Proposal: add Blink-related triaging of the UI component

[Chris Harrelson]

+Robert Liao, TL of several UI components for the browser.

There is some evidence that some bugs filed under component:UI have Blink relevance.

I suggest we add a task to this rotation to review this query and add Blink components and UI subcomponents as appropriate, with a goal of moving the untriaged count in component:UI to zero. As part of that we could recruit help from non-Blink engineers to do some of the UI triage into subcomponents. Currently there are 1315 bugs in this list, so it'd take a while to get it to zero (which is fine).

Alternatively, we could "just" find the Blink bugs among the untriaged UI bugs. Downside is that we'd be doing repetitive work as we wouldn't know that a previous Blink triager had already looked. 

Chris

[Johnny Stenback]

I think we should help out here, yes. And while us Blink engineers aren't necessarily great at figuring out the appropriate UI subcomponents we could add a label, something like UnlikelyWebPlatform, to indicate that a Blink triager looked at one of those bugs and determined that it's unlikely to be a Blink (or other Web Platform) bug. Then we could update the query that we use in the Blink triage rotation to exclude those bugs.

Thoughts?

Cheers,

Johnny

--
You received this message because you are subscribed to the Google Groups "Component:Blink bug rotation" group.
To unsubscribe from this group and stop receiving emails from it, send an email to crblink-rotati...@chromium.org.
To view this discussion on the web visit https://groups.google.com/a/chromium.org/d/msgid/crblink-rotation/CA%2BN6QZs%3D%3D9uVTu_5VcJr8aoFXTnse0_o4B4rBH4u%2BusUjciOpA%40mail.gmail.com.

[Chris Harrelson]

On Thu, May 6, 2021 at 11:16 AM Johnny Stenback <jste...@chromium.org> wrote:

I think we should help out here, yes. And while us Blink engineers aren't necessarily great at figuring out the appropriate UI subcomponents we could add a label, something like UnlikelyWebPlatform, to indicate that a Blink triager looked at one of those bugs and determined that it's unlikely to be a Blink (or other Web Platform) bug. Then we could update the query that we use in the Blink triage rotation to exclude those bugs.

Thoughts?

Agreed, that could work also.

[Chris Harrelson]

Anyone else with an opinion? The options are:

1. Triage all component:UI bugs into sub-components and/or Blink components to the best of your ability

2. Triage all component:UI -label:UnlikelyWebPlatform to see if any Blink components should be added; add that label after triaging

3. Do nothing and leave 1300 untriaged bugs in component:UI for other teams to own, or get lost, or rely on hero engineering

In my opinion, option 1 looks better than 2 because my guess is that the amount of work is similar, and it achieves a better outcome. But I could be wrong - e.g. it could be that there a number of component:ui bugs that have no clear place to be put.

Without objection, I suggest that we pilot option 1 via a volunteer sub-group (I volunteer to be one of them) and then come back to the group with data.

[Dave Tapuska]

Did you intend the query operator to be ":" as opposed to "=". I believe the queries should be component=UI not component:UI as the latter includes sub-components of UI.

To view this discussion on the web visit https://groups.google.com/a/chromium.org/d/msgid/crblink-rotation/CA%2BN6QZsrEEMS6OEVu536xa4L_V79eKk3F%2BeZSXfe5JZEt0rUtw%40mail.gmail.com.

[Chris Harrelson]

On Mon, May 10, 2021 at 9:52 AM Dave Tapuska <dtap...@chromium.org> wrote:

Did you intend the query operator to be ":" as opposed to "=". I believe the queries should be component=UI not component:UI as the latter includes sub-components of UI.

Good correction. component=ui is what I intended.

[Stephen Chenney]

I lean toward (1) with (2) in cases of no decent subcomponent. That provides a signal as to where we might have a hole in ownership. I guess it's really (2) discouraging the use of the label.

Stephen.

To view this discussion on the web visit https://groups.google.com/a/chromium.org/d/msgid/crblink-rotation/CA%2BN6QZvNbb9p3AMqSBitKnh4cDOcS7Kq9EHoVwwPyVQhHAthkA%40mail.gmail.com.

[Chris Harrelson]

Not hearing any objection, I went ahead and added this as an optional task here. I'm on rotation today and tomorrow, will give it a shot and report back what I find.

[Chris Harrelson]

Update:

I have triaged about 20-25 of the issues. I found it was very easy to find other components in almost all cases - triaging those 25 issues took only half an hour.

In the process I found about 5 of them were actual Blink bugs. Examples:

https://bugs.chromium.org/p/chromium/issues/detail?id=1208277

https://bugs.chromium.org/p/chromium/issues/detail?id=1197833

I think this is clearly high-value.

Chris

p.s. There were about 5 bugs that were very poor-quality reports regarding something about Android OEMs that wasn't sure what to do with, so dumped them in a few places.

[Chris Harrelson]

On Wed, May 12, 2021 at 10:50 AM Chris Harrelson <chri...@chromium.org> wrote:

Update:

I have triaged about 20-25 of the issues. I found it was very easy to find other components in almost all cases - triaging those 25 issues took only half an hour.

In the process I found about 5 of them were actual Blink bugs. Examples:

https://bugs.chromium.org/p/chromium/issues/detail?id=1208277

https://bugs.chromium.org/p/chromium/issues/detail?id=1197833

I think this is clearly high-value.

I think it's so valuable that we should make this a required task. Any objection?

[Stephen Chenney]

On Wed, May 12, 2021 at 2:13 PM Chris Harrelson <chri...@chromium.org> wrote:

On Wed, May 12, 2021 at 10:50 AM Chris Harrelson <chri...@chromium.org> wrote:

Update:

I have triaged about 20-25 of the issues. I found it was very easy to find other components in almost all cases - triaging those 25 issues took only half an hour.

In the process I found about 5 of them were actual Blink bugs. Examples:

https://bugs.chromium.org/p/chromium/issues/detail?id=1208277

https://bugs.chromium.org/p/chromium/issues/detail?id=1197833

I think this is clearly high-value.

I think it's so valuable that we should make this a required task. Any objection?

I agree. Required to triage Compoonent=UI issues, with a target of processing 20 issues?

More, less?

Stephen.

[Chris Harrelson]

On Wed, May 12, 2021 at 11:48 AM Stephen Chenney <sche...@chromium.org> wrote:

On Wed, May 12, 2021 at 2:13 PM Chris Harrelson <chri...@chromium.org> wrote:

On Wed, May 12, 2021 at 10:50 AM Chris Harrelson <chri...@chromium.org> wrote:

Update:

I have triaged about 20-25 of the issues. I found it was very easy to find other components in almost all cases - triaging those 25 issues took only half an hour.

In the process I found about 5 of them were actual Blink bugs. Examples:

https://bugs.chromium.org/p/chromium/issues/detail?id=1208277

https://bugs.chromium.org/p/chromium/issues/detail?id=1197833

I think this is clearly high-value.

I think it's so valuable that we should make this a required task. Any objection?

I agree. Required to triage Compoonent=UI issues, with a target of processing 20 issues?

More, less?

I don't think we need a quota, we just need to keep the number of untriaged bugs decreasing. If we all triage even 30 of them over our two days on duty, it'll take about 90 days to get to zero (not counting new bugs coming in, but there really aren't an enormous number of them), which I think is just fine. We don't need to be in a rush to triage the backlog.

[Johnny Stenback]

Great to see this moving the needle, and doing so fairly easily! I'm supportive of making this a required step.

Cheers,

Johnny

[Chris Harrelson]

Update from today: I triaged about 20 more bugs, and it was once again pretty easy and I found a few that seemed actionable. Currently there are 1251 bugs, let's get it to zero!

I've now updated it this to be a required step:

Task 2: Handling Component=UI issues (mandatory, daily)

1) Search for Component=UI

2) Read the issue description and add comments or move to sub-components of UI or other components (including Blink sub-components as appropriate. Set priorities as needed

[TAMURA, Kent]

LGTM!

To view this discussion on the web visit https://groups.google.com/a/chromium.org/d/msgid/crblink-rotation/CA%2BN6QZtD8f7161bbn_MK5f4iBJoL7OERjhWiP4JQqv7R2z-eCQ%40mail.gmail.com.

--

TAMURA Kent
Software Engineer, Google

[Chris Harrelson]

Thanks!

I also noticed that we're down to 989 untriaged tests. Not bad to go down by close to 300 in only a week.

[Dave Tapuska]

Yes I removed the UI component where it was set on a subcomponent and that made sense...

Like component = UI, UI>Browser>Omnibox.  I removed the UI component.

dave.

To view this discussion on the web visit https://groups.google.com/a/chromium.org/d/msgid/crblink-rotation/CA%2BN6QZsMjyRcV2Xa67%2B9%3Dm1DFYcS73fZ0%2B%3Dw3_3nhVAmcOVP1g%40mail.gmail.com.

[Rick Byers]

Wow, that sounds way more valuable than I initially assumed it would be. Thanks for pushing on it Chris, sounds like a great change! I think my expectation of Component:UI being a dumpster-fire of low signal was coloured by my long ago feedback on ChromeOS UI team looking at product feedback reports along with bugs.

To view this discussion on the web visit https://groups.google.com/a/chromium.org/d/msgid/crblink-rotation/CAHgVhZW%2BT5WaMJdQ%3DVSyU4t3-2-oBgM_0S9EkCN4dKo5G4a0SQ%40mail.gmail.com.

[Chris Harrelson]

FYI we just today got down to zero open issues in the UI component.

Thanks everyone! Now we can focus on maintaining that via continued steady-state triaging of the UI component into subcomponents and other components.