-queuebot:#ubuntu-release- New binary: calamares-settings-ubuntu [i386] (cosmic-proposed/universe) [17] (no packageset) | 01:20 | |
-queuebot:#ubuntu-release- New binary: calamares-settings-ubuntu [s390x] (cosmic-proposed/universe) [17] (no packageset) | 01:20 | |
-queuebot:#ubuntu-release- New binary: calamares-settings-ubuntu [ppc64el] (cosmic-proposed/universe) [17] (no packageset) | 01:20 | |
-queuebot:#ubuntu-release- New binary: calamares-settings-ubuntu [amd64] (cosmic-proposed/universe) [17] (no packageset) | 01:21 | |
-queuebot:#ubuntu-release- New binary: calamares-settings-ubuntu [armhf] (cosmic-proposed/universe) [17] (no packageset) | 01:21 | |
-queuebot:#ubuntu-release- New binary: calamares-settings-ubuntu [arm64] (cosmic-proposed/universe) [17] (no packageset) | 01:21 | |
tsimonq2 | slangasek: ^ can I get approval so I get it rolling in the dailies? :) | 01:24 |
---|---|---|
=== tacobat` is now known as tacobat | ||
=== tacocat` is now known as tacocat | ||
-queuebot:#ubuntu-release- New: accepted android-platform-external-boringssl [armhf] (cosmic-proposed) [8.1.0+r23-1] | 03:56 | |
-queuebot:#ubuntu-release- New: accepted android-platform-external-boringssl [amd64] (cosmic-proposed) [8.1.0+r23-1] | 03:57 | |
-queuebot:#ubuntu-release- New: accepted android-platform-external-boringssl [i386] (cosmic-proposed) [8.1.0+r23-1] | 03:57 | |
-queuebot:#ubuntu-release- New: accepted android-platform-external-boringssl [arm64] (cosmic-proposed) [8.1.0+r23-1] | 03:57 | |
-queuebot:#ubuntu-release- New: accepted android-platform-external-boringssl [ppc64el] (cosmic-proposed) [8.1.0+r23-1] | 03:57 | |
-queuebot:#ubuntu-release- New: accepted hvac [source] (cosmic-proposed) [0.5.0-0ubuntu1] | 05:52 | |
-queuebot:#ubuntu-release- New binary: hvac [amd64] (cosmic-proposed/none) [0.5.0-0ubuntu1] (no packageset) | 05:54 | |
LocutusOfBorg | tsimonq2, bad ffmpeg upload is bad :D | 06:16 |
* LocutusOfBorg exports the variable and reuploads :p | 06:17 | |
acheronuk | ah, that's why the ppc64el autotest still fails! | 06:18 |
tsimonq2 | LocutusOfBorg: baaaaaah. | 06:19 |
* tsimonq2 blames Laney :P | 06:20 | |
LocutusOfBorg | yep, I see the export does actually pass the -O2 to the build (on DOM) | 06:20 |
LocutusOfBorg | lets see if we got it right :) | 06:20 |
* tsimonq2 sends giraffes at Laney as a present. | 06:20 | |
LocutusOfBorg | we hate giraffes! | 06:20 |
tsimonq2 | Exactly. :D | 06:21 |
LocutusOfBorg | but not as much as horses I would say | 06:21 |
LocutusOfBorg | horses are evil | 06:21 |
acheronuk | ???? | 06:21 |
acheronuk | oh. ic | 06:22 |
* LocutusOfBorg finds the video | 06:22 | |
LocutusOfBorg | https://www.youtube.com/watch?v=ojgrEGmDSCg | 06:23 |
LocutusOfBorg | this one! | 06:23 |
tsimonq2 | (I'm blaming Laney because I blindly copied what he did in #ubuntu-devel. :P) | 06:23 |
tsimonq2 | LocutusOfBorg: hahaha | 06:23 |
LocutusOfBorg | tsimonq2, I see from build log that it might be not enough, because it passes -O2, but there is another -O3 passed by dpkg itself | 06:25 |
LocutusOfBorg | it might win the latter, so -O3 | 06:25 |
tsimonq2 | LocutusOfBorg: Please follow up on the Debian bug then, if that's not the case. | 06:26 |
acheronuk | LocutusOfBorg: seems a fair amount of logic in source configure to set different optimisations | 06:43 |
acheronuk | depending on the env | 06:43 |
acheronuk | and --optflags=OPTFLAGS override optimization-related compiler flags | 06:48 |
tsimonq2 | Man, where is this all documented? | 06:48 |
tsimonq2 | This is slightly confusing to me; even as a MOTU. | 06:48 |
tsimonq2 | LocutusOfBorg: Bad LocutusOfBorg is bad and doesn't link docs. :D | 06:49 |
acheronuk | that is in configure in the source | 06:49 |
tsimonq2 | Anyway, I should try to 😴 o/ | 06:49 |
tsimonq2 | acheronuk: Right; confusing. ;) | 06:50 |
acheronuk | it is if you depend on the BS and never run ./configure --help yourself! | 06:51 |
tsimonq2 | hahahaha | 06:51 |
LocutusOfBorg | bad tsimonq2 exporting -O2 everywhere! :) | 07:02 |
tsimonq2 | 😆 | 07:03 |
LocutusOfBorg | acheronuk, I was looking at "--extra-cxxflags=ECFLAGS" instead | 07:04 |
acheronuk | LocutusOfBorg: I only skimmed it, so could be | 07:05 |
LocutusOfBorg | I'm trying to get the order of all that stuff, in the meanwhile I uploaded a build on my ppa | 07:05 |
LocutusOfBorg | https://launchpad.net/~costamagnagianfranco/+archive/ubuntu/locutusofborg-ppa/+sourcepub/9325527/+listing-archive-extra | 07:05 |
acheronuk | I still see -O3 after -O2 in the logs there | 07:07 |
LocutusOfBorg | acheronuk, https://trac.ffmpeg.org/ticket/2816 | 07:10 |
LocutusOfBorg | you are my best bet | 07:10 |
LocutusOfBorg | acheronuk, that build was to see if only ppc64el was using the code | 07:11 |
LocutusOfBorg | https://launchpad.net/~costamagnagianfranco/+archive/ubuntu/costamagnagianfranco-ppa/+packages | 07:11 |
LocutusOfBorg | this one uses your approach | 07:11 |
LocutusOfBorg | I grepped math-errno and it is indeed passed by the configure script, not something that comes from gcc or dpkg | 07:12 |
LocutusOfBorg | and the -O3 we want to override is added in the same chunk of configure code | 07:12 |
acheronuk | makes sense | 07:14 |
acheronuk | ambiguous configure options. sigh | 07:21 |
acheronuk | google shows me a fair few examples where people have -O2 and other options passed via --optflags. whether it effective is another matter | 07:25 |
LocutusOfBorg | yep | 07:29 |
acheronuk | optflags gives me 'CFLAGS= blah blah blah -O2' in ffbuild/config.mak after local configure, vs -O3 without, so seems to be effective there | 07:33 |
LocutusOfBorg | doesnt work in ppa | 07:35 |
LocutusOfBorg | https://launchpadlibrarian.net/382847674/buildlog_ubuntu-cosmic-ppc64el.ffmpeg_7%3A4.0.2-1ubuntu3_BUILDING.txt.gz | 07:35 |
acheronuk | sigh | 07:35 |
LocutusOfBorg | does it look ok to you= | 07:35 |
LocutusOfBorg | ? | 07:35 |
acheronuk | it appears not ok | 07:37 |
LocutusOfBorg | but the configure line? | 07:38 |
acheronuk | looks fine compared to what worked locally here | 07:41 |
LocutusOfBorg | lol acheronuk it wasn't | 07:50 |
LocutusOfBorg | ahahaha got it | 07:50 |
LocutusOfBorg | CONFIG += --optflags='-O2' | 07:50 |
LocutusOfBorg | CONFIG += "--optflags='-O2'" | 07:50 |
LocutusOfBorg | this was the difference | 07:50 |
LocutusOfBorg | and now it works | 07:50 |
LocutusOfBorg | bad configure is bad | 07:50 |
LocutusOfBorg | https://launchpad.net/~costamagnagianfranco/+archive/ubuntu/costamagnagianfranco-ppa/+build/15254125 | 07:51 |
acheronuk | lol. not sure how I missed the " " | 07:53 |
acheronuk | looking to hard at the arg, and not what was around it I guess! | 07:53 |
acheronuk | LocutusOfBorg: ooh. you uploaded. fingers crossed then | 08:04 |
acheronuk | on the tests | 08:05 |
acheronuk | it bugs me that on each gcc invocation you still see -O3, but followed by -O2 :/ | 08:09 |
acheronuk | oh, but that is a just a reversal of what happened without optflags, so just how this one rocks I guess | 08:15 |
LocutusOfBorg | arm64 what about failing to build? | 08:45 |
acheronuk | oh ffs | 09:00 |
acheronuk | LocutusOfBorg: retried and succeeded? | 09:26 |
LocutusOfBorg | ssssshhhh we didn't see it failing, ever! :) | 09:30 |
ginggs | LocutusOfBorg: are you going to try setting -O2 for ghostscript? | 09:41 |
acheronuk | LocutusOfBorg: autotest still fails: https://autopkgtest.ubuntu.com/packages/ffmpeg/cosmic/ppc64el | 10:13 |
ginggs | i tried in my PPA twice https://launchpad.net/~ginggs/+archive/ubuntu/testing/+sourcepub/9320256/+listing-archive-extra | 10:15 |
ginggs | what I didn't try was export DEB_CFLAGS_MAINT_STRIP=-O3 / export DEB_CFLAGS_MAINT_APPEND=-O2 together with CONFIG += --optflags=-O2 | 10:16 |
ginggs | i.e. maybe we need to override the -O3 that ubuntu put in, as well as what configure puts in | 10:17 |
LocutusOfBorg | ginggs, I'm leaving shortly | 10:21 |
LocutusOfBorg | lets try a new ffmpeg spin then | 10:25 |
LocutusOfBorg | ginggs, sorry I didn't get your sentence above, did that approach work or not? | 10:26 |
ginggs | i tried DEB_CFLAGS_MAINT_STRIP=-O3 / export DEB_CFLAGS_MAINT_APPEND=-O2 in ~ppa1 | 10:27 |
ginggs | then CONFIG += --optflags=-O2 in ~ppa2 | 10:27 |
ginggs | i haven't tried both together | 10:27 |
ginggs | i can try that in my PPA now, at least it might make it easier to spot in the log if -O3 creeps in from elsewhere | 10:28 |
ginggs | building now https://launchpad.net/~ginggs/+archive/ubuntu/testing/+sourcepub/9325607/+listing-archive-extra | 10:35 |
LocutusOfBorg | I did fix it | 10:37 |
LocutusOfBorg | objdump shows good code with -O3 and -fno-wrapv on ppc64el pbuilder chroot | 10:38 |
LocutusOfBorg | I have to leave, uploaded | 10:57 |
LocutusOfBorg | fingers crossed! | 10:58 |
ginggs | LocutusOfBorg: it worked! http://autopkgtest.ubuntu.com/packages/f/ffmpeg/cosmic/ppc64el | 11:53 |
ginggs | tkamppeter: ghostscript uploaded. you almost had it, the DEB_CFLAGS_MAINT bit had to come before include /usr/share/cdbs/1/rules/debhelper.mk | 11:58 |
acheronuk | aaaah | 12:01 |
acheronuk | ginggs: thank you | 12:03 |
-queuebot:#ubuntu-release- Unapproved: gnome-shell-extension-system-monitor (bionic-proposed/universe) [35-1 => 35-1ubuntu0.18.04.1] (no packageset) | 12:12 | |
Laney | tsimonq2: I said export in there. | 13:01 |
Laney | and it would have been good to get it all passing in a PPA instead of using the real archive to iterate | 13:02 |
Laney | anyway, hope it fixes it, thx for working on that | 13:02 |
ginggs | running the autopkgtests in a PPA is quicker too... | 13:07 |
ginggs | if you can get the URLs right https://wiki.ubuntu.com/ProposedMigration#Testing_against_a_PPA | 13:08 |
Laney | ginggs: what was wrong with DEB_CFLAGS_MAINT_APPEND in ghostscript btw? I just uploaded with that and it built on ppc64el OK https://launchpad.net/~laney/+archive/ubuntu/ppa/+build/15254447 https://paste.ubuntu.com/p/M6nrXF5ybB/ | 13:22 |
ginggs | [13:58:22] <ginggs> tkamppeter: ghostscript uploaded. you almost had it, the DEB_CFLAGS_MAINT bit had to come before include /usr/share/cdbs/1/rules/debhelper.mk | 13:23 |
ginggs | Laney: ^ | 13:23 |
ginggs | cdbs... shudder | 13:24 |
Laney | ginggs: right, but I'm talking about using STRIP/PREPEND which is a bit unusual | 13:24 |
ginggs | Laney: DEB_CFLAGS_MAINT_APPEND probably would have worked, i prefer STRIP/PREPEND as it makes the logs easier to read | 13:25 |
slangasek | ddstreet: it seems your nodejs SRU FTBFS on all archs, as noted by ehashman on the bug. can you follow through on this? Specifically, parallel/test-tls-server-verify test fails | 17:25 |
tsimonq2 | Laney: :P | 19:06 |
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!