* stgraber waves | 15:57 | |
mdeslaur | o/ | 15:58 |
---|---|---|
* slangasek waves | 15:58 | |
infinity | \o | 15:59 |
mdeslaur | looks like kees is MIA | 16:00 |
mdeslaur | let's get started | 16:00 |
mdeslaur | #startmeeting | 16:00 |
meetingology | Meeting started Tue May 23 16:00:05 2017 UTC. The chair is mdeslaur. Information about MeetBot at http://wiki.ubuntu.com/meetingology. | 16:00 |
meetingology | Available commands: action commands idea info link nick | 16:00 |
mdeslaur | [topic] Action review | 16:00 |
mdeslaur | infinity, slangasek: any progress on your items? | 16:00 |
slangasek | none | 16:00 |
slangasek | but my guilt levels are rising appropriately | 16:01 |
infinity | As usual, I reminded myself of action items 5 minutes before the meeting. I've pinged roaksoax to get the MaaS thing finally sorted. | 16:01 |
mdeslaur | heh, ok | 16:01 |
infinity | The other item, I did a ninja wiki edit of s/16.04.2/16.04.3/ :P | 16:01 |
slangasek | heh | 16:01 |
mdeslaur | infinity: perhaps just bump it to 16.04.5 right away? :) | 16:01 |
infinity | Shush you. | 16:01 |
mdeslaur | hehe | 16:01 |
mdeslaur | #topic Mailing list archive | 16:01 |
mdeslaur | doesn't look like there's anything new | 16:02 |
mdeslaur | #topic Community bugs | 16:02 |
mdeslaur | no open bugs | 16:02 |
mdeslaur | #topic AOB | 16:02 |
mdeslaur | anyone have anything to discuss? | 16:02 |
slangasek | not I | 16:03 |
stgraber | nothing here | 16:03 |
infinity | Can we drop Go from the archive? | 16:03 |
infinity | If not, then I have nothing to discuss. :P | 16:03 |
mdeslaur | heh | 16:03 |
mdeslaur | #topic Next chair | 16:04 |
mdeslaur | slangasek with stgraber as backup | 16:04 |
* slangasek nods | 16:04 | |
mdeslaur | #endmeeting | 16:04 |
meetingology | Meeting ended Tue May 23 16:04:37 2017 UTC. | 16:04 |
meetingology | Minutes: http://ubottu.com/meetingology/logs/ubuntu-meeting-2/2017/ubuntu-meeting-2.2017-05-23-16.00.moin.txt | 16:04 |
mdeslaur | thanks! | 16:04 |
slangasek | kees: LATE | 16:04 |
infinity | kees: You missed it. | 16:04 |
kees | whoops, late | 16:04 |
kees | haha | 16:04 |
slangasek | :) | 16:04 |
kees | saw the "thanks" and started laughing | 16:04 |
slangasek | fwiw in 2 weeks time our team is sprinting, so I may be an unreliable meeting narrator | 16:05 |
infinity | kees: You have AOB before we all /part? | 16:05 |
kees | infinity: nope, all good. | 16:05 |
mdeslaur | slangasek: I'll handle it if you don't show up | 16:05 |
infinity | kees: You might be happy to know that we're finally PIE on all arches. | 16:06 |
slangasek | mdeslaur: isn't stgraber supposed to? :) | 16:06 |
kees | infinity: \o/ \o/ | 16:06 |
slangasek | infinity: has that flip been switched? | 16:06 |
slangasek | (and I never got an answer to my email asking who would coordinate rebuilds) | 16:06 |
infinity | slangasek: Yeah, doko got all keen and did it on Saturday, I followed up with a dpkg so the world didn't explode. | 16:06 |
kees | https://outflux.net/ubuntu/hardening/ubuntu-hardening-5years.png | 16:06 |
slangasek | ok | 16:06 |
stgraber | slangasek: I'll be sprinting too :) | 16:07 |
mdeslaur | slangasek: won't he be at the same sprint? | 16:07 |
slangasek | mdeslaur: not the same one no | 16:07 |
mdeslaur | oh, heh | 16:07 |
stgraber | different sprints, same week | 16:07 |
mdeslaur | well, we'll figure it out | 16:07 |
stgraber | it was supposed to be the same sprint originally, but that didn't quite happen | 16:07 |
kees | I changed how my PIE-analysis tool credited PIE packages, so there was quite a bump recently (if package is all .so's and no binaries, count as PIE. If a single ET_EXEC exists, no PIE for you) | 16:07 |
slangasek | yeah, conference bookings went a bit sideways across the reorg | 16:08 |
infinity | stgraber: Sadness. I don't think we've hung out in a long while. | 16:08 |
slangasek | kees: ahh ok | 16:08 |
stgraber | infinity: been a while indeed, Seville was the last one I think | 16:08 |
slangasek | I was going to say, I know I didn't have *that* many in my mass-rebuild | 16:08 |
kees | slangasek: this is only amd64, so it'll likely not show much change | 16:09 |
slangasek | kees: I mean the mass-rebuild we did right around 17.04 release | 16:09 |
slangasek | which is probably the second smaller tick on your graph | 16:09 |
kees | oh! that likely did help, yes. | 16:09 |
slangasek | (main-only) | 16:09 |
infinity | kees: Any urge to run stats for i386, so we can see where we need to apply abuse? | 16:10 |
infinity | (or armhf or arm64, they should all be in the same state) | 16:10 |
kees | I would suspect the two steps here are rebuild and script-adjust: https://outflux.net/ubuntu/hardening/ubuntu-hardening-month.png | 16:10 |
kees | infinity: In theory I should be able to do i386, just requires some script-bashing | 16:11 |
infinity | slangasek: I admit to being somewhat ignorant about the bits that need staged rebuilds. Something to do with static linked bits? | 16:12 |
slangasek | or some bash scripting? | 16:12 |
kees | https://outflux.net/ubuntu/hardening/main/ had some unreliable data that I didn't try fixing | 16:12 |
slangasek | infinity: yes | 16:12 |
infinity | I don't think I was involved in that part for the previous arches. | 16:12 |
kees | but the _current_ result is accurate, I think, for amd64 main | 16:12 |
slangasek | infinity: basically, when we turned on PIE, we had a long tail of opaque build failures that required tracking down and rebuilding static libs in order to fix | 16:12 |
kees | i.e. 96% PIE | 16:12 |
slangasek | and nobody really took systematic notes on this | 16:13 |
infinity | kees: Nice. I, too, am 96% pie. | 16:13 |
slangasek | hahaha | 16:13 |
kees | haha | 16:13 |
slangasek | so while we could just say "if it has a static lib that's not PIE, rebuild it", that's going to hit pretty much every library package in the archive instead of the ones that actually have static revdeps | 16:14 |
infinity | slangasek: So, I guess maybe the next step, if we have no notes, is a rebuild test that we can unwind. | 16:14 |
slangasek | infinity: I was going to suggest doing archive analysis for -dev build-deps that don't translate into runtime deps | 16:15 |
slangasek | should we take this to #-devel? :) | 16:15 |
infinity | slangasek: Sure. | 16:18 |
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!