From MozillaWiki
Jump to: navigation, search


Scribed by Ksenia

How to change the web-bug process for the better? (james)

Now with even less resources, is there anything we can do to our processes to make them more efficient? Is there any data we can collect to inform our work and prioritize issues?

  • james: we should probably think what we can do now that there is even less resources.
  • dennis: diagnosis is going to be a problem that Karl is gone. Tranco list is being used to set priority label. Severity label is used for how much breakage is seen.
  • james: 80% of issue are severity critical. 20% are priority critical or important. Looking at only top sites would be a trivial way of looking at diagnosis, but we can miss some things. Wonder what the outcome of outreach.
  • dennis: Maybe about 10% respond
  • tom: Maybe if the sites don't respond, we could just add interventions
  • dennis: yes, we have action-needssitepatch label, so we could do that, but we also need resources
  • james: is there a way to know there is an intervention in the console panel?
  • dennis: yes, but not for css interventions or all UA overrides. See [this patch](, which was blocked by [this bug](, and we (Karl and I) deproritized that a month ago.
  • james: I'm not sure to what extent google is willing to help with outreach
  • joe: I have a question about an outcome of diagnosis. One part is QA function and another part is whether we're contributing any new information to the core team
  • james: there are a few types of breakage - 1) a site is detecting UA and not supporting Firefox 2) breakage is based on a core issue (it depends on a feature that is not implement or a behavioural differnce)
  • joe: is there a way to detect a core issue early and stop diagnosis?
  • ksenia: It's hard to know what's causing an issue. Recently, I tried building a minified testcase, and couldn't find the underlying cause. I wasn't able to figure out, and neither were Emilio and Daniel Holbert without figuring out what's causing is.
  • joe: how many flexbox issues are there?
  • dennis: dholbert has been working on fixing a few issues, maybe around 70 on bugzilla
  • james: maybe we can stop early and ping layout team and let them know that it's a flexbox issue, for example
  • joe: it appears that most core engineers don't pay attention to webcompat priotiry flag on bugzilla, wonder if it is useful
  • james: maybe we weren't too consistent with using it
  • joe: this is ok for your team to change severiry of a bug
  • james: perhaps when we're setting webcompat priority, we can raise the severity too
  • joe: different teams have different definition of what's important and it's hard to change
  • dennis: I'm happy to try and see if any of the issues are related to flexbox and close them early
  • joe: we have to tell ourselves that it's ok to not diagnoses everything
  • paul: I'm curious about the QA function of the team
  • joe: For example, if we detect an issue on a bank website, we can signal about it early enough and take action

Firefox 100 Recap (so far) (dennis)

Brief summary of how Version 100 is going so far.

  • dennis: We had a few busy weeks between Chrome's and Firefox release, when we shipped a lot of overrides, but there aren't a lot of reports of new breakage at the moment.