[06:25] <CoderEurope> good morning hobbits
[08:49] <CoderEurope> How do I contribute to English Canadian on launchpad ?
[09:25] <CoderEurope> What is the bug list for ubuntuOS artful aadrvark 17.10 ?
[09:46] <tshirtman> hello, i messed up updating the ppa o
[09:46] <tshirtman> f our project, and pushed with the wrong version in changelog
[09:46] <tshirtman> (10.0 instead of 1.10.0) now i'd like to correct it, but build got rejected because it's lower than the previous successful build
[09:47] <tshirtman> (kivy-stable ppa)
[09:47] <tshirtman> any insight of how to correct that?
[09:52] <tshirtman> hm previous related questions are not encouraging https://answers.launchpad.net/launchpad/+question/126403 https://answers.launchpad.net/launchpad/+question/12218 https://answers.launchpad.net/launchpad/+question/19614
[09:52] <tshirtman> i guess i'll ask for clearing the wrongly build package
[09:57] <tshirtman> (https://answers.launchpad.net/launchpad/+question/644496)
[10:08] <cjwatson> I've answered your question there.
[10:18] <tshirtman> thanks a lot, found the page :)
[16:01] <nacc> cjwatson: what is the default sort order for getPublishedSources if order_by_date is not set? is it in version order (newest to oldest)?
[16:07] <cjwatson> nacc: source package name (lexicographical), SPPH ID (highest to lowest), version (highest to lowest, omitted if you specify a version)
[16:09] <nacc> cjwatson: thanks! is that specified somewhere? i'm going to be leveraging that in our code and want to document it, if possible
[16:16] <cjwatson> nacc: only in code at the moment ...
[16:16] <cjwatson> nacc: are you sure you want to rely on this?  tell me more about what you're doing
[16:16] <rbasak> Won't SPPH ID be always unique, and so version based sorting will never happen? Or am I missing something?
[16:17] <cjwatson> rbasak: I think the version fallback is indeed probably pointless, but it's there in the code at the moment
[16:17] <cjwatson> oh, I quoted it wrongly
[16:17] <cjwatson> it's actually source package name (lexicographical), version (highest to lowest, omitted if you specify a version), SPPH ID (highest to lowest)
[16:17] <cjwatson> that makes more sense
[16:17] <cjwatson> nacc: ^-
[16:17] <rbasak> Ah. Thanks!
[16:18] <rbasak> I'm not sure of the exact logic nacc's relying on so I'll leave it to him to answer your question.
[16:18] <cjwatson> (a given source package can be copied, which creates a new SPPH, so the effect is that newer copies of the same package show up first)
[16:18] <rbasak> I know we were looking for version-based sorting. I'm not sure exactly why.
[16:18] <rbasak> Possibly to walk backwards to find something that has a matching orig tarball or something.
[16:18] <nacc> cjwatson: well for the purposes of our 'dpkg-buildpackage -S' wrapper, I want to find the most recent version (if it exists) that is the same upstream version to grab the orig tarball for. But I want to shortcut out of the search (as it can take a while) if the upstream version we are building is newer than any published version (e.g., for a uupdate)
[16:19] <cjwatson> I guess it indeed doesn't make sense to do your own sort then
[16:19] <nacc> I don't need it to be in version order for the general case, and really don't want it that way for the general case as we're often based upon the publishing order (e.g., a fix to the most recent xenial publish). But this one case (uupdate) is special :)
[16:19] <cjwatson> It's unlikely to change, but it would probably be a good idea to file a bug (or just an MP ...) to document the current sort order
[16:19] <nacc> cjwatson: ack, will do
[16:20] <cjwatson> Your general case basically sounds like order_by_date
[16:20] <nacc> cjwatson: yep, and that's what we were using
[16:20] <nacc> cjwatson: *are using
[16:20] <cjwatson> which is technically Desc(SPPH.datecreated), Desc(SPPH.id), but should normally amount to the same thing except in edge cases of nearby transactions
[16:21] <nacc> cjwatson: yep, makes sense
[16:22] <nacc> cjwatson: unless you can think of a saner way for me to query launchpad for (is my upstream version newer than any published upstream version in (ubuntu, debian))
[16:22] <cjwatson> FWIW, on the teach-to-fish principle, you can find this in lib/lp/soyuz/model/archive.py and then grep for def getPublishedSources
[16:23] <nacc> cjwatson: excellent, thanks!
[16:23] <cjwatson> I don't think LP has any more immediate way to do that at the moment
[16:23] <cjwatson> that I can think of anyway
[16:23] <nacc> cjwatson: sure, that's fine -- I appreciate you taking the time to teach this fisherman-wannabe
[16:23] <cjwatson> the interface declaration that shows up in the apidoc and such is in lib/lp/soyuz/interfaces/archive.py
[16:24] <cjwatson> that's a general pattern for most things
[19:11] <nacc> cjwatson: thanks! https://code.launchpad.net/~nacc/launchpad/document-getpublishedsources-sort/+merge/326177
[23:46] <cjwatson> nacc: merged, thanks
[23:47] <nacc> cjwatson: thank you!
[23:47] <nacc> rbasak: --^ fyi (public documentation, if it helps with your review for the tooling change)
[23:57] <rbasak> Thanks!