19:01:52 <mikeperry> #startmeeting tbb-dev
19:01:52 <MeetBot> Meeting started Mon Nov  9 19:01:52 2015 UTC.  The chair is mikeperry. Information about MeetBot at http://wiki.debian.org/MeetBot.
19:01:52 <MeetBot> Useful Commands: #action #agreed #help #info #idea #link #topic.
19:02:19 <kernelcorn> ok, well I guess I'll start off by mentioning that I just created #17569 a few seconds ago
19:02:25 <mikeperry> hello everyone, welcome to the TBB dev meeting
19:02:38 <special> nickm: at work at an office today, can't chat
19:04:12 <arthuredelstein> hi everyone
19:04:18 <mikeperry> kernelcorn: I suggest we table that until later. that's not a flamewar I feel like having right now
19:04:19 <nickm> special: okay, no worries.  Just wanted to know if you'll be coming back to #15588 or if I should try to find somebody else to bang on it. :)
19:04:40 <kernelcorn> mikeperry: fine by me, I have no intention of starting a flamewar
19:04:45 <special> let's say yes unless somebody else is being blocked by it
19:04:52 <special> I want to, just struggling to find time and energy lately
19:05:13 <nickm> special: afaik ricochet is the main thing that wants it.  But IMO it's a very good idea.
19:05:22 <mikeperry> ok, anyway, last week, I wrote our status report, helped with the hardened release blog post, but was otherwise consumed by ED interviews and related things
19:05:55 <mikeperry> this week I am going to try to page back in a bunch of tor work related to padding, guard discovery, and load balancing
19:06:09 <mikeperry> so I won't be doing too many TBB things. I will try to keep an eye on tickets though
19:06:14 <mikeperry> that's it for me
19:10:08 <mikeperry> anyone here? who wants to go next?
19:10:48 * mcs will go next
19:10:53 <mcs> This past week, Kathy and I reviewed the patch for #9659 and created a revised patch.
19:11:00 <mcs> Based on a comment from a cypherpunk, our patch may need to be revised. Other opinions are welcome.
19:11:07 <mcs> We also reviewed the patch for #16990, triaged assorted bugs, etc.
19:11:12 <mcs> Next on our list:
19:11:17 <mcs> - Revisit the #9659 patch.
19:11:22 <mcs> - Look at #17248.
19:11:27 <mcs> - Tor Browser 5.5 work: develop patches for #17442, #16940, and #17344.
19:11:59 <mcs> That's all for us.
19:12:33 <GeKo> here is what i did:
19:12:49 <GeKo> i worked on the releases, #17406 and #16990
19:13:25 <GeKo> started with #15578 and filed #17568 for the remaining clean-up our controller code
19:14:16 <GeKo> this week i plan to get #15578 fixed and i plan to do some more code reviews
19:14:23 <GeKo> (patches are piling up)
19:14:24 <s7r> are we still thiking about if we should have SR keys or not? (scrollback)
19:14:58 <GeKo> it seems we have two regressions in the alpha series: #17550 and #17561
19:15:11 <GeKo> we shuold address them before the next release is coming out
19:15:16 <GeKo> *should
19:15:21 <GeKo> that's it for me for now
19:15:24 <mikeperry> GeKo: if you need help reviewing C++ patches, I can do that.
19:15:53 <GeKo> yeah, i'll ping you in this case
19:16:59 <mcs> r.e. #17550, does anyone on the Tor Browser team have access to a Windows 10 system?
19:17:15 <mcs> (we do not but could work on solving that problem)
19:17:38 <GeKo> not me
19:17:56 <mikeperry> we need windows 10 boxes to also test for signing issues, too
19:18:04 <mikeperry> hrmm
19:18:13 <mcs> #17550 seems like a font metric problem (definitely a strange problem though)
19:18:14 <GeKo> ywaning said that the signature was fine on his box
19:18:31 <GeKo> so, 2:1 for the signing is working :)
19:19:23 <mikeperry> mcs: Yawning is usually around on IRC
19:19:48 <mcs> mikeperry: I know; thanks.
19:19:53 <mikeperry> not sure what timezone he operates on
19:20:47 <mikeperry> ok, who's next?
19:21:22 * arthuredelstein can go
19:21:38 <arthuredelstein> Last week, I wrote a patch for #17446.
19:21:56 <arthuredelstein> Also I wrote a fix for GeKo's find in #16990.
19:22:07 <arthuredelstein> I submitted patches to Mozilla for bugzil.la/1174386 (aka #5926),
19:22:12 <arthuredelstein> bugzil.la/1217290 (#16005),
19:22:16 <arthuredelstein> bugzil.la/1222285 (#15646),
19:22:22 <arthuredelstein> and bugzil.la/1216882 (#14716).
19:22:47 <arthuredelstein> I also spent some time working to revise our patch for bugzil.la/1121643 (#13313). The font code in Firefox is simply crazy.
19:22:54 <arthuredelstein> Also I submitted #17565 this morning.
19:23:12 <arthuredelstein> This week I'll work more on bugzil.la/1121643, revise some more Mozilla patches that need revising, and have a look at #17568.
19:23:34 <arthuredelstein> that's it for me.
19:25:25 * boklm can go next
19:26:06 <boklm> This past week I updated the testsuite to do a "SIGNAL NEWNYM" beetween test retries, to remove false positives caused by slow circuits
19:26:09 <boklm> I made a patch for torbrowser-launcher to make it use our updater's xml to get the stable version number
19:26:18 <boklm> I looked at why the tests fail on the hardened tor browser. One reason is mbox, so I started looking at #17325
19:26:29 <boklm> After disabling mbox, the selenium tests are running, but not the mozmill ones, so I opened #17567
19:26:41 <boklm> This week I'm planning to work on #17325 and investigate #17567
19:26:54 <boklm> I also noticed on http://trychooser.pub.build.mozilla.org/ that they have a linux64-asan option so I will try it
19:27:03 <boklm> that's all for me
19:27:11 <GeKo> interesting re 17567
19:31:19 <mikeperry> boklm: did #17567 give a longer stacktrace, or just that?
19:31:34 <boklm> mikeperry: only that
19:31:41 <mikeperry> oh:
19:31:45 <mikeperry> ==12494==AddressSanitizer: while reporting a bug found another one. Ignoring.
19:31:52 <mikeperry> sad
19:33:02 <mikeperry> can you maybe get one from gdb? https://trac.torproject.org/projects/tor/wiki/doc/TorBrowser/Hacking#Usinggdb
19:33:48 <boklm> mikeperry: I will try that
19:35:09 <mikeperry> boklm: also giving an example way to reproduce the crash on the bug will be helpful too
19:35:26 <boklm> ok
19:35:27 <mikeperry> boklm: like what args to https://people.torproject.org/~boklm/tbbtests-doc/usage.html you used (if you even used that to trigger this)
19:36:20 <boklm> ok, I'm adding that to the ticket
19:37:27 <mikeperry> cool, thanks
19:39:37 <mikeperry> boklm: btw someone filed https://trac.torproject.org/projects/tor/ticket/17570. Do we still regularly run the tests to check the HTTP vs HTTPS script inclusion?
19:40:12 <mikeperry> the ones from #13053
19:41:07 <boklm> yes, we are running them
19:41:38 * boklm can't see the content of #17570
19:43:45 <boklm> the tests we are doing are in: https://gitweb.torproject.org/boklm/tor-browser-bundle-testsuite.git/tree/mozmill-tests/tbb-tests/noscript.js
19:44:28 <mikeperry> boklm: it basically sounds like a #13053 regression. their repro test case isn't very good either, so seeing the GPG content probably won't help you very much
19:46:37 <mikeperry> aha
19:47:46 <mikeperry> I think I can turn their janky example into something for Giorgio. I got it to work
19:48:16 <GeKo> !
19:48:41 <mikeperry> maybe. it's actually not clear what I got working yet
19:49:30 <mikeperry> I think just some CSS has loaded here, not actual scripts.. the test site they are using has a self-signed cert, so it's all kind of jank to figure out what is blocking what where
19:51:15 <boklm> are they including the http script using <script src="http://..." /> or doing something else ?
19:53:32 <mikeperry> ok, yeah, all that allowing the self-signed cert seems to change is that CSS is loading.. it's not loading the JS for me still
19:53:53 <GeKo> the same here
19:55:28 <mikeperry> yeah, at first I thought that because the site used <base href="https://site.com"> and then a relative <script src="/somescript.js"> that NoScript might have gotten confused
19:55:37 <mikeperry> but that does not appear to have actually happened
19:56:30 <mikeperry> ok, well I will comment on the bug saying so
19:56:36 <mikeperry> anything else for this meeting?
19:57:27 <mikeperry> boklm: I will add you to Cc, and ask that if they can get a test case that is easier to verify, that they encrypt it to you as well
19:57:43 <boklm> ok, thanks
19:59:37 <mikeperry> alright, thanks everyone!
19:59:45 <mikeperry> #endmeeting *baf*