IRC logs for #baserock for Friday, 2015-11-20

*** rdale has quit IRC01:50
*** gtristan has quit IRC06:09
*** gtristan has joined #baserock06:40
*** fay_ has joined #baserock08:01
*** toscalix has joined #baserock08:10
straycatoh dear08:27
straycatmason failed with since the llvm upgrade we no longer build shared libs08:28
straycatlooking at the llvm docs i see that shared libs aren't enabled, so perhaps it would be better to compile mesa with --disable-llvm-shared-libs08:28
straycatsorry, i mean to say that "shared libs aren't recommended"08:29
*** bruce_ has joined #baserock08:31
*** toscalix has quit IRC08:45
*** toscalix has joined #baserock08:45
*** rdale has joined #baserock09:06
pedroalvarezyeah, I saw that error yesterday night :/09:08
pedroalvarezI want pre-merge CI more than ever :)09:09
straycati've patched mesa to use --disabled-llvm-shared-libs and it compiles, just waiting for the rest of the build09:10
* pedroalvarez has no idea about graphics09:11
straycatit shouldn't make a difference, the libraries will just be statically linked instead09:12
pedroalvarezdo you want me to test graphics with that patch?09:12
straycatis that a lot of work?09:14
radiofreellvm is only used for the swrast stuff, so as long as weston-simple-egl works in a vm that's a good enough test09:15
*** Lachlan1975 has joined #baserock09:17
straycatgenivi build finished with
*** tiagogomes has joined #baserock09:28
* pedroalvarez tests09:37
*** Lachlan1975 has quit IRC09:37
*** jonathanmaw has joined #baserock09:40
*** locallycompact has joined #baserock09:50
* gtristan looks at audio-bluetooth... and considers that pulseaudio needs x11 and GTK+... and groans09:56
* gtristan facerock09:57
straycathrm, 232 commits in definitions since the last release10:01
*** ssam2 has joined #baserock10:06
*** ChanServ sets mode: +v ssam210:06
*** DavePage has left #baserock10:08
straycatgtristan, for the release notes, it would be cool to write about all the gnome stuff, do we pretty much have a basic gnome system now?10:10
gtristanstraycat, I'm struggling a little to find the definition of 'basic gnome system', maybe I strive too much for perfection I'm not sure10:12
straycatthe vagueness there is mostly due to my lack of gnome knowledge, i see there's a gnome system in definitions and i seem to remember you saying that gdm, gnome-shell, gnome-terminal are there?10:13
gtristanstraycat, basically - in my local branch goa is working very well, most of the platform is (albeit I think I need to pay attention to pulseaudio)10:14
straycatokay that's awesome then10:14
gtristanexcept that the last 2 days I've been struggling to make the initial setup work... it works except that the credentials for online accounts are lost in the handoff from the gnome-initial-setup user to the user it creates10:15
gtristanwhich is a delicate dance to begin with10:15
gtristanother than that, if you re-authorize your online accounts, next login the keyring automatically unlocks (and updates when you change the password and such)10:16
gtristanand then say, if you have your google account there, gnome-shell's calandar will show your events10:16
gtristanwhich is nice, online desktop UX pretty much works except for initial setup10:16
gtristanalso localization & input methods work very well10:17
* gtristan still has to add those chinese fonts10:17
pedroalvarezoh!  for the notes, LOCALES!10:17
gtristanto be honest this system will be the first I know of where asian input methods work out of the box10:17
straycatwe also seem to have put the xfce system into the ci, which is cool10:19
gtristansystem still lacks apps of course, but that is really the easy part10:19
gtristanI havent tested audio at all though, looking at this baby: ...10:19
gtristanlooks like we have some traveling before audio really works as expected10:19
pedroalvarezstraycat: weston-simple-egl works, we are good to go10:22
ssam2release!!! this is exciting10:32
*** locallycompact has quit IRC10:33
*** locallycompact has joined #baserock11:17
straycatpedroalvarez, i've pushed the release branch11:18
pedroalvarezgood :)11:19
*** locallycompact has quit IRC11:27
*** locallycompact has joined #baserock11:28
straycatdraft release notes are at let me know if you think i've missed anything or if i've made them to informal/excitable this time.11:45
pedroalvarezThey look good to me :)11:47
straycatthere is a point at the end about artifacts11:50
straycatand it mentions x86 only, but we build jetson arm artifacts now do we not?11:50
pedroalvarezwe do, but the mason instance that is doing that is not public, so nobody can check that it is still providing artifacts11:51
pedroalvarezs/is doing that/is creating arm artifacts/11:51
straycatokay, because iirc we used to upload x86 and arm artifacts to, so that all releases were just fetched, without even needing to set artifact-cache-server11:53
straycati guess we still do that, and that it's just users of that will have x86-only, the real problem being here that we can't specify a list of cache servers, i guess11:54
pedroalvarezhaving 2 cache artifacts was a bit confusing in my opinion11:54
pedroalvarezI stopped uploading artifacts to g.b.o to be honest11:55
straycatoh, okay11:55
* rjek puts in a request for a tarball of the ARM file system to be readily available11:55
straycatwell i guess for most people this makes no difference, so...11:55
straycatrjek, that's a standard part of the process so it should be there11:56
jmacs_My devel system seems to be having trouble with SSL certificates. Do I need to do anything special to install root CAs on a recent baserock build?11:59
*** jmacs_ is now known as jmacs11:59
straycatjmacs, what trouble exactly?11:59
jmacs"Could not fetch URL connection error: [SSL: CERTIFICATE_VERIFY_FAILED] certificate verify failed (_ssl.c:590) - skipping"11:59
straycatokay you should be able to fix that by upgrading12:00
straycatca-certs got broken by the move to python3, which got fixed quite recently12:00
jmacsAh, ok, I'm on a Nov 8 definitions12:00
rjekstraycat: Ah, I thought on prior releases only ARM images made it12:01
jjardonstraycat: about "a tarball of the ARM file system to be readily available"; its new in this release, so maybe its worth a note12:01
straycatjmacs, 4b9f6341f0d0a890c6a112a1db077506a86f4fb8 should be the commit that fixes that12:02
pedroalvarezthanks for mentioning the arm rootfs, I was forgeting about it12:02
straycatrjek, sorry i misread, pedroalvarez seems to know about it anyway12:03
straycatjjardon, okay i'll add a note12:04
ssam2straycat: release notes look fine12:13
ssam2except, I'd prefer if the 'definitions compatibility' section matched previous ones12:13
ssam2it's quite a complex thing to reason about so I think saying "Baserock references systems now use version 6." alone isn't so helpful12:14
ssam2in it's a bit clearer what this all means for the user, I thikn12:15
*** bwh_ has quit IRC12:17
*** gtristan has quit IRC12:17
*** bwh has joined #baserock12:18
*** bwh has quit IRC12:21
*** bwh has joined #baserock12:22
straycatssam2, okay12:30
straycatfixed and added a note about arm rootfs12:36
straycatpedroalvarez, alright mason failed for some reason12:40
straycatsome error running strip commands when building mesa12:40
*** toscalix has quit IRC12:42
jjardonI've never seen that before12:49
straycati don't really know what's going on there, i've built that system on x86-64 on my machine here12:52
ssam2straycat: I meant all of the text under 'definitions compatibility, shall I add the rest there?12:53
straycatoh i see, no i can do that12:54
pedroalvarezMason is fine, it only failed because of a transfer error12:59
straycatok cool13:01
straycatssam2, should be fixed now?13:01
ssam2perfect, thanks13:03
*** gtristan has joined #baserock13:17
pedroalvarezsystems are being deployed right now :)13:32
*** toscalix has joined #baserock13:46
jmacsIf we were to start up the Baserock deterministic builds thing properly, how difficult would it be to get a web server at "" I could push pages to?13:54
persiapedroalvarez ssam2 gary_perkins_  franred ^^14:00
*** persia_ has quit IRC14:03
*** persia_ has joined #baserock14:03
jmacsHmm, clang does take a while to build14:28
rjekWhat you need is a faster compiler14:29
* straycat thinks we'll have improvements to the format solve that problem14:30
franredjmacs, what do you want to push? cache artifacts? definitions? documentation?14:35
jmacsBuild reports, which will be static HTML + a bit of CSS mainly14:35
franredI fear that ssam2 and pedroalvarez will be your best bullets then14:38
persiajmacs: At a high level, it means putting a description of such a server in the infrastructure definitions repo, and then deploying it, just to provide a bit of frame.14:39
straycatpedroalvarez, what's the state of the builds?14:46
pedroalvarezstraycat: the last image is being compressed14:47
straycatawesome :)14:47
pedroalvarezsorry I can't help now jmacs, but if you manage to set up a server somewhere, you can ask gary_perkins_ to create the DNS entry14:47
*** gary_perkins_ is now known as gary_perkins14:48
*** gary_perkins has joined #baserock14:49
jmacspedroalvarez: No problem, it's not urgent.14:57
jmacsI have my own web server which I'm using at the moment - if you have sufficient trust in me to point a subdomain at it, then that would work fine.14:57
pedroalvarezI'm ok with that14:59
pedroalvarezstraycat: once they are uploaded, do you want me to update the symlinks in
straycatpedroalvarez, that would be good, shall i tag this now?15:02
pedroalvarezyes :)15:02
pedroalvarezOk. that's done15:04
jmacsgary_perkins: If you're ok with that DNS entry, my server is
gary_perkinsjmacs: yep, no problem15:14
jmacsThere's just a placeholder there at the moment (if I've set up apache properly)15:14
gary_perkinsjmacs: ok, done. May take a bit of time to propagate, of course15:15
jmacsOf course.15:16
gary_perkinsalready! has address
jmacsOh, it kind of works. Although my HTML is broken.15:18
jmacsThanks gary_perkins15:18
radiofreethe table looks lovely jmacs15:38
rjekThat's pretty shiny15:39
jmacsIt'd be good to have something actually related to Baserock on there. I'm working on that now.15:42
tiagogomeswhat is the name of the thingy that provides a git API ? is it morph-cache-server?15:56
ssam2what do you mean by git API ?15:57
ssam2morph-cache-server is
ssam2it adds a couple of HTTP methods that expose some info about cached git repos15:58
tiagogomesfor example to resolve the tree sha1 for a given commit sha1 for certain repoi15:58
ssam2right, that's morph-cache-server15:58
ssam2that lets you do it over HTTP15:58
straycat15.47 is released!16:09
straycatthanks to pedroalvarez mostly i think16:09
* straycat runs to catch a train he will probably miss16:09
straycatssam2, i don't know whether the mail i sent to announce will need moderation or not, possibly16:11
* straycat disappears16:11
*** bruce_ has quit IRC16:22
* tiagogomes would like that more morph changes were mentioned in the release notes16:24
*** ssam2 has quit IRC16:27
tiagogomesmmm, running morph without workspaces causes build branches to persist16:58
persiatiagogomes: Hrm?  How do you mean?17:00
persiaI would have thought that not having workspaces helped reduce spurious persisting branches (or at least, that was my experience)17:01
tiagogomespersia morph creates a temporary branch when it does a build. It is supposed to delete that branch when the build finishes or fails17:02
persiaI thought morph stopped creating those temporary branches when workspaces were dropped.17:02
persiaAt least, that was the idea (and part of the point)17:02
tiagogomesno, a temporary branch is created on definitions only unless you pass --local-changes=ignore17:05
persiaOh, it might be a workflow thing.17:07
persiaFor the non-workspace model, I think the recommendation was to build against a clean commit of definitions17:07
persia(with the idea that one can `git commit --append` if adjustment is required)17:08
persiaThe temporary branches are just to ensure the reflog contains something that allows one to recover the results of the build if one likes it.17:08
persia(to make up for not having that be stored in the workspace)17:08
tiagogomesmmm. it could become annoying to the user if he has to amend a commit several times in a short time frame17:09
persiaDepends on the user.  My personal git workflow involves running git commit --amend before `make` every time.17:16
persiaJust so I have a reflog of everything that happened, and can get back to an earlier state within the commit if I didn't like what I did.17:16
persiaAnd yet I only have one commit to submit, avoiding any merge commits, as my commits can be fast-forwarded safely.17:17
jjardonjmacs: maybe a good idea to add a link to in ?17:17
jmacsI'm going through the baserock page at the moment; I'll be updating it17:17
jmacsCurrently does not have any baserock-related content on it17:18
jjardonjmacs: nice, great work btw!17:18
jmacsThank you17:19
jmacsI should patch 'df' so it says "Check system-version-manager list" in flashing letters17:26
*** locallycompact has quit IRC17:50
*** jonathanmaw has quit IRC17:57
*** rdale has quit IRC17:59
straycat16:24 ! tiagogomes would like that more morph changes were mentioned in the release notes18:02
straycattiagogomes, perhaps you should have said something in the 3 or so hours i gave for people to review the notes?18:02
*** toscalix has quit IRC18:09
tiagogomessorry, I missed the link18:11
straycatapology accepted :)18:13
*** fay_ has quit IRC18:29
*** edcragg has quit IRC18:50
*** franred has quit IRC19:05
*** tiagogomes has quit IRC19:27
*** edcragg has joined #baserock22:42

Generated by 2.15.3 by Marius Gedminas - find it at!