[11:23] <rbasak> May I have some help understanding Launchpad's model against getChangesUrl() in source_package_publishing_history please?
[11:23] <rbasak> AFAICT, it sometimes returns None, which makes sense when the original upload came from Debian.
[11:23] <rbasak> When there's a pocket copy (eg. to release from a proposed upload), the source_package_publishing_history entry for the release pocket does return a URL from getChangesUrl()
[11:24] <rbasak> So is there any case I need to track back through copied_from_archive?
[11:24] <rbasak> I wrote the code to do it, but I'm now wondering if it's even necessary.
[11:24] <rbasak> I've not managed to find a case where getChangesUrl() returns None and the upload didn't originally come from Debian.
[11:35] <rbasak> It looks like getChangesUrl() is also following back through a pocket copy from a PPA by itself
[11:38] <cjwatson> rbasak: Yes, unless the upload was created by the "gina" importer (which we only use for Debian in practice), there'll be a corresponding PackageUpload row and SPPH.changesFileUrl() will be able to find the .changes file from that
[11:40] <rbasak> Great. Thanks!
[13:47]  * ogra wonders why https😕/launchpad.net/~build.snapcraft.i👋+snap/323809a9ca4f03013b934353a1358c6f/+build/1254491 did actually succeed ... (it runs git clone after ~2h in a loop during build (pulling submodules) and normally the proxy access is timed out by that time...)
[13:47] <ogra> grmbl
[13:47] <ogra> silly emoji stuff
[14:16] <cjwatson> The logs indicate that the build started at 22:42:11, and the last connection to the proxy was 01:41:01.  The timeout is currently 3h, so that just scraped in under the wire.
[14:17] <cjwatson> [5~3[6~[6~ 
[14:17] <cjwatson> (oops)
[14:23] <ogra> heh, lucky one then
[14:24] <ogra> (this is really funny, because i added another part *before* everything, so it should actually kick in later, not earlier)