed or select a file/folder. 3. Press Numpad-Enter to hear your selection; you don't hear anything. 4. Press Orca+Numpad-Enter to hear which app has focus, you don't hear anything. Actual results: As I mentioned above, nothing happens. You should hear your selection or the name of the focused app. Expected results: I expect to hear my selection or the name of the app that has focus. Does this happen every time? It happens all the time. Other information: I use a 1 Ghz machine with 240 MB RAM. I use Orca 2.19.1pre with braille and speech but no magnification. I use the latest Espeak synth - version 1.22.
*** Bug 430280 has been marked as a duplicate of this bug. ***
Thanks Hermann! Lynn, our "Where am I" guy was out last week. He should be back today, and I'm assigning this bug to him.
Created attachment 86575 [details] [review] Patch to hopefully fix all the "where am I" problems. Several changes. No longer need to pass a speech context into the whereAmI() routine to get the application name. Application specific code is now in the app specific scripts. Moved the Evolution specific "where am I" code to the Evolution script. This also fixes bug #427481 – "where am I" does not speak Evolution Mail...Tasks toggle buttons. The problem here was that those objects were toggle buttons and the whereAmI() routine did not have a generic catchall clause at the bottom to handle such objects. I haven't committed the patch yet. Could other Orca team members please try it and provide feedback. Thanks.
This is much better than it was. I still need to do more work to find the places where the spec isn't being applied but checking in this patch will I think solve several user issues.
Thanks Mike. Okay, patch committed. Hermann, can you give it a try now too please and make sure it fixes all the problems you were finding with it? Move state to "pending".
Comments from Hermann sent to the Orca mailing list: "I've updated Orca and tested the patch: In general it works with two exceptions: 1) When i select files/folders in Nautilus the selection is not spoken. Orca speaks the last item I'm on. I have to check the status bar in order to see what I've selected. 2) In OO-writer and in Evolution Orca speaks only selected text within paragraphs. When I select a block that is over several paragraphs, it only speaks the first. Gedit is still OK."
> When i select files/folders in Nautilus the selection is not spoken. > Orca speaks the last item I'm on. I have to check the status bar in > order to see what I've selected. My understanding is that the "where am I" functionality works only on the object that currently has focus. It doesn't handle multiple objects. Therefore it will only speak about the last icon in this case. Mike/Will, am I misunderstanding the way that "where am I" should work? There is nothing in the Orca specification in section 3.15 "Where Am I command", about objects of role ICON. Currently they are being caught by the new catchall generic object routine in the where-am-I code that I've just added, which just speaks their label and role. Mike, what additions need to be spec'ed here for objects of role ICON? > 2) In OO-writer and in Evolution Orca speaks only selected text within > paragraphs. When I select a block that is over several paragraphs, it > only speaks the first. Again, my understanding is that the "where am I" functionality only works on the object that currently has focus. It doesn't handle multiple objects. Therefore it will only speak about the last paragraph in this case. When I tried this, the paragraph which had focus (i.e. where the text caret was) was described and it also correctly spoke the selection state. Mike, does the where-am-I functionality need to be changed to handle these two situations described above?
To be Honest, I didn't know you could select multiple icons. I'm not really sure what we should be doing here. I'm open to any ideas. I always use the list view for file management so am not at all sure how this is supposed to work.
To clarify the purpose of the "Where Am I" function two questions: 1) Why does the text selection work in Gedi? No matter how many paragraphs I select, they all were spoken. 2) If the purpose of the "Where Am I" function is just for orientationn and not to review selected text/files/folders, what about creating a feature that does exactly this: Speak/braille selected/highlighted text or objects? I think this is rather urgent. How can someone check out what he/she has selected if the selection is more than one paragraph or object? This is important for people who have to use the OO-writer, for example.
> 1) Why does the text selection work in Gedi? No matter how many paragraphs > I select, they all were spoken. Because all that text in gedit is a single object. Those "paragraph" are really just two newlines right together. > 2) If the purpose of the "Where Am I" function is just for orientationn and > not to review selected text/files/folders, what about creating a feature > that does exactly this: Speak/braille selected/highlighted text or objects? I'll let Mike answer that. When it comes to Orca UI design, I am but a humble minion. :-)
(In reply to comment #9) > To clarify the purpose of the "Where Am I" function two questions: > 1) Why does the text selection work in Gedi? No matter how many paragraphs > I select, they all were spoken. This is actually a bug and should be fixed. > 2) If the purpose of the "Where Am I" function is just for orientationn and > not to review selected text/files/folders, what about creating a feature > that does exactly this: Speak/braille selected/highlighted text or objects? > I think this is rather urgent. How can someone check out what he/she has > selected if the selection is more than one paragraph or object? This is > important for people who have to use the OO-writer, for example. I'll be updating the spec to reflect this and make the intended behavior more clear. > (In reply to comment #9) > To clarify the purpose of the "Where Am I" function two questions: > 1) Why does the text selection work in Gedi? No matter how many paragraphs > I select, they all were spoken. > 2) If the purpose of the "Where Am I" function is just for orientationn and > not to review selected text/files/folders, what about creating a feature > that does exactly this: Speak/braille selected/highlighted text or objects? > I think this is rather urgent. How can someone check out what he/she has > selected if the selection is more than one paragraph or object? This is > important for people who have to use the OO-writer, for example. >
There were two remaining issues for where-am-I related to handling multiple object selections. Two new specific bugs have been opened to handle these to allow us to close this general where-am-I bug (which I'm doing). These are: Bug #435223 - Where-am-I doesn't correctly identify multiple selected objects in Nautilus. Bug #435226 - Where-am-I doesn't correctly handle multiple selected paragraphs in OOo Writer and Evolution.