[tor-project] Tor Browser team meeting notes, 20 Feb 2018
gk at torproject.org
Tue Feb 20 20:35:00 UTC 2018
Here are the meeting notes for the Tor Browser meeting on 20 Feb (we had
no meeting last week).
The chat log can be found at:
And here come the notes from our pad:
Tuesday, February 20, 2018
Spent a lot of time on Timer Rounding and Fuzzing. Spending yet more
time on it, but made good progress..
Got a patch written to hide the canvas prompt unless triggered by
The Keyboard layout patch has problems.
It's still leaking data:
It breaks a lot of stuff (see Duplicates also):
I.... don't really understand this one. Arthur can you walk me
through it? (Arthur says: Sure. After the tbb meeting?)
Made headway on "Why doesn't MinGW run" - but need assistance debugging
Added a MinGW 32 Opt build to TC and am working on a x64 Build
There is a draft for the Fingerprinting Protection Help Page
mcs and brade:
- Completed some code reviews:
- #15599 (Range requests used by pdfjs are not isolated).
- #22614 (Make e10s/non-e10s Tor Browsers indistinguishable).
- #22659 (Changes to `intl.accept.languages` get overwritten).
- For #23136 (Moat integration):
- Confirmed that the BridgeDB Moat distributor is now working
- Posted a patch for #19910 (Rip out optimistic data socks
handshake variant (#3875)).
- Found and fixed #25266 (PT config should include full names of
executable files on Windows).
- Created some test builds for the Moat feature and posted them.
Planned for this week:
- Debug #23136 (Moat integration) which is not working for Georg.
- Review some more patches.
- Triage the Tor Launcher bug list (set priorities, close outdated
- Worked on #21777 (I finally got a Tor Browser build somewhat
running on Windows by using a clang-based cross-compiler)
- Release preparations for the alpha
- More work on #21777 (trying to get mozilla-central built with the
new clang-based cross-compilation toolchain)
- Getting 8.0a2 out
- Think about summarizing the security slider discussion we had so
far and moving the topic forward
- Finally writing the tests for my mixed content blocker .onion
- Worked on #25260 - merging moz build files into tor-launcher
- This ticket is becoming something...different, so I may repurpose it
- We had an unofficial group meeting on Monday and discussed some
stuff about mobile
- Worked some more on the TBA roadmap
- This week
- Finish TBA roadmap
- Continue working on #25260
- Start porting Orfox patches onto mozilla-central
- Should we takeover responsibility for Orfox (and release Orfox
updates) while we work on TBA?
synced with Pari on February's spreadsheet regarding Tor Browser
Worked on #25112 - sandboxing issue on Vista
Tracked down point of failure (ntdll patching), investigating fix
GK: could you ping me offline with the info you mentioned having on
getting debug symbols working in windows builds
basically printf debugging atm which is fine for now but would like
to be more efficient in the future [Full-fledged debug builds are broken
right now (but mozilla-central should have that fixed
(https://bugzilla.mozilla.org/show_bug.cgi?id=1372958 has the details).
Your best bet with a debugger right now is using the optimized build we
usually build, however with stripping disabled; make sure you have
--disable-strip in your .mozconfig and if you package a build you need
--disable-install-strip as well; bobowen is the windows related
sandboxing guru and e.g. in #boxing and pretty helpful if you want
feedback for patch ideas GeKo]
pospeselr: --disable-install-strip was the missing piece, thank you!
confirmed fix for #24341 (ubuntu build issue)
uplifting patch for #22794, fix for #25112
more bug fixes for the bug gods
- Investigated the #25247, I was not able to reproduce neither
in my machine nor in a debian virtual machine. (Tried different
resolutions and pixel densities)
- Looked into the
https://bugzilla.mozilla.org/show_bug.cgi?id=942652 , the ticket is out
to date, good part of the firefox code is already updated to use
except the sync code. I am now looking into the sync background
service, there is a bug about using a different http lib
- sysrqb: I assume we'll want to disable Sync at
compile-time, so we may not need to do a lot
- igt0: interesting, I found these tickets
https://trac.torproject.org/projects/tor/ticket/10368 , i am not sure it
reflects the current state. (where it can be enabled for sync bookmarks,
passwords and history)
- Help sysrqb to write the TBA roadmap/documentation
- Contact the Mozilla team working in the sync service about
#942652, to make sure we are aligned what is the proper solution.
- worked on some patches for:
- Bug 25020: add a tbb_version.json file
- Bug 25178: re-enable windows-x86_64 build in 'make testbuild'
- Bug 24879: enable fetching of new commits by default for
- started working on updating binutils (#16472) and gcc (#25304)
- help with build of the new alpha release
- continue working on:
- toolchain updates (#16472 and #25304)
- shipping auto-updates for Tor Browser nightly channel (#18867)
- testsuite updates/fixes
- Uplifted patches to Mozilla (checked in)
Also worked on (but not finished):
Current remaining patches for uplift: https://torpat.ch/short
Two bugs mentioned above, and try to finish/uplift
- Submitted on Feb 12 SOI to DRL for TB desktop work (small grant -
nothing too crazy, implement ux work from anto/otf grant and keep up
with mozilla releases/security stuff)
- organize team roadmap prep work for Rome - will share more this
week (let me know if you have any questions)
- moving some work w/ TB team to March because ux team has to get
website work done by rome
- prep for team meeting day in rome (should we create pads and share
stuff before hand or prepare an agenda for the meetings?)
-------------- next part --------------
A non-text attachment was scrubbed...
Size: 833 bytes
Desc: OpenPGP digital signature
More information about the tor-project