Hi all!
Here are the notes from our weekly Tor Browser meeting which we had earlier today. The chat log can be found at
http://meetbot.debian.net/tor-meeting/2018/tor-meeting.2018-04-30-18.00.log....
and our pad items were:
Monday April 30, 2018 Discussion question(s): -igt0 asks: I know we have talked about accessibility before, about mobile, how much we care right now? I am asking because of ticket #25902. Android accessibility services allow password managers to listen for user events. e.g. typing a password in a login form. -Upcoming releases -What is our way forward with our extensions for ESR60? [sysrqb and arthuredelstein will look next week into what is needed to get our extensions integrated into a working ESR60 build]
mcs and brade: Last week: - Helped with triage of incoming tickets. - Continued rebasing Tor Browser updater patches for ESR60 (part of #25543). - Tested on macOS. - Participated in the UX/Tor Browser "sync" meeting. This week: - Finish rebasing Tor Browser updater patches for ESR 60. - Test on Linux. - Test on Windows. - Monitor #25807 (Can not request bridges from torproject.org (App Engine is broken for moat)).
igt0: Last week: - #25810: Backported few fixes affecting Orfox to tor-browser-52.7.3esr-8.0-1. - #25974: Make sure Android Oreo(API level 26) autofill feature is disabled (looks like it doesn't work on FF, i am trying to make autofill to work on simulator to be 100% sure) This week: - Review and smoke test rebased Orfox patches branch. - Build and test on android the rebased tor browser patches for ESR60 to make sure we are not going to have problems in the alpha release.
tjr - Found and bypassed a crash bug for MinGW. Looking at two, maybe three crashes currently. - 1) A crash in https://searchfox.org/mozilla-central/source/nsprpub/pr/src/threads/prtpd.c#... that happens on TC - 2) This assert: https://searchfox.org/mozilla-central/source/layout/svg/nsSVGIntegrationUtil...
- Have been learning more about debugging symbol-less (in WinDBG). Feel pretty comfortable identifying where I am in assembly (assuming Debug asserts are present...) [GeKo: Do you feel you could write up what you learned and, say, add this to the Hacking doc we have? I guess this would help us a lot, too.] - Wrote it up at https://trac.torproject.org/projects/tor/wiki/doc/TorBrowser/Hacking/Debuggi...
GeKo Last Week: - wrote a patch for the rustc cross-compilation for Windows - fixed the rustc cross-compilation for macOS (still need to clean-up the patch for review) - helped with Sponsor4 reports (yes! those were still a thing!1!) - wrote/backported small patches for tor-launcher (getting Moat to work again) and for tor-browser (fixes an off-by-one error) - started to review the patch rebase for ESR 60 - gave the circuit display patch (#24309) another round of testing - patch reviews/merges (#25898, #25458, #25810, #25973) This week: - finish up the rustc compilation patch for macOS - rebase review - update macOS toolchain - plan to take days off (5/1 and 5/3+5/4)
boklm: Last week: - patches for #25817 (add ansible roles for nightly builds) and #25318 (adding email notification for nightly builds) are now ready for review - installed some VMs for running the Tor Browser testsuite, and started working on some ansible roles for the setup - worked on a patch for #25876 (Source release tarballs for Tor Browser) to automate creation of source tarballs - worked on #25862 (Clean up wrapper script/CFLAGS and friends mix on Windows) which is now ready for review - investigated the reproducibility issue with #16472 (binutils update) - started reviewing #25894 (Get a rust cross-compiler for Windows) and #25832 (Enable pthread support for mingw-w64) This week: - help with building the new releases - continue investigating issue with binutils update (#16472) - continue work on testsuite VMs setup
sysrqb: Last week: - Worked on TBA patch rebase (#25741) - Worked on updating Orfox's https-everywhere addon (#25603) This week: - Found -8.0-1 branch is busted when building Orfox, fixing now (#25980) - Merge #25603 after we fix bustage - Continue testing #25741, put in needs-review - Start working on another TBA ticket
pospeselr: Last week: - Worked on #23247 (Communicating security expectations for .onion), most of the way there but need to plumb down some logic exposing mixed-mode info for .onion to work correctly This Week: - Finishing up #23247 - Entertaining Shane this afternoon - Flying to the east coast to visit family Friday afternoon
arthuredelstein: Last week: - https://trac.torproject.org/25938 (backport 1334776) - Updated https://trac.torproject.org/25543 (Rebase Tor Browser patches for ESR60) - Revised patch for https://trac.torproject.org/24309 (Activity 4.1: Improve how circuits are displayed to the user) - Revised patch for https://trac.torproject.org/22343 (Save as... in the context menu results in using the catch-all circuit); will post after more testing - Did more cspace calculations for https://trac.torproject.org/25575 (Server space request for hosting Tor Browser downloads) updated by arthuredelstein
This week:
- Try to get a preliminary desktop build working for TBB/ESR60
- Continue revising https://trac.torproject.org/25543 branch
- Permissions patch revision: https://bugzilla.mozilla.org/show_bug.cgi?id=1330467
Georg
Hi all!
Here is another round of notes from our weekly Tor Browser meeting, latest edition today.
The chat log can be found at
http://meetbot.debian.net/tor-meeting/2018/tor-meeting.2018-05-07-17.59.log....
and our pad items were:
Monday May 7, 2018
Discussion: -meeting invites (specific meeting day one day earlier?) [GeKo: We try to have a dedicated tor browser team day before the official team meeting day this time and see how it goes] -how much should we care about nightlies being busted (build- and function-wise) while having all the transitions to ESR-60 underway? [GeKo: we'll update to newer toolchains starting with Linux presumably breaking the other platoforms. I'll annouce that to tbb-dev and we'll fix it as fast as we can. That way, we get the transition to ESR60 with no additional effort while at least having nightlies on Linux available all the time]
GeKo: Last Week: -reviews and release preparations -finished macOS cross-compilation of rustc (https://trac.torproject.org/projects/tor/ticket/25975) -started looking into updating macOS toolchain (https://trac.torproject.org/projects/tor/ticket/24632) -further rebase review (https://trac.torproject.org/projects/tor/ticket/25543) -made small progress on https://bugzilla.mozilla.org/show_bug.cgi?id=1390583 (stylo build bug for windows); still need to figure out some missing pieces -took 1 1/2 days off -Richard: How is the monthly meetup with Pari going on getting up-to-date about issues users are facing? [GeKo: not sure, Richard will ping Pari] This Week: -release help (signing etc.) -being of the month admin work (ticket keyword updates/roadmap updates etc.) -finish rebase review -macOS toolchain update -start with the network code review ESR52-ESR60 -further bug triage -I plan to be afk on Thursday (public holiday) and, potentially, Friday
tjr
RE https://trac.torproject.org/projects/tor/ticket/26019
I think it should be possible to easily reintroduce separate optimization settings so long as the setting exists on the JS Compartment.
I did this for Timer Precision Reduction: https://reviewboard.mozilla.org/r/226564/diff/10#index_header
MinGW Work: https://bugzilla.mozilla.org/show_bug.cgi?id=1389967 [GeKo: so, you get the browser running and now are hitting the shutdown crashes? Do you have a patch set somewhere I could try locally to bypass the crashes I see?]
Actually, this is in TaskCluster. Locally it seems to run (I don't seem to be hitting the SVG Asert anymore, although I haven't changed anything.) I have not tried esr60 actually, this one a commit midway through 61.
However, https://hg.mozilla.org/try/pushloghtml?changeset=3508e0987cd1d38223005637887... is my patchset, of which the most important one is
https://hg.mozilla.org/try/rev/ee28ff6445d0 to avoid runtime crashes
More investigation is needed on my part before I feel comfortable telling you "Yes, try it, you won't be wasting your time"
pospeselr:
Last Week:
- continued work on #23247 (Communicating security expectations for .onion)
- have a patch working in ~90% of cases
- built test environments for various mixed-mode scenarios
This Week
- more #23247
- there are a coupe of edge cases not explicitly covered by the google doc, will send out an email later with details
- Arthur: one thing not accounted for with regards to this ticket is the hanger menu off of the info/lock icon shows 'Connection is Not Secure' and the wrong icon for https(s) onion domains
Since you've had to mess around in there for the new onion circuit UI, could you point me to where this rendering logic is handled?
[Arthur writes:
Here's where the security view and security subview are implemented in XUL:
https://dxr.mozilla.org/mozilla-central/source/browser/components/controlcen...
https://dxr.mozilla.org/mozilla-central/source/browser/components/controlcen...
A lot is done in CSS:
https://dxr.mozilla.org/mozilla-central/source/browser/themes/shared/control...
And I think most of the logic is here:
https://dxr.mozilla.org/mozilla-central/source/browser/base/content/browser-...
]
mcs and brade: Note: To accommodate a series of family events, Kathy and I will have limited availability for Tor work from May 16 - May 30. Last week: - Finished rebasing Tor Browser updater patches for ESR 60; tested on Linux and Windows. - Spent some time thinking about #25694 (improve updater UX) and made a comment on that ticket. - Participated in the UX/Tor Browser "sync" meeting. This week: - Review Matt's changes for #25750 (update Tor Launcher for ESR 60). - Do some testing of the revised Tor Launcher in an ESR60-based browser. - Follow up on some ESR60 updater loose ends. - File a Bugzilla bug for #25909 (disable updater telemetry) - Start to work on #22074 (Review Firefox Developer Docs and Undocumented bugs since FF52esr)
igt0: Last Week: - Investigated a bit more about accessibility on Fennec(#25902). - it fires events just when the Exploration by Touch feature is enabled. - Exploration by Touch changes how the user interacts with the apps and it must be enabled by the user, so IMO this bug doesn't have a high priority.
- Fennec is leaking the user's OS language (#26018)
- I sent a patch to mozilla: https://bugzilla.mozilla.org/show_bug.cgi?id=1459089
- HLS player on Android is not using the central Proxy Selector (#21863)
- Sent a patch to mozilla: https://bugzilla.mozilla.org/show_bug.cgi?id=1459420
This Week: - More linkability/fingerprinting/proxy bypass investigation on Android.
- Tor button with 60ESR Desktop/Mobile.
sysrqb: Last week: - Patch for building Orfox (#25980) - Investigated using Tor-Launcher with 60ESR and put patches in-review (#25750) - Merged https-everywhere update for Orfox (#25603) This week: - Return to testing TBA patches (#25741) - Read about igt0's investigation results on fingerprinting vectors - Begin designing Tor Launcher for Android - Pick up Bug1440789 again (upstream feature add-on support bug)
arthuredelstein: Last week: - Finished a new patch for #22343 (Save as... in the context menu results in using the catch-all circuit) - Built Tor Browser releases and confirmed and signed hashes - Worked on getting tor-browser-build to work with rebased patches This week: - Today I will post the latest 25543 branch with updater patches from mcs and brade) - Finish getting tor-browser-build branch for Linux [OS X will wait for #9711] - https://bugzilla.mozilla.org/show_bug.cgi?id=1330467 (When "privacy.firstparty.isolate" is true, double-key permissions to origin + firstPartyDomain) - https://trac.torproject.org/projects/tor/ticket/25794 (Sanitize Pointer Events), https://bugzilla.mozilla.org/show_bug.cgi?id=1363508
boklm: Last week: - helped build the new releases - worked on testsuite VMs setup - updated patches for #25817 (Add ansible scripts for setup of nigthly build server ) and #25318 (Add Tor Browser nightly builds email notification) after review comments - made some progress on the binutils bisect to find commit responsible for reproducibility issue in #16472 (Upgrade Binutils to 2.25+) This week: - Will be afk on Thursday and Friday - publishing the new releases on Wednesday - continue bisecting the binutils issue - continue work on testsuite VMs setup
Georg
tor-project@lists.torproject.org