15:01:42 <mikeperry> #startmeeting
15:01:42 <MeetBot> Meeting started Fri Jul 18 15:01:42 2014 UTC.  The chair is mikeperry. Information about MeetBot at http://wiki.debian.org/MeetBot.
15:01:42 <MeetBot> Useful Commands: #action #agreed #help #info #idea #link #topic.
15:03:00 <mikeperry> so this week, I set up three build machines in my apartment, made a branch for mjuarez for website traffic fingerprinting, but then stalled out due to the missing langpacks for this upcoming Firefox release
15:03:47 <mikeperry> I started looking into security slider prefs also, and hope to have a list of the up today
15:04:43 <mikeperry> as hopefully everyone is now aware, there is a tor security release coming, along with the FIrefox security release
15:05:13 <GeKo> well, at least the latter is coming next week...
15:05:24 <GeKo> see Roger's last mail
15:09:52 <mikeperry> the last one I saw said fridayish
15:10:49 <boklm> there was one today
15:11:09 <boklm> <20140717225404.GC7437@moria.seul.org>
15:12:47 * GeKo wonders about that reaaaaally long July 17
15:12:51 <GeKo> ;)
15:13:32 <boklm> ah, that was today, or yesterday depending on timezone :)
15:17:45 <mikeperry> well I still can't find this mail.. anyway, we should keep going
15:19:31 <GeKo> mikeperry: so what would be good for next week would be a tor browser branch for ESR 31
15:19:47 <GeKo> no need to have some rebased patches yet
15:20:01 <mikeperry> ok. it looks like they just tagged a secodn build of it. I got that mail :)
15:20:03 <GeKo> just something for me to test our toolchains
15:20:32 <n8fr8> next week? grr, i am going offline next week :(
15:20:55 * n8fr8 interrupts rudely
15:22:15 <GeKo> mikeperry: or maybe I put one in my user repo... might work as well
15:24:26 <mikeperry> yeah, it should be possible to change the TBB remote URL location easily now
15:24:54 <GeKo> then I put that on my plate
15:25:22 <GeKo> until we have an official esr31 branch
15:26:29 <mikeperry> should we enable baseline jit for this release next week btw?
15:26:53 <GeKo> hmm...
15:27:53 <GeKo> I'd like to enable that stuff when we make the JIT pages non-writable
15:28:12 <GeKo> escpecially as we have no new user report complaining ASAICT
15:28:18 <GeKo> *reports
15:28:28 <GeKo> AFAICT, too
15:28:43 <mikeperry> well the tails people posted a complaint about sharedcoin on blockchain.info not working
15:29:17 <GeKo> and aneabling baseline jit fixed that?
15:29:22 <GeKo> *enabling
15:30:28 <mikeperry> not sure. I could ask. we also need to move all of this jit conversation out of the security slider ticket and into its own ticket I think
15:30:39 <GeKo> good idea
15:31:02 <GeKo> so, when this really fixes the problem, then yes, as baseline jit is not so critical security-wise
15:31:19 <GeKo> IIRC the talk with nicholas
15:32:20 <GeKo> i.e. better tested etc.
15:35:39 <mikeperry> ok, I will file that ticket. who wants to go next?
15:35:44 <GeKo> me
15:36:50 <GeKo> I hopefully fixed #12391 and #12381
15:37:12 <GeKo> the latter is not merged yet as I thought we have somebody testing it first
15:37:38 <GeKo> if that is not happening I'll merge the patch anyway and give it to our 3.6.3 QA
15:38:12 <GeKo> I think it definitely should make it into 3.6.3 if possible
15:38:42 <GeKo> then I debugged the VTV of #10599 further and it seems at least one reason of not getting proper debugging output is
15:38:59 <GeKo> a bug in the gdb shipped with Ubuntu Precise
15:39:27 <GeKo> compiling an own, new version helps but I am still investigating with the help of Google engineers
15:39:52 <GeKo> hopefully I can file a Mozilla bug with proper output next week...
15:40:24 <GeKo> then I looked at the TBB roadmap which came out of the meeting and I updated it with tickets I filed
15:40:38 <mikeperry> do you remember the JIT write protect tickt on bugzilla btw?
15:40:55 <GeKo> I think only two things a re missing ticket-wise: the mixed content blocking thing
15:41:10 <GeKo> and one test ticket
15:41:25 <GeKo> mikeperry: see #12523
15:41:35 <GeKo> https://bugzilla.mozilla.org/show_bug.cgi?id=977805
15:42:03 <GeKo> then I looked at #3455 part 1 and merged and reviewed some tickets
15:42:09 <GeKo> next week:
15:42:35 <GeKo> I try to do some more review; part 3 is realisitc not sure whether i get tp part 2 of #3455
15:43:04 <GeKo> then the VTV issue + I try to finally get rid of #9531 and the like
15:43:26 <GeKo> and, of course, the release is coming...
15:43:32 <GeKo> that's it so far from me
15:45:12 <arthuredelstein> I can go next.
15:45:21 <arthuredelstein> I've been working on #5752 and descendants. I have posted patches for #3455 and #8405 and getting lots of very helpful feedback. I'm hoping to get started on #8641 now (browser circuit status + exit UI), and if anyone has any thoughts on how that should look, I'd be glad to hear.
15:46:47 <arthuredelstein> Also I'm available for #12620 in whatever capacity
15:47:28 <arthuredelstein> That's all for me.
15:47:46 <GeKo> nice work btw, arthuredelstein
15:47:53 <arthuredelstein> thanks
15:48:03 <arthuredelstein> thanks for all your help
15:48:05 <mikeperry> yeah
15:48:21 <GeKo> yw, I am really looking forward to get this deployed
15:48:45 * boklm can go next
15:48:45 <mikeperry> I could see a "New IP" or "New Circuit" button in the #8641 menu.. it could for example just ad a new nonce into a hash for the SOCKS username+password for a tab
15:49:28 <mikeperry> that way people could pick a new circuit for just a given site
15:49:31 <GeKo> but we are not hvaing a per tab things
15:49:37 <mikeperry> I don't think the info panel should be global though
15:49:52 <mikeperry> well, for that url bar I mean
15:50:01 <GeKo> so getting the binding to the domain part properly translated into UI
15:50:15 <GeKo> for the normal user seems not easy to me
15:50:20 <arthuredelstein> One problem with picking a new circuit for one site is that the new and old identities are linkable unless we can wipe everything for that URL Bar domain.
15:50:30 <GeKo> right
15:50:44 <arthuredelstein> ...as GeKo and lunar pointed out to me
15:51:02 <mikeperry> yeah.. I think people compain about missing New Identity from Vidalia because they really want a way to get a new exit IP
15:51:14 <mikeperry> which is not exactly the same as New Circuit/NEWNYM
15:51:34 <GeKo> yes
15:51:54 <arthuredelstein> I guess sometimes, we want a new exit that *is* linkable. Because the old exit is too slow
15:53:03 <mikeperry> or banned by the site...
15:53:04 <arthuredelstein> or otherwise unacceptable
15:53:31 <arthuredelstein> The distinction between new identity and new exit would need to be very clear
15:55:04 <boklm> So this week I started integration of xpcshell tests into our test suite
15:55:17 <boklm> I ran it on a mozilla commit: https://people.torproject.org/~boklm/tmp/tests/r/browser-155c4c79792c997b8d70/browser-155c4c79792c997b8d70.html
15:55:25 <boklm> And on tor-browser-24.7.0esr-3.x-1 branch: https://people.torproject.org/~boklm/tmp/tests/r/browser-c660ca4cb98351bbce6c/browser-c660ca4cb98351bbce6c.html
15:55:45 <mikeperry> in #9442, I suggested that we provide a hidden pref for people who want "New Circuit", so we can get the technical people to stop complaining all the time
15:56:13 <GeKo> that might be a part of the solution, yes
15:57:04 <selimthegrim> what's the latest tor-experimental build version number, out of curiosity?
15:57:20 <selimthegrim> not 2.5.x?
15:57:21 <mikeperry> boklm: so the stock FF fails lots of tests on our build setup... hrmm..
15:57:46 <selimthegrim> i'm seeing if I can reproduce kpdyer's bug
15:57:46 <boklm> mikeperry: actually, some of the fails are on parts of the tree that are not built
15:57:50 <mikeperry> perhaps we just filter those tests out for now so we can focus on the ones we broke?
15:57:55 <boklm> yes
15:58:04 <boklm> I should add a filter
15:58:21 <boklm> Next week I'm planning to add a diff of test failures between current and previous commit, and launch it on all of our commits (on esr27 branch first, and restart it on esr30 when it's available), to find which commit is responsible for which test failure
15:58:52 <selimthegrim> i'll presume 3.6.2.
15:59:15 <mikeperry> boklm: awesome
16:00:13 <boklm> And then see how we can also run those tests on the binaries produced by the nightly builds of the bundle
16:01:52 <boklm> that's it for me
16:01:55 <GeKo> I need to leave now but am reading the backlog later...
16:02:19 <mikeperry> ok
16:03:17 <mikeperry> any support people here?
16:03:39 <mikeperry> Pearl Crescent is absent this week, but posted a report on TBB dev
16:05:32 <mikeperry> if there's no one else, I think we might be done anyway then?
16:05:53 <boklm> yes
16:07:19 <mikeperry> ok
16:07:40 <mikeperry> #endmeeting *baf*