1422674 - IAccessible::accChild on parent process accessibles fails with positive indexes

Status: RESOLVED FIXED

(Core :: Disability Access APIs, defect)

Description

[James Teh [:Jamie]]

Spun off bug 1422201 comment 13.

STR:
1. Start NVDA and Firefox.
2. Open about:support.
3. Ensure the document is focused.
4. Open the NVDA Python console with NVDA+control+z.
5. Enter: focus.IAccessibleObject.accChild(1)
Expected: An IAccessible.
Actual: Exception.
6. Open a remote document; e.g. mozilla.org.
7. Ensure the document is focused.
8. Open the NVDA Python console with NVDA+control+z.
9. Enter: focus.IAccessibleObject.accChild(1)
10. Observe that you get an IAccessible as expected.

I think that this is because AccessibleWrap::GetRemoteIAccessibleFor calls GetRemoteIAccessibleFor if this isn't a chrome accessible:

if (XRE_IsParentProcess() && !IsProxy() && !sIDGen.IsChromeID(varChild.lVal)) {

but it doesn't first check that this is a negative id. I don't think !sIDGen.IsChromeID knows about positive ids, so it's going to answer that it isn't a chrome id. Thus, GetRemoteIAccessibleFor gets called, which will of course fail. This is all educated guesswork right now, though; I haven't done any actual debugging.

Low priority because I don't know of any clients that actually need this.

Comment 1

[James Teh [:Jamie]]

Regression introduced by bug 1304449 (Firefox 52). Trivial fix, so I'm just going to do this now while it's in my head.

Comment 2

[James Teh [:Jamie]]

Attachment: Bug 1422674: Fix IAccessible::accChild on parent process accessibles with positive indices.

Previously, in the parent process, we were treating positive child ids as remote unique ids.
This of course failed when searching remote documents and returned early.
Make sure we only treat ids as remote if they are less than 0.
Ids above 0 are child indices and are handled later in the code for both local and remote children.

Review commit: https://reviewboard.mozilla.org/r/205632/diff/#index_header
See other reviews: https://reviewboard.mozilla.org/r/205632/

Comment 3

[Marco Zehe (:MarcoZ)]

Comment on attachment 8934752 [details]
Bug 1422674: Fix IAccessible::accChild on parent process accessibles with positive indices.

https://reviewboard.mozilla.org/r/205632/#review211290

Comment 4

[Pulsebot]

Pushed by [e-mail]:
https://hg.mozilla.org/integration/autoland/rev/ce401a7898ee
Fix IAccessible::accChild on parent process accessibles with positive indices. r=MarcoZ

Comment 5

[Raul Gurzau (:RaulG)]

https://hg.mozilla.org/mozilla-central/rev/ce401a7898ee

Comment 6

[Ryan VanderMeulen [:RyanVM]]

Given where we are in the cycle and that this isn't a new regression, I'm calling this wontfix for 58. Feel free to set the status to affected and nominate it for Beta approval if you feel strongly otherwise.