From MozillaWiki
Jump to: navigation, search


Scribed by Karl

Relman - Webcompat Working better together (karl)

Dennis and I had a meeting with relman last Thursday. Relman discovered we had a Webcompat Priority Flag. They thought we were not sending any signal at all. So they want us to have more visibility in the organization and have less unplanned regressions when releasing. Let's discuss.

  • James: Release management wants to avoid regression. There is the severity issue which is slightly different from the webcompat priority flag. If they start looking at WebcP1 flag and say they can't release, we will not be able to set WebcP1. We can't use the WebcP1 as a strong signal for relman.
  • James: if we know this changing will break some sites, we should set the tracking flag and should be set as S1 (Severity). I'm a bit concerned about using WebcP1.
  • Dennis: We should do both. We should get better at tracking flag for regression. relman should probably not be overly concerned to block release over webcp1.
  • Karl: Probably we should share the same process in between us. We don't currently.
  • Ksenia: When we take too much time to diagnose, the regression has already "passed". So the tracking is useless if we are not diagnosis in advance.
  • Dennis: Let's do both.
  • Thomas: I usually don't set the flags myself, and I try to ping the person. Setting the tracking could be done, I usually try to raise the awareness.
  • Dennis: Relman looks at the tracking flag on a daily basis. If we see a regression, we need to make sure that relman is seeing it, hence tracking.
  • James: if we know that something will break this site, and we say that it breaks a big site, we are in a better place to inform on severity, because the webcompat team has the context.
  • Ksenia: Do we set tracking flags? for regression? for top sites? for everything?
  • Dennis: Probably we should limit the huge site.
  • James: Relman is asking: "should Firefox ship?". If we know it will break this top site but will not block this top site.
  • Thomas: Sometimes we are being asked for diagnosis from the engineers, and nobody sets the flag.
  • Dennis: It seems that relman thinks that Webcompat knows about regression, but they don't.
  • Karl: Sometimes we identify a regression for a small site, but we don't always know if we break big sites.
  • Thomas: Severity is probably a good way to inform people about the priority.
  • James: We have the risk of setting a flag on all bugs and we will not be helpful.
  • Dennis: We would be helpful if we are tracking the regression. Currently we don't track everything, the same way.
  • James: People want our expertise. Flagging too much stuff, then we don't help. We want to add information. We do not want quantity, but quality. We will probably adjust the process with time.
  • James: We have the triage meeting. We could use also this meeting to set the tracking.
  • Ksenia: There are certain bugs where the engineer release to test the water, to see if it breaks something. What should we do?
  • Dennis: relman wants to have these features shipped to beta. If same-site lax breaks bing, we need to make sure that there is a bug.
  • James: Do we have enough information to know if it's a regression in advance.
  • Ksenia: Raul and Oana are adding regression since recently. Diagnose fast, Track quick!

The Brag Document(tm) (dennis)

Let's briefly talk about the brag document idea from last year. (Next week)

Two Minutes ( 👹 )

The useful, the silly, the things you did since the last meeting. Pro and Perso accepted.

  • Dennis: Mostly diagnosis.
  • James: Interop 2022 blog post &c. WebDriver touch and wheel events.
  • Karl: Firefox/Chrome 100 blog post. Webcompat Diagnosis. UA freezing meeting. Management thingies. Meeting with relman. Mingei exhibition (folk artcraft as arts, soetsu yanagi).
  • Ksenia:
  • Oana: Triage and moderation. Finished Bugzilla issues triage and continue checking top Alexa sites with Firefox.v100. I did a bit of research on Automation with Python. Had an awesome hike weekend with warm and sunny weather.
  • Raul: Triage and moderation. Continue checking top Alexa sites with Firefox.v100. Snowboarding -did my first ride through the forest trails and first run doing kick jumps (short in jump height->30-40cm - but seemed extremely high to me)
  • Thomas: Finally almost done with the automated test framework; focusing on tests now. Prepping some slides for the next all-hands, and dealing with snow and cold (finally a proper Canadian winter!)
  • Joe:
  • Paul: Paternal leave.