19:01:56 <mikeperry> #startmeeting tbb-dev
19:01:56 <MeetBot> Meeting started Mon Feb 23 19:01:56 2015 UTC.  The chair is mikeperry. Information about MeetBot at http://wiki.debian.org/MeetBot.
19:01:56 <MeetBot> Useful Commands: #action #agreed #help #info #idea #link #topic.
19:02:09 <mikeperry> Let's get started
19:02:56 <mikeperry> Last week, I merged things for 4.5-alpha and 4.0, and tagged that release. Matching builds for 4.5-alpha-4 and 4.0.4 should be up in my homedir on people. I still need to write a tor-qa post for 4.5-alpha-4
19:03:05 <mikeperry> and we need to sign the packages and MARs
19:03:13 <mikeperry> I also tagged several tickets with tbb-4.5-alpha that I feel are important to make it in to the 4.5 series: https://trac.torproject.org/projects/tor/query?keywords=~tbb-4.5-alpha&status=!closed
19:03:24 <mikeperry> March 31st is the next Firefox point release, so 4.5-alpha-5 will be then, and we should aim to have 4.5-stable out shortly after that, and close as many of those tickets as we can for that release.
19:03:44 <mikeperry> This week we need to have the blog posts and release bundles signed by Thursday, when most of us take off for Tor dev. Or at least GeKo and I do
19:04:00 <boklm> I see that GeKo wrote a tor-qa post for 4.5a4
19:04:10 <GeKo> yes
19:04:31 <mikeperry> feel free to add things to that tag you think should make it in to the 4.5 series before we stabilize it, and we can try to discuss assignments of tickets during the dev meeting next week
19:04:44 <mikeperry> ah, ok, good. I hadn't fetched all my mail yet. thanks!
19:04:56 <GeKo> mikeperry: I added the child tickets of #9387 to the tbb-4.5-alpha tag to make the workload explicit
19:06:02 <MarkSmith> #14270 seems like kind of a big thing for 4.5 but maybe it is important?
19:06:12 <GeKo> I am still torn whether we should have #14270 in 4.5 at all
19:06:14 <GeKo> yes
19:06:22 <GeKo> it is quite late in the alpha cycle
19:06:31 <GeKo> and this feature deserves quite some testing I think
19:06:52 <GeKo> apart from the fact that we are probably running out of time due to the dev meeting etc.
19:06:56 <mikeperry> yes, it probably only makes sense as an option
19:07:50 <mikeperry> at best
19:08:40 <MarkSmith> I guess someone would need to look at #14273 and see what the core browser work will be.
19:08:41 <mikeperry> anyway, that's it for me.
19:09:04 <mikeperry> yeah
19:09:24 <mikeperry> I think I will leave it on our radar until that happens, I guess?
19:09:47 <GeKo> sounds good to me.
19:10:41 <GeKo> here is what I did:
19:11:24 <GeKo> I spent some time reviewing different patches (#14059, #5698) and worked on various release related things
19:11:42 <GeKo> today I spent most of my time looking at the windows signing
19:11:59 <GeKo> it turns out there is no tool for our purposes yet, we need to create custom code
19:12:10 <GeKo> I am currently investigating the best way to do that
19:12:34 <GeKo> I'll continue doing that this week + work on the security slider
19:12:55 <GeKo> helping with the release comes to mind as well
19:12:58 <GeKo> that's it for now
19:14:03 * MarkSmith can go next
19:14:11 <MarkSmith> This past week, Kathy and I reviewed changes for #14203 and #13882.
19:14:20 <MarkSmith> We made progress on #14631.  We expect to make changes available for review soon.
19:14:27 <MarkSmith> We also did some bug triage, e.g., #14923, #14947, #14971.
19:14:39 <MarkSmith> And we did a quick TB 4.0.3 -> 4.0.4 incremental update test on Mac OS.
19:14:47 <MarkSmith> I said this last week too, but this week we hope to finish #14631.
19:14:55 <MarkSmith> If we finish that, we could look at #12827 or #13458 if no one else gets to them first.
19:14:59 <toralf> 8a9d86b..d74a78c  -> 5/359 TESTS FAILED. (0 skipped)
19:15:13 <toralf> here at a hardened stable Gentoo
19:15:15 <MarkSmith> err #13548
19:15:24 <MarkSmith> We will also test 4.5a3 -> 4.5a4 updates once the release is available (we do not currently have a strategy for testing signed updates until after the MAR files are signed).
19:15:28 <MarkSmith> That's all for us.
19:15:46 <GeKo> the MathMl and SVG pref would be helpful, yes
19:15:52 <GeKo> *prefs
19:16:21 <MarkSmith> We will try to get to them.  If we start working, we will take ownership of the ticket.
19:17:08 <MarkSmith> Who's next?
19:17:10 * boklm can go next
19:17:25 <boklm> Last week I merged and updated the searchengine test for #11236
19:17:30 <boklm> I updated the testsuite to find version-buildN directories on people.tpo
19:17:38 <boklm> I added some test pages for the noscript globalHttpsWhitelist option: https://trac.torproject.org/projects/tor/ticket/13053#comment:8
19:17:49 <boklm> While testing this option I noticed that http URLs sourced from https pages are not blocked when globalHttpsWhitelist and cascadePermissions are true
19:18:22 <boklm> This week I'm planning to fix #14992, #14959 and continue adding tests for #13053
19:19:00 <boklm> that's all for me
19:20:20 <mikeperry> ah, yes. we ran into #14992 and related issues during this build
19:20:59 <toralf> git bisect blames 10ae9b9bf58822 to be a bad commit to break test cases
19:22:57 <mikeperry> do we have an arthuredelstein? or support people?
19:26:12 <mikeperry> hrmm. well, we should discuss MAR and release signing then, I suppose
19:26:52 <msvb-lab> nsTransferable landed in mozilla-central, in time for ESR38. Can I ask a question?
19:27:27 <GeKo> sure
19:27:42 <mikeperry> GeKo: do we have documentation for doing that yet? I can try to offline sign everything, but having docs on excactly what tools I need might make it easier
19:27:56 <GeKo> yes, we have
19:28:13 <GeKo> see the README.build
19:28:23 <GeKo> there is the explanation of the signmars.sh script
19:28:32 <GeKo> I tried to be verbose
19:28:51 <GeKo> if something is unclear I think we should fix that there
19:30:31 <GeKo> you need the signmar tool as well (which gets built during release)
19:30:56 <mikeperry> ok, I can give that a shot
19:32:10 <GeKo> MARTOOLS_ZIP="$WRAPPER_DIR/../../gitian-builder/inputs/mar-tools-${OSNAME}.zip"
19:32:33 <GeKo> is relevant here, too. I had to put that relative to my signing dir in order to get it working
19:32:55 <mikeperry> hrmm.. yeah, does it need to be per-OS?
19:33:19 <mikeperry> sounds like I'm going to need to transfer a huge directory tree in order to do this offline
19:34:07 <GeKo> well, I just have the gitian-builder/inputs/mar-tools* on my offline machine
19:34:18 <boklm> $OSNAME is probably always linux64
19:34:58 <GeKo> and then I do the signing in tor-browser-bundle/gitian
19:36:12 <mikeperry> ah, ok
19:38:29 <mikeperry> ok, well anything else for the meeting today then?
19:38:44 <GeKo> mikeperry: looking at the january report mail we switch this time to the new key for the stable releases, too, right?
19:38:51 <GeKo> Or did I get that wrong again?
19:38:58 <GeKo> *release
19:39:32 <arthuredelstein> Hi all, sorry I'm 40 minutes late!
19:39:39 <GeKo> hi
19:40:43 <msvb-lab> mikeperry: Just want to close out nsTransferable before end of meeting if that's okay.
19:40:45 * arthuredelstein is typing report now...
19:41:59 <mikeperry> GeKo: hrm.. I think yes, though we should probably update https://www.torproject.org/docs/verifying-signatures.html.en and https://www.torproject.org/docs/signing-keys.html.en
19:42:49 <GeKo> the latter has the new key
19:42:57 <GeKo> I can update the former tomorrow
19:44:33 <msvb-lab> Alright, so the logic to correct #9701 has been in for a while, and Mark's idea to include a mochitest became the unimplemented #14255.
19:44:41 <msvb-lab> However, both tickets (including mochitest) are implemented and landed at Mozilla [1] now.
19:44:43 <msvb-lab> [1] https://bugzilla.mozilla.org/show_bug.cgi?id=1123480
19:44:47 <msvb-lab> The question is if we want to copy the Mozilla mochitest to the Tor trunk as well.
19:44:53 <msvb-lab> Since I wrote a chrome test rather than HTML/browser test it might not integrate that well, I'm not sure.
19:45:14 <msvb-lab> Any opinions? Like relating to release engineering/alpha builds on ESR38?
19:45:56 <arthuredelstein> msvb-lab: We do have a few mochi-browser tests already.
19:46:05 <arthuredelstein> So I don't see any reason why not
19:46:24 <msvb-lab> arthuredelstein: Yes, but I think they're not testing components, and don't use the chrome window.
19:46:27 <msvb-lab> Or do they?
19:46:45 <mikeperry> I am inclined to postpone updating the tests until we start the ESR38 switch, which hopefully we will begin right after 4.5 becomes stable in mid-april (since we'll have a FF38 beta to work off of by then)
19:47:26 <msvb-lab> mikeperry: If that's the conclusion, then it would make sense to just not copy the tests over... since they'll be integrated implicitly later.
19:48:14 <msvb-lab> I might not get to it until April anyway.
19:48:22 <msvb-lab> Okay, thanks that answered my question. Over and out.
19:49:23 * arthuredelstein can go now
19:49:33 <arthuredelstein> Last week I worked on #14429, #14555 and #14956. This week I'll mostly be afk, but the following week I'm hoping to finish #13670 and upstream some fingerprinting patches to Mozilla.
19:49:56 <arthuredelstein> That's it for me.
19:50:44 <GeKo> thanks again for looking so quickly at the HTTPS-E I was afraid we'd need some additional last minute fix + a rebuild.
19:51:02 <GeKo> s/HTTPS-E/HTTPS-E bug/
19:51:26 <arthuredelstein> no problem!
19:54:50 <mikeperry> ok, anything else then? I am not sure if trying to have the IRC meeting next monday during the dev meeting will work out or not
19:55:22 <mikeperry> likely not, as it will be after hours and we may be ajourning to dinner and/or kicked out of the space
19:55:32 <toralf> nickm: git bisect blames 10ae9b9bf58822 to be a bad commit to break test cases
19:56:49 <chema9355> i need some help with chutney does anybody control it?
19:56:52 * MarkSmith will tentatively plan on "no meeting" but will try to be on IRC a lot.
19:57:18 * arthuredelstein as well
19:57:58 <mikeperry> ok, sounds good
19:58:12 <mikeperry> we may want to ask things on IRC as we do roadmapping, etc
19:58:21 <MarkSmith> right
19:59:41 <mikeperry> I think that wraps it up for today then at just under an hour. thanks everyone!
19:59:50 <mikeperry> #endmeeting *baf*