Compatibility/Meetings/Sync w Mike

From MozillaWiki
< Compatibility‎ | Meetings
Revision as of 14:09, 22 April 2020 by OanaArbuzov (talk | contribs) (minor changes)
Jump to navigation Jump to search

April 22 2020


Monthly Intervention check (SV)

Issue https://bugzilla.mozilla.org/show_bug.cgi?id=1623375 Salesforce Community intervention tracking bug - which is a "meta" bug, contains a list of webcompat issues https://github.com/webcompat/web-bugs/issues?utf8=%E2%9C%93&q=doNotShowUnsupportedBrowserModal

Proposal:

1. Test all webcompat issues from the list and create a Bugzilla issue to remove the intervention if the issue is no longer reproducible (for each domain)? The meta bug remains open until all issues no longer reproduce without the intervention.

2. Test 1-2 webcompat issues and if one of them is still reproducible with interventions off, consider the meta bug still reproducible.

Mike: I think testing a few sites is sufficient.

Oana: And if we find those are fixed, we can test all of them to verify.

Interventions are added based on site popularity.

SV team can also propose sites to which interventions could be added - https://github.com/webcompat/web-bugs/labels/action-needssitepatch

UA Overides documentation: https://github.com/mozilla/webcompat-addon/blob/master/src/data/ua_overrides.js

Focus issues not reproducible on Firefox Preview Nightly (SV)

We noticed that you closed the issue that are reproducible in Focus but not in Fenix as WON'T FIX.

Suggestions:

1. We proceed on closing similar future issues as WON'T FIX.

2. (also) Forward the issues to the Focus repo.

Sergiu: how about we open a new Focus bug, and close the webcompat one:

https://github.com/mozilla-mobile/focus-android/issues/new

Mike: COOL

Tracking Protection Strict issues - desktop (SV)

Previously, François Marier told us he is not interested in TP-Strict issues since it is expected to break stuff. Also, the default ETP setting in desktop is Standard.

We've been reporting Bugzilla issues which are reproducible on desktop with ETP-Strict enabled. Should we continue doing that or should we just skip them?

Action item for Mike: Get clarity on this.

SV Team: Continue reporting them for now.

Moderation improvements?

https://github.com/webcompat/webcompat.com/issues/3291

Action item for SV Team - document typical work flow, improvements - leave a comment.



April 10 2020


Tracking Protection same domain (SV)

Meta bug: https://bugzilla.mozilla.org/show_bug.cgi?id=1609867 Embedded content (Twitter/Instagram/Facebook) is not displayed with ETP Strict enabled

- Should we report different issues related to missing embedded Twitter/Instagram/Facebook domains for every site we encounter or add them to the "See also" of the Meta bug

Action item: Let's ask Tom (we pinged him on the internal channel).

Tom reply: Add separate issues fro every domain

Interventions & Overrides recheck (SV)

Beta has been migrated to Fenix. We will run the montly Interventions and Overrides on Release, unless we can get a Fenix build without the interventions (?).

Action item: Test on Release.

SV team - run Interventions on Monday/Tuesday (with Release)

mike: We should ask Dennis what we can do here... how we can disable before about:compat is ready. Custom build?

sergiu: Another solution would be to take the last known Beta build from FTP server and test with that (before the Fenix migration).

Pull to Refresh issues on Fenix (SV)

We are receiving lately a bunch of issues tied to page refreshing while performing different actions on a page.

- Should we consider closing them (if that is the case) as duplicates of the meta-bug: https://github.com/mozilla-mobile/fenix/issues/9766 ?

- Maybe find a more relevant duplicate from the list: https://github.com/mozilla-mobile/fenix/issues?q=is%3Aissue+is%3Aopen+pull+to+refresh+ ?

Action item: Mark them as duplicate of https://github.com/mozilla-mobile/fenix/issues/9766 (according to Petru Lingurar comment)



March 25 2020


(FYI) Q2 OKR planning (SV)

We added our ideas/suggestions in the Q2 OKRs planning in a comment

https://docs.google.com/document/d/1NUVw7AQL8t8cAciKZjDeovtIsCFgaeboajbOfDnYU74/edit#heading=h.y0hwvz9mtx90

Covid-19 exploratory testing (SV)

The sites and the progress is visible here https://docs.google.com/spreadsheets/d/1vwe2GxrS-mbAG4m15PJEfV80jEsml5hvb4scGD3RxQQ/edit#gid=0

- add new sites to the list (maybe we can crowdsource with the team, or with the larger Mozilla community?)

- maybe we could ask for help testing + diagnosis, outreach, etc.

- there's a #covid19 channel on slack, we could scrape the comments / link for other sites (or ask them for suggestions)



March 12 2020


Timing of Interventions verification & Bugs (Mike)

Checklist for Intervention Release: https://docs.google.com/document/d/1hyIv14aKDzGixuPz3rFHmVSXs-DlTFfJoleQxILwGws/edit?ts=5e61bdaf#

Meta bug for each release: https://bugzilla.mozilla.org/show_bug.cgi?id=1621062

Release Calendar: https://wiki.mozilla.org/Release_Management/Calendar

If we identify sites that no longer require interventions or overrides, can we please file a bug in Bugzilla in Web Compatibility::Interventions & Tooling to remove it? Following along in GitHub comments can get confusing.

It's probably a good idea to create a link to the original landing patch via See Also. And leave a comment.

We should probably time new testing to happen after we land a new version -- that way we have a few weeks to work on removing uneeded interventions.

Issues related to Login (SV)

We have noticed a lot of Login related issues. Most of them we cannot reproduce.

mike: This sounds like https://bugzilla.mozilla.org/show_bug.cgi?id=1618610

mike: There's a pref as well `network.cookie.sameSite.laxByDefault`

Action item: link the Bugzilla bug to the webcompat Login issues

explanation here: -> https://web.dev/samesite-cookies-explained/

(FYI) ML is helping a lot (SV)

(11 issues closed in the last 24 hours)

The great sickness (World)

We're trying to protect ourselves as best as we can

(FYI) Test Party Webcompat (SV)

- successful testing event - 13 participants - 2h - 45issue reported

- participants enjoyed having the report site issue directly on the browser



February 19 2020


FYI Anonymous reporting feedback (SV)

Overall it is working fine, moderating takes about 1h for 50-60 issues.

It would be useful to have the public link in the issue, in order to moderate and triage it at the same time - Karl is working on it.

Concern: issues will pile up like crazy since a lot of them are being reported (seems to be more than before /day), and we also need to focus on the OKRs, and also take PTO soon in this quarter - ML can help with this.

mike: will sent an email to Nemo about ML, CC SV team in the email

Anonymous reporting closed issues before being accepted (SV)

Some issues get closed before being accepted due to GH's lag/or by mistake and no longer end up in Public repository (thus they get lost).

Is there any mechanism to bypass this behavior so no issue will get lost?

mike: reopen a private closed issue and try to accept it again (test on staging)

Action item: SV team to check the flow for private closed issues, log issue in case it does not work as expected

Fenix issue on Tablet layout (SV)

Tablets receive phone UA layout on Fenix istead of tablet layout.

We recommended to test on Firefox Release or Firefox Beta until Fenix issue is fixed. What do you think ?

mike: continue testing on Firefox Release or Firefox Beta

New Webcompat form (SV)

Any news if the new form will remain or will it be closed/disabled?

mike: leave it on and collect data

oana: maybe request feedback from the users

mike: ksenia is writing tests for the new form, probably she will re-write the new form (JS - implementation)

mike: make small improvements

mike: meeting with Open Innovation and discuss about it next Monday



February 5 2020


Anonymous reporting on Staging (SV)

Are there manual tests needed here?

mike: is this on staging yet?

sergiu: not sure, but we saw it's merged.

mike: I'll get in touch with Karl, and we'll get some documentation. some light testing would be good.

mike: I'll deploy to staging in a few hours.

Action item: Testing anon-reporting tomorrow (06.02.2020) on staging

Action items following the All Hands (SV)

Are there any action items for us, as a result of the Berlin talks?

Any Tracking Protection testing news?

mike: keep doing what we are doing

mike: no news on TP for now

mike: focus on interventions/TSCI?

Presentations (good to watch):

Berlin 2020 - Firefox Web Technologies All Hands https://onlinexperiences.com/Launch/Event.htm?ShowKey=44908&DisplayItem=E356040 - get a sense of future directions

Berlin 2020 - Thursday Plenary - different cultures & communication https://onlinexperiences.com/Launch/Event.htm?ShowKey=44908&DisplayItem=E354347

Interventions testing (SV)

Is it ok to test the Interventions on Fennec Beta?

mike: it's probably good enough!

sergiu: we still have issues that are reproducible in beta/release but not Fenix Nightly.

mike: let's go ahead and close those.



January 22 2020


Non-compat issues rule (SV)

https://github.com/webcompat/web-bugs/issues/47700 - involved payment, valid data was needed to create account, not reproducible after clearing cookies.

Update on (anon) private reporting for webcompat.com (Mike)

mike: we hope to have this ready by the end of the week!

Update on Sergiu's surfing trip.

Went snow surfing.



January 09 2020


Enterprise TSCI (SV)

More context on https://github.com/mozilla/webcompat-team-okrs/issues/117

sergiu: what's up with this?

mike: this is a list of sites that are important for Enterprise. We may build a TSCI report for these sites this quarter, or propose a new Outreachy mentorship for it.

sergiu: probably we can come back to these after OKRs are finished.

oana: probably the next sites we test are 10 to 20.

Top Site Testing (10 - 20) (Mike)

mike: let's focus on severity-critical and -important!

Content Moderation (Mike)

mike: if anything illegal or harmful comes up during triage, please ping miketaylr immediately. no need to comment on issues, mike will take care of it until we have a better process in place.

mike: please read the below resources as well, they may be helpful.



December 11 2019


ML issues with >95% probability (Sergiu)

We find it quite useful and we should continue closing these issues without checking them. We did not see any users that commented on closed "ml-probability-high" stuff.

Although, we ran into a contradictory labeling situation on a valid login related bug where the issue was filed by a user (https://github.com/webcompat/web-bugs/issues/46318) which was not labeled with `ml-probability-high` and another issue filed anonymously (https://github.com/webcompat/web-bugs/issues/45304) which was labeled (`ml-needsdiagnosis-false` and `ml-probability-high`).

Quality of reports (miketaylr)

mike: Rosana wondered if the subcategories were helpful. sergiu: When we don't have enough info from the description, it can be helpful. But so many are invalid.

oana: It's useful to have that icon, nice for the user.

Form v2 Metrics (miketaylr)

Valid/invalid: https://docs.google.com/spreadsheets/d/1tQkQ3NvvwWv6KMKZN4LUhfo7icBEfl8EZbttP6IZpGE/edit#gid=0

GA completion metrics: https://docs.google.com/spreadsheets/d/1SIhjfxyioHSfnMlm3H6iCK3YyWzzl1i2VUw5M6ZOAOs/edit#gid=777563269

TP (resource blocking) & ETP (cooking blocking) site breakage (Oana)

Are there any news from Security team, related to mobile device issues? Should we continue logging them on "Firefox - Protection UI" ?

mike: We can continue to file mobile ones there, according to Privacy team.



November 20 2019


ML issues probability > 95% (Cipri)

Starting next week we could close all the issues that have a probability above 95% to be invalid.

mike: what’s the rationale for waiting a week? ciprian: just wanting to gather some feedback, or know if it’s safe. mike: let’s go ahead and close them.

New Webcompat form (Oana)

We see more and more issue reported using the new form. We did not see any improvement in the quality of the issues reported, the content (description) is still vague and with few details.

oana: it’s probably easier to report an issue. looks nicer, but the content – it’s still as vague as ever. it’s probably better to focus on how to improve quality. mike: matt (the UX designer) shared some thoughts here: https://docs.google.com/document/d/1Qq0xTXTyy8xr9EE0SA7Abxesp39UaOiWRjlS0F5ff4I/edit

Incompat add-on (Sergiu)

https://github.com/andreicristianpetcu/incompat

sergiu: we got in touch with the developer, and now besides its original functionality, when we’re on Github looking at a web-bug it tells us how many issues are open and closed for a given domain. we can take a look at closed issues and see if the current bug is a duplicate.

mike: wowowowow, we should send him something.

Triage backlog

OI team trying to understand v1 vs v2 numbers (Mike)

https://docs.google.com/spreadsheets/d/1SIhjfxyioHSfnMlm3H6iCK3YyWzzl1i2VUw5M6ZOAOs/edit#gid=0



November 6 2019


webcompat.com testing

oana: Will there be another iteration on the new form? If needed, we can take a shot at it.

Action item: test the new form while doing TSCI top site testing, report bugs and suggestions.

ML add-on

sergiu: Do we have some data about how the ML performed? Is there a way for it to go live?

Proposal: turn the system on and have it

sergiu: has the system ever predicted something at 100%?

Another idea: automatically close (and label) issues at 95%, flag ones at 90% for manual verification, etc.

Facebook testing

cipri: You mentioned some time ago, about a session of Facebook testing. Are we still on?

Action item: mike to send profile of test account to st peter. Also, ask FB to flag the user to not get banned since it’s a test account.



October 16 2019


Secure Connection failed issues

Can we close them if the certificate is not A grade? mike: yes, except TSCI level sites.

ML add-on

How come apparently identical issues get different results from the add-on? mike: here's an older document with the experiment design, https://docs.google.com/document/d/1XFNVNx_0Eb3OozzZyg2wrfCOzNQH6KI5RfBAp5Tq0K8/edit mike: some word salad about statistics mike: how accurate is it? sergiu: it's about 75% accurate oana: for duplicates, it's not useful. sergiu: is the bot already trained? or still being trained? https://github.com/mozilla/webcompat-ml-webext https://github.com/mozilla/webcompat-ml-webext/blob/master/src/background.js#L1 oana: let's continue testing for 2 weeks. sergiu: we could improve this addon by having some kind of interaction method. when it's wrong, give it some feedback as to why it's wrong (or right). mike to follow up with nemo about this. oana: this is only possible when we get a notification.

Webcompat.com - new design

oana: Can I have the link to the new form? I want to fiddle around with some automated tests, as part of my own curiosity. mike: send oana the cookie name/value for staging.webcompat.com

IRC follow up

mike to send email to dennis