mozilla :: #accessibility

17 Feb 2017
09:16firebotNew Core - Disability Access APIs bug 1340458 filed by
09:16firebot NEW, Crash in mozilla::a11y::DocAccessibleParent::RecvShowEvent
15:06tbsaundedavidb: what do we want to do about aurora?
15:06tbsaundeI guess given linux we should backport crash fixes erg
15:10davidbtbsaunde: aurora is 53 currently, I don't expect a lot of windows crashes because we won't ship on by default there
15:11davidbare we having a lot of linux instability?
15:11tbsaundebut see linux, so my conclussion was can't really hurt I guess
15:11tbsaundeI don't know
15:12tbsaundeand I don't run aurora or beta so no real idea
15:13davidbok, it is news to me we don't know.
15:13tbsaundewell, you've been looking at crash stats
15:13tbsaundemore than me
15:13davidblinux is buried
15:14davidbjoanie: we might need your help, do you know if many (any?) orca users use Firefox beta, aurora, or nightly?
15:14tbsaundeyeah, I think the question is what does it look like if you only look at linux crashes?
15:14tbsaundewhich I would thing is a thing crash stats can show you
15:15davidbtbsaunde: here is one day on nightly
15:15davidbtop crash is 8 crashes on RecvHideEvent
15:15davidbso concerns are valid it seems
15:16tbsaundeok, so back to "might as well"
15:16davidbtbsaunde: where do I look to confirm where linux e10s a11y is shipping?
15:18tbsaundethere's some stuff in toolkit/xre/ and I guess IPCAccessibilityActive though we have disabled there in a while iirc
15:18tbsaundeerr, I mean we haven't disabled there in a while, sorry
15:20davidblooking at bug 1312100
15:20firebot FIXED, Enable a11y+e10s by default for OSX/Linux in Fx52
15:24davidbtbsaunde: I'm not seeing a11y crashes anywhere near the top here in 52
15:25davidbbut then, isn't helpful
15:27tbsaundedavidb: yeah...
15:27tbsaundeI didn't see much for 53 when I searched for that either
15:54davidbtbsaunde: dxr shows about 7 instances of MOZ_DIAGNOSTIC_ASSERT(CheckDocTree())
15:54tbsaundedavidb: and?
15:55davidbtbsaunde: i really don't know if that should worry me
15:55davidbI think you know which are causing crashes, and I guess the others shouldn't
15:55tbsaundewell, I turned all or most of them into normal MOZ_ASSERTS a couple hours ago
15:56tbsaundeI don't think we'll learn much by leaving some enabled
15:56davidbtbsaunde: is keeping them as MOZ_ASSERTS helpful? I guess for your debugging?
15:57tbsaundeit can't hurt since they only do something in debug builds
15:57tbsaundedoesn't help me immediately, but maybe some day someone will find a test case
15:57davidbi guess it will only hurt our developers
15:58tbsaundeit helps fuzzzzzer people so that's something
15:58tbsaundenot that fixing the bugs they've already filed is going well
16:20firebotNew Core - Disability Access APIs bug 1340575 filed by
16:20firebot NEW, Crash in mozilla::a11y::HyperTextAccessibleWrap::QueryInterface
16:24firebotNew Core - Disability Access APIs bug 1340579 filed by
16:24firebot NEW, Crash in PLDHashTable::Search | mozilla::a11y::DocAccessibleParent::GetAccessible
22:22davidbtbsaunde: ryan was asking about uplifting bug 1334673. Looks like there is an NI still waiting on you.
22:22firebot FIXED, Crash in IPCError-browser | PDocAccessibleParent::RecvShowEvent invalid index
22:22davidbyou are Mr popular :)
23:22firebotNew Core - Disability Access APIs bug 1340733 filed by
23:22firebot NEW, get 'attributes' from xpcom accessible hangs the browser in windows.
23:58aklotzDoes anybody know where the M(a) test suite came from?
23:58aklotzie, which bug or developer?
18 Feb 2017
No messages
Last message: 216 days and 10 hours ago