| === ePierre_ is now known as ePierre | ||
| === mthaddon` is now known as mthaddon | ||
| === chihchun_afk is now known as chihchun | ||
| === chihchun is now known as chihchun_afk | ||
| smoser | hey... wondering if there is a easy solution for this that we're missing. | 16:51 |
|---|---|---|
| smoser | we have a recipe building curtin and publishing to a 'daily build ppa'. | 16:51 |
| smoser | then, we have code that is testing the package build from that daily build ppa (on a nightly basis) | 16:51 |
| smoser | we have a test harness that lives in the source. | 16:51 |
| smoser | so we apt-get install curtin | 16:52 |
| smoser | and apt-get source curtin | 16:52 |
| smoser | and then run the test harness from the source | 16:52 |
| smoser | the issue is that we find skew between source version and binary version | 16:52 |
| smoser | and then we have expected failures because the harness isnt at same level as the code | 16:52 |
| nacc | smoser: why not use curtin= ? | 17:09 |
| nacc | smoser: to force them to be the same | 17:09 |
| smoser | nacc: source package ? | 17:10 |
| smoser | confused. | 17:10 |
| smoser | apt-get source | 17:10 |
| smoser | and apt-get install | 17:10 |
| smoser | get different versions | 17:10 |
| smoser | thats my problem. (because the binary just isnt published or possibly built yet) | 17:10 |
| smoser | originally i guess i assumed that the source would only ever be published if the binary was built. | 17:11 |
| nacc | smoser: if you can apt-get install curtin | 17:11 |
| nacc | then you can find what version you have installed | 17:11 |
| nacc | then apt-get source curtin=<version> | 17:11 |
| nacc | presumingn the ppa has kept that source avaialblle | 17:11 |
| smoser | oh. yeah... and do you think that will always be there? | 17:12 |
| nacc | smoser: i forget the rules for ppas | 17:12 |
| smoser | do the ppa archives keep more than one ? | 17:12 |
| smoser | thats worth a try though. thanks. | 17:12 |
| nacc | smoser: it won't go back super far regardlless, but i'm assuming this only matters for the most recnet build(s)? | 17:14 |
| nacc | smoser: you can check the source list on the ppa page, or using lplib, i thinnk | 17:14 |
| nacc | smoser: you could see what import-ppa does (from git-ubuntu) for your ppa | 17:14 |
| smoser | nacc: thanks | 17:15 |
| smoser | rharper: https://code.launchpad.net/~smoser/curtin/+git/curtin/+merge/337290 | 17:24 |
| rharper | nice | 17:25 |
| nacc | smoser: yeah that seems reasonablle. I wouldl expect a PPA to keep the sourcse used to build any packages that are currently installlable; obviously this is a race (i'm guessing) where a new source is being built, so it's possible they would disappear midway, but it seems a bit unlikely | 17:39 |
| cjwatson | smoser,nacc: it probably won't be published in Sources files, but you should be able to do dget https://launchpad.net/~OWNER/+archive/ubuntu/PPANAME/+files/curtin_VERSION.dsc (or moral equivalent) | 23:10 |
| nacc | cjwatson: ah good to know | 23:12 |
| smoser | cjwatson: ok. thank you. harder than i'd like it to be. but ok. | 23:45 |
| cjwatson | I miiiight be wrong, but I thought this was in my list of "boy I sure wish this worked differently, but I don't feel like spending the non-trivial effort involved in fixing it" | 23:51 |
| wgrant | cjwatson is correct. Only the most recent source remains published on ppa.launchpad.net. | 23:54 |
| wgrant | The only case where multiple versions of a particular source or binary are published in an (archive, suite, arch) is atomic arch-indep domination. | 23:55 |
| wgrant | A similar mechanism could probably be used for sources, now that atomic arch-indep domination exists, but nobody's ever done that. | 23:55 |
| cjwatson | The Debian archive has done it for a while. | 23:56 |
| wgrant | Yeah | 23:57 |
| wgrant | I mean nobody's looked at it seriously in LP's history AFAIK | 23:57 |
| wgrant | It also ties into the whole Built-Using thing. | 23:57 |
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!