Number of items no longer announced in tree views by NVDA

[Neurrone]

Issue Type: Bug

In explorer view, navigate through the tree view.

Info about the number of items e.g, "node_modules 2 of 5" used to be read but somewhere in the past few months this is now gone.

This isn't that big of a deal but if there aren't technical reasons why this is not set I'd rather have it because this is actually useful in some cases. For example, working through a find and replace job across multiple files.

I have explorer.compactFolders set to false but that shouldn't affect this.

VS Code version: Code - Insiders 1.46.0-insider (d487078, 2020-05-07T16:19:54.327Z)
OS version: Windows_NT x64 10.0.18363

System Info

Item	Value
CPUs	Intel(R) Core(TM) i7-8550U CPU @ 1.80GHz (8 x 1992)
GPU Status	2d_canvas: enabled flash_3d: enabled flash_stage3d: enabled flash_stage3d_baseline: enabled gpu_compositing: enabled multiple_raster_threads: enabled_on oop_rasterization: disabled_off protected_video_decode: enabled rasterization: enabled skia_renderer: disabled_off_ok video_decode: enabled viz_display_compositor: enabled_on viz_hit_test_surface_layer: disabled_off_ok webgl: enabled webgl2: enabled
Load (avg)	undefined
Memory (System)	15.86GB (2.77GB free)
Process Argv
Screen Reader	yes
VM	40%

Extensions (12)

Extension	Author (truncated)	Version
notes	can	1.0.0
vscode-markdownlint	Dav	0.35.1
vscode-eslint	dba	2.1.5
rest-client	hum	0.23.2
python	ms-	2020.4.76186
remote-wsl	ms-	0.44.2
cpptools	ms-	0.27.1
Go	ms-	0.14.1
vsnotes	pat	0.7.1
rust	rus	0.7.5
markdown-preview-enhanced	shd	0.5.3
vscode-webhint	web	1.5.5

[isidorn]

@Neurrone thanks a lot for filling this. I can reproduce this with NVDA and Orca. We properly set the aria attributes aria-setsize and aria-posinset. So I am not sure why screen readers are not reading this content
@joanmarie can you maybe investigate on the Orca side?

It might be that our "weird" tree selection confuses screen readers, since we set aria-selected = false on focused elements.
fyi @LeonarddeR @feerrenrut

[LeonarddeR]

I think it would help to know where the regression started to occur. It could also be a bug in Chromium.

[Neurrone]

When focus was over one of the tree items, I did nvda+f1 to bring up debug info and this is what NVDA reports for the focused object:

roles:treeitem; level:4; setsize:0; posinset:0

This is probably the culprit -  a bug in how the values are computed.

[joanmarie]

Presumably this Chromium bug which is already fixed. https://bugs.chromium.org/p/chromium/issues/detail?id=1066632

In case someone reports something similar regarding the selected item count, that would presumably be this other Chromium bug which is already fixed.
https://bugs.chromium.org/p/chromium/issues/detail?id=1058961

[isidorn]

@joanmarie awesome!

My feeling is that this is not super important so we will not back port. Though please let me know if someone is passionate about this and we can see with @deepak1556 about potentially backporting the fix.

[deepak1556]

@joanmarie thanks for the links to the CL and also thanks for fixing them.

Have backported them in electron

electron/electron#24061
electron/electron#24060

[isidorn]

@deepak1556 thanks a lot for backporting. I see the backport has a 8 target, so we should have this once we update to electron 8?

[deepak1556]

yup I plan to push Electron 8 forward this iteration :) Should be in insiders this week.

[isidorn]

@deepak1556 awesome!

[isidorn]

As @deepak1556 informed me we plan to go to electron 9 in the next milestone, thus pushing this out to August.

[isidorn]

Seems like we will shipt with Electron 9, thus closing this optimistically.

[Neurrone]

This is still happening on latest insiders build (electron 9) on my other machine that does not have this yet.

Can this be reopened?

[Neurrone]

Oh wait sorry, my bad. I got confused between this issue and a separate one, which is not this one. Please ignore the previous comment.