GNOME Bugzilla – Bug 697969
Deprecate atspi_text_get_text_{before,at,after}_offset()
Last modified: 2015-01-21 09:56:36 UTC
As described in this message [1], atk_text_get_text_{before,after}_offset should be removed. It's redundant to atk_text_get_text_at_offset() and confuses those implementing AtkText support. Orca is the primary, and possibly only, consumer of these additional redundant methods. Or, rather, it was. As of today, Orca no longer calls the pyatspi2 methods getTextBeforeOffset() and getTextAfterOffset(). Let's deprecate these methods for the 3.10 cycle and remove them in 3.12. [1] https://mail.gnome.org/archives/gnome-accessibility-devel/2013-April/msg00000.html
We have finally decider to go ahead with other approach (see [1])and deprecating the three functions in favour of a new one, not break the ABI, so renaming the bug accordingly... [1] https://mail.gnome.org/archives/gnome-accessibility-devel/2013-August/msg00001.html
Created attachment 250990 [details] [review] 1. Deprecate AtspiTextBoundaryType
Created attachment 250991 [details] [review] 2. Deprecate atspi_text_get_text_*_offset()
Created attachment 251478 [details] [review] Patch proposal New (unique) patch to fix this issue after agreeing on the get_string_at_offset new name
Created attachment 251479 [details] [review] Patch proposal Actually, this is the one
[Mass-resetting default assignee, see bug 705890. Please reclaim this bug report by setting the assignee to yourself if you still plan to work on this. Thanks!]
Created attachment 253842 [details] [review] Patch proposal (In reply to comment #5) > Created an attachment (id=251479) [details] [review] > Patch proposal > > Actually, this is the one Mike have already committed a similar patch to the one proposed one as part as the work done in bug 705581, but forgot to deprecate the "after" and "before" flavours of these methods, so I'm now attaching a new patch here just to take care of that extra last step.
Review of attachment 253842 [details] [review]: After all this months the patch doesn't apply cleanly anymore. Could you update it? In the same way, not sure if it makes sense to move the implementation of atspi_get_text_at_string around. Is not really related with the reason behind the commit (so not really atomic) and I don't think that is a rule-not-written the need to keep the same order of the header file on the source file. Mike, do you have an opinion about that?
Created attachment 294907 [details] [review] Patch Proposal For some reason, this bug fell off my radar a long time ago and, for another reason, I happened to find it last week, so here is a patch updated. Note: I also think it makes no sense to reorder the implementation file
Review of attachment 294907 [details] [review]: Looks good.
This problem has been fixed in the development version. The fix will be available in the next major software release. Thank you for your bug report.