/srv/irclogs.ubuntu.com/2010/04/24/#ubuntu-mozillateam.txt

asacfta: there?12:06
ftaasac, yep but not for long12:06
asacfta: ho ;) ... great12:07
asacfta: so debian folks go nuts ... ;)12:07
asacfta: they said they cant wait longer for chromium ... and then after discussing they said that a security team guy of debian wants to maintain it12:08
asacactually he wanted to take it over12:08
asaci suggested that he joins your team12:08
asacand told him to chat with you and me about what the rules would be in a branch12:09
asacerr in our bzr branch12:09
asacfta: anyway. first step would be to upload this weekend.12:10
asacfta: which a debian developer needs to do at first, but after that and applying for Dm status in debian for that package you can upload to debian directly and then sync to ubuntu12:11
asaci just wanted to ask if uploading to debian directly is ok with you12:11
fta*sigh*12:11
ftamore bureaucracy12:12
asacfta: but after that you get a guy that said he can deal with the security mess for chromium ;)12:12
asacand not pain because debian starts to build their own chromium way of packaging12:13
asacor something stupid like that12:13
ftawell, the problem i see here is that if that security guy works on it, it will be in the beta part, he won't maintain head and dev12:14
asacfta: is that a problem?12:14
asaci mean in general i see that its better if he could help everywhere12:14
asacbut isnt help on one side an improvment?12:15
ftathat means even more merges between the branches12:15
ftahelp is good, i just feel i won't be able to keep track, that's it12:16
asacfta: well. what i think what he will do is a) keep his own branch12:16
asacfor stable package maintenance and do cherry picks12:16
asacor12:16
asache would just do release management ... in the sense that he uploads beta or stable releases that are done anyway12:16
asace.g. we do all stable/beta bumps in our branches anyway, dont we?12:17
asacso he could take some version he thinks is good and upload that and do the paperwork for debian security team at least12:17
asacthen we can check if he does good work and use the same tags as he does (so no need to figure exactly what to upload)12:17
ftaok, i have no strong feeling about all that, it's meant to happen anyway so be it12:18
asaci think he should come here and you can talk about what would be possible/not possible.12:18
asacfta: ok. you said you have not much time today? when should i get him here?12:19
asacfta: by the way. did you ever hear if chromium linux is supposed to be on stable for 512:21
ftanope12:21
asacthis beta madness really .... ;)12:22
asacbut welll thats how it is.12:22
asacfta: let me know when i should get giuseppe in here ;)12:24
asacDerevko: hey ;)13:21
Derevkohi13:21
asacfta: ^^ ... he is debian peer i spoke about13:21
asacwe were talking about what to release to debian usually13:22
asacDerevko: so currently beta channel is the best we have13:22
asacDerevko: that means if my suggestion makes sense we would look in the releaseblog what was last announced13:22
asacand consider packaging that13:23
Derevkoasac: yes, I agree13:23
asachowever, the problem then begins ... we dont know what the releases you see on that RELEASE url13:23
asacwere about.13:23
asacthey could a) be experiments/improvments etc.13:23
Milos_SDHi. When will dom.ipc.plugins.enabled start working on Firefox 3.7 from this PPA? :(13:23
asacor b) reression fixes for that push13:23
asacregression fixes for that push ;)13:23
asacso thats where i am still in the dark ;)13:23
asacwe could assume that regression fixes would be annonuced13:24
asacI will take the action to check that with the chromium folks13:24
asacDerevko: so ok; whats last on beta release bog?13:24
asac5.0.342.913:24
asacso if my theory is riht, we should be able to find that on the beta branch13:25
asachmm13:25
asaci think we should make fta do release commits for all versions if thats a good scheme13:25
asacgreat. launchpad gives internal server error13:27
* asac branches13:27
asachmm.13:28
asacfta: i think the tagging could be improved13:28
asacfta: this is what i see with bzr log --include-merges on .beta branh13:29
asachttp://paste.ubuntu.com/421645/13:29
asacso there you seemed to have merged the packaging for the .9 release13:29
asacfrom .head13:29
asac(so seems they release .9 everywhere)13:29
asacup to beta13:29
asacanyway.. i think the .head commit should be something like ...0ubuntu1~head ... then we could have tagged the real commit on .beta branch like 0ubuntu1~beta13:30
asace.g. in a release commit after the .beta merge13:30
asacDerevko: so atm i think its the latest .beta packaging ... which is the upstream version we have on release blog13:31
asacDerevko: we could maintain a downstream branch from .beta ... and whenever we pick something merge it don13:32
asacdown13:32
asacas we might need to change Maintainer: etc. though in the end we could do tha ton the real branches ... but hav eto sync with fta as his bot might need to get fixed to do the versioning etc. right13:33
* asac talks too much13:33
asacfeels like swamping Derevko ... maybe play around and build the branch ;)13:33
Derevkoasac: I was under the impression that lp:chromium-browser was the main branch from where the Ubuntu lucid packages (and so also the Debian packages will?) were built, but perhaps I'm wrong.13:44
asacDerevko: do you have that branched? please paste bzr info inside13:45
asaclet me see13:46
asacthats .head indeed13:46
asacchecking out13:46
Derevkoasac: http://paste.debian.net/70437/plain/7043713:47
asacDerevko: yeah. seems thats currently the case13:47
asacDerevko: so we need to talk to him13:47
asacas long as we have a single branch thats fine. but imo that should be .beta13:48
ftasorry, i was off (and i'm back just for ~10min)13:48
asacno problem13:48
asacfta: can you explain why you release beta releases from .head?13:48
ftai tag and release in .head,13:48
ftathen i merge in the beta branch just for the beta ppa13:49
asacand if you release _all_ beta releases from .head or if there might be a beta release from .beta for some circumstances13:49
asacfta: what do you do if they do a cherry pick release on beta channel13:49
asacand dev etc. have already moved on?13:49
asacjust a retroactive release?13:49
ftawell, it tried many things, releasing from beta then merging back to head, or the opposite13:49
asacfta: why do you need to merge back beta releases?13:50
asaci would think its just one commit on the beta branch after the merge13:50
ftachangelog mostly13:50
asactrue. so ok13:50
asacfor the case i metntioned you could probably just branch out an old revision13:50
asacdo the release there13:50
asacmerge back into .head?13:50
asacand into .beta ;)13:50
asac(for ppa)13:51
asaci guess something like that hasnt happened yet?13:51
ftai'm constantly improving .head, but head is sometimes/often ahead, so it's a merge vs cherry pick fight13:51
ftaone way or another, i have to merge one and either cherry-pick or revert the other13:52
fta(by revert, i don't mean uncommit, just commit a reversed patch)13:52
ftarollback13:53
asacfta: so to understand when head has moved away, you branch out the old versoin that was closest? do the release there and merge that back into .head?13:53
asacwouldnt that make senes?13:53
asacwe do that for firefox when that moved ahead and it works quite well13:53
asac(though chromium is a bit of a different situation with there super volatile versions ;))13:53
ftanope, i bzr diff from the previous release, and drop everything that is too far ahead13:54
asacfta: in the branch?13:54
ftayes13:54
asacso the brnach goes backwards sometimes?13:54
ftathat's what your paste is showing13:54
* asac reads opens it again13:54
ftanot really, it recommit the changes on top just after the release13:54
fta-it+I13:55
asacfta: why not doing it like i suggested? that would be kind of a logical way.13:55
asacits kind of not nice changelog wise, yes. but it works i think ;)13:55
ftait's not perfect, but that's the best way i could find to manage 3 channels in parallel with less differences13:55
ftaas i said, the master branch is .head13:56
ftano merge in there13:56
asacso for the last beta release we could hav branches 51213:56
asacwe bake a beta release there. then just merge it back because of changelog and also merge that into the .beta branch13:56
asacpreparing the release is a manual operation anyway13:57
asacits just a retroactive mini branch13:57
ftait's possible, it's just not the way i did the last 4~5 releases13:57
asacyeah. maybe think about it ... i think thats less hazzle that way13:57
asacanyway i think i know what to do then13:58
asacfor upload we release just the last version announced ;)13:58
asaceasy enough13:58
asacthen when something is in stable releases we only upload that what was announced either as regression fix or a security update13:58
asacnow we have to find out what that is ;)13:59
asacalso we hope that regression fixes are announced ;)13:59
asacand that those bumps without announce (if there are any) usually is for adding new stuff13:59
ftajust to clarify, when i fix a bug, i do it in .head so it's immediately available to the crowd, then it will be in the next release if it's not urgent otherwise i create an incremental release immediately14:00
fta(i have to run now)14:00
asacfta: ok. so you decide what you backout based on importance? e.g. you dont revert all?14:00
asacthanks14:01
asacttyl14:01
asacDerevko: so that makes some sense.14:02
asacDerevko: at least for now ... pick the tag associated with the release we want14:02
asace.g. branch that ... then change the debian/control as we need and upload ;)14:02
asacthen next time ther eis a release we want, merge that tag to that branch and upload14:03
asacmaybe we should collapse changelogs14:03
asacthen we can at some point think about how to do the debian packaging in the branch so it becomes upstream there14:03
ftayou should branch one of those: http://paste.ubuntu.com/421663/14:04
asac(in the .head branc)14:04
asacyeah14:04
asacDerevko: add fta, me, you as the uploders and the chromium alioth mailinig list as maintainer ;)14:06
asacDerevko: are you good at perl?14:06
asaci messed up the licensecheck.pl thing ... e.g. there is a sort problem ... so generating licenseinfo updates is bad to read in diff14:07
asacnever had time to fix it ;)14:07
asacmy @values = values %{$$data{$dir}{$license}};14:09
asaci think that need to be sorted14:09
Derevkoasac: no sorry14:10
asacheh. np14:11
Milos_SDHi. When will dom.ipc.plugins.enabled start working on Firefox 3.7 from this PPA?14:31
chrisccoulsonMilos_SD - i will fix that tomorrow if nobody has done that before then16:40
chrisccoulsonoh16:40
chrisccoulsonhe didn't hang around ;)16:40
=== nikolam_ is now known as nikolam
ftaasac, i can teach you19:01
ftaasac, by what do you want to sort the hash?19:01
=== nikolam_ is now known as nikolam
BUGabundoeven|ng21:07

Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!