[jbosstools-issues] [JBoss JIRA] (ERT-708) Orca's flat review mode can not be used [EBZ#543949]

[Eric Williams (Jira)]

     [ https://issues.jboss.org/browse/ERT-708?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]

Eric Williams reassigned ERT-708:
---------------------------------

      Sprint: devex #162 February 2019
    Assignee: Eric Williams


> Orca's flat review mode can not be used [EBZ#543949]
> ----------------------------------------------------
>
>                 Key: ERT-708
>                 URL: https://issues.jboss.org/browse/ERT-708
>             Project: Eclipse Release Train
>          Issue Type: Task
>          Components: Platform
>            Reporter: Friendly Jira Robot
>            Assignee: Eric Williams
>            Priority: Major
>              Labels: SWT, bzira
>
> Recently I reported on the orca list a problem where every time I
> tried to use flat review mode, my machine was completely frozen.
> The problem has been fixed but now I can not use the flat review mode. According to Joanie, the person who corrected the problem, the cause is due to a bug in the eclipse. 
> I am copying the text of the email I received because I believe it
> will give more details.
> ----------
> Hey again.
> I have good news and bad news. The good news is that I have fixed the
> desktop lockup, at least for me. The bad news is that until the Eclipse
> developers fix their bug, flat review will not work in impacted windows.
> Eclipse is giving an insane bounding box (x, y, width, height) for a
> panel which seems to be the child of the scroll pane in that dialog. I
> get, for example, (618552120, 32709, 618552176, 32709). I should be
> getting things like (598, 353, 720, 328). In theory one should be able
> to test this in Accerciser, but Eclipse seems to still have that bug
> where it takes 15 or so seconds (followed by a DBus timeout to become
> responsive). And that seems to make Accerciser very unhappy. At least on
> my machine.
> Because flat review is constructing a 2-dimensional collection of
> objects based on their bounding boxes, we need accurate (not to mention
> sane) bounding boxes.
> So the way the Eclipse developers can test this is to grab the latest
> Orca master and perform the steps to reproduce. This time the expected
> result is that Orca should present the window via flat review and the
> actual result is that it just presents the name of the window ("Eclipse
> Platform Launcher"). To see why, look in your friendly debug.out. Here's
> what I get for reading the current line:
> vvvvv CONSUME ATSPI_KEY_PRESSED_EVENT: 'KP_Up' (80) vvvvv
> [normal, non-concerning stuff goes here]
> 00:47:16 - INFO: Extents for [panel | ] are: (618552120, 32709,
> 618552176, 32709)
> 00:47:16 - INFO: [panel | ] seems to have bogus coordinates
> [normal, non-concerning stuff goes here]
> ^^^^^ CONSUME ATSPI_KEY_PRESSED_EVENT: 'KP_Up' (80) ^^^^^
> You can quickly find these strings by searching for "bogus coordinates"
> and then looking at the previous line.
> ----------
> The steps to reproduce this problem are:
> 1. Activate orca but you need to grab the latest from git.
> 2. Launch eclipse.
> 3. When eclipse ask for the workspace try to explore the screen using the flat review mode.
> Expected result:
> Orca should read the content of the screen according to the commands used.
> Actual result:
> Orca reads only Eclipse IDE Launcher.
> Important: If you use an old version of orca, probably yourmachine will freeze.



--
This message was sent by Atlassian Jira
(v7.12.1#712002)