957306 - atk_text_get_text() should return an empty string when the start and end offsets are the same

Status: RESOLVED INVALID

(Core :: Disability Access APIs, defect)

Description

[Joanmarie Diggs]

Attachment: test case: text.html

Steps to reproduce:
1. Launch the attached accessible-event listener in a terminal
2. (Re)load the attached test case

Expected result: The text returned for character range (0, 0) would be an empty string.

Actual result: The text returned for character range (0, 0) is the element's text.

Comment 1

[Joanmarie Diggs]

Attachment: accessible-event listener: document.py

Comment 2

[alexander :surkov (:asurkov)]

can't confirm on xpcom level, I assume Nightly is tested?

Comment 3

[Joanmarie Diggs]

(In reply to alexander :surkov from comment #2)
> can't confirm on xpcom level, I assume Nightly is tested?

29.0a1 (2014-01-07)

Comment 4

[alexander :surkov (:asurkov)]

Trev, would you mind to look? I don't really see how it may be so different from XPCOM part

Comment 5

[Trevor Saunders (:tbsaunde)]

(In reply to alexander :surkov from comment #4)
> Trev, would you mind to look? I don't really see how it may be so different
> from XPCOM part

I tried, but only got one call to getTextCB with start 0 and end -1 and then before I could tell gdb to continue the document.py script  printed some gerror timeout thing and no more calls to stuff in gecko happened.

Comment 6

[alexander :surkov (:asurkov)]

the script doesn't use -1? Did you try to modify the script to reveal the issue?

Comment 7

[Trevor Saunders (:tbsaunde)]

Ok, so I figured out how to debug this, but when atk-bridge calls into gecko for the text.getText(0, 0) call gecko is passed (text, 0, -1) for some reason, so I claim this isn't our bug (I don't have symbols for libatk-bridge setup so I'm not sure why that happens)

Comment 8

[alexander :surkov (:asurkov)]

Thanks, Trev! Cc'ing Alejandro.

Comment 9

[Joanmarie Diggs]

Found it. Apologies. https://bugzilla.gnome.org/show_bug.cgi?id=721860