mozilla :: #accessibility

16 Feb 2017
14:50davidbwe don't own the current nightly top crasher!
17:34stephendheya yzen - I've searched; there doesn't appear to be a docker image for axe-core, right?
17:37yzenstephend i m not sure there is, i only heard of axe-cli on npm
17:38stephendI'm interested in looking into it: my specific use case, to start, would be to hook it up to Jenkins, and prompt for a base_url (target site) it could run against, self-contained
20:17yzentbsaunde so regarding the compatibility stuff, this part i believe will not work in content:
20:17yzenand i think there's one more place related to plugins
20:17yzenunless that only happens in parent
20:21tbsaundeyzen: yeah, I guess so, that's so much sadness there
20:22yzentbsaunde so my thinking was to pass/set consumers on a11yActivate between content parent and child
20:24tbsaundeyzen: why bother? seems to me it probably wouldn't be that terrible to just drop the jaws requirement for that weird hack
20:24tbsaundethe other obvious thing I see is FireWinEvent() but that should only happen in the main process
20:25tbsaundethere's the IA2Enabled thing, but I think that only matters for really old jaws
20:25tbsaundeif at all
20:26yzentbsaunde and this one ?
20:30tbsaundeyzen: yeah I guess that happens in the child process
20:30tbsaundebut given e10s changes around plugins a bit I wonder if its still needed
20:31tbsaundeanyway I wouldn't hurry to do a bunch of work just to make that work
20:32yzentbsaunde ok ill take care of the service provider case
20:57davidbi spoke to soon :)
20:57davidbwe have the top two crashers again
20:58davidbbug 1197181 wins the day
20:58firebot REOPENED, crash in mozilla::a11y::DocAccessibleParent::RecvHideEvent(unsigned __int64 const&)
21:05firebotNew Core - Disability Access APIs bug 1340284 filed by
21:05firebot ASSIGNED, Audit instances where Compatibility is used in content process.
21:06tbsaundedavidb: and no str bleh, lets see what happens if we downgrade that assert I guess :/
21:09davidbtbsaunde: sgtm
21:14tbsaundeok then I'll push the obvious patch and we'll see if it just causes a problem somewhere else
21:15yzentbsaunde also there's one in rootaccessiblewrap RootAccessibleWrap::DocumentActivated where if dolphin, we hide non active window..
21:17tbsaundeyzen: that spot isn't going to work for e10s
21:17tbsaundeI sort of thought you handled dealing with this already somewhere else
21:18yzentbsaunde was not needed really, as focus issue (which is where that came up) was caused by wrong hwnd when firing events
21:19tbsaundeyzen: ok, well at any rate would need to be someplace else
21:19tbsaundeby which I actually mean I think this needs to happen in the main process
21:20yzentbsaunde ok ill open a bug for htat
21:23tbsaundedo we have data on how many dolphin users there are? fixing stuff specific to it seems rather low priority
21:29yzendavidb ^ ?
21:36davidbyes low priority relatively
21:36* davidb looks
21:39davidbWe have about 1 Dolphin user for every 12 Jaws users
21:40tbsaundewow that many?
21:40davidbyeah surprised me too
21:45yzenhm ok
21:47davidbactually more like 1 in 9
17 Feb 2017
No messages
Last message: 5 days and 2 hours ago