Some SourcePackageRecipeBuild attributes seem redundant
| Affects | Status | Importance | Assigned to | Milestone | |
|---|---|---|---|---|---|
| | Launchpad itself |
Medium
|
Paul Hummer | ||
Bug Description
I've just added a dia diagram of the current source package recipe tables in db-devel:
https:/
and can't see why we've got sourcepackagename and distroseries on SourcePackageRe
Related branches
- Curtis Hovey (community): Approve (rc) on 2010-05-27
- Tim Penhey (community): Approve on 2010-05-25
- Björn Tillenius (community): Approve (db) on 2010-05-24
- Stuart Bishop: Approve (db) on 2010-05-24
-
Diff: 1966 lines (+472/-387)39 files modifieddatabase/sampledata/current-dev.sql (+79/-65)
database/sampledata/current.sql (+76/-62)
database/schema/comments.sql (+0/-2)
database/schema/patch-2207-58-0.sql (+12/-0)
lib/canonical/config/__init__.py (+4/-0)
lib/canonical/launchpad/readonly.py (+1/-1)
lib/canonical/launchpad/webapp/configure.zcml (+6/-0)
lib/canonical/launchpad/webapp/publication.py (+22/-18)
lib/canonical/launchpad/webapp/tales.py (+13/-0)
lib/canonical/launchpad/webapp/tests/test_publication.py (+8/-0)
lib/lp/archiveuploader/nascentupload.py (+1/-1)
lib/lp/archiveuploader/tests/test_recipeuploads.py (+3/-3)
lib/lp/buildmaster/manager.py (+4/-1)
lib/lp/buildmaster/tests/test_buildqueue.py (+29/-29)
lib/lp/buildmaster/tests/test_manager.py (+3/-1)
lib/lp/code/browser/sourcepackagerecipe.py (+1/-4)
lib/lp/code/browser/tests/test_sourcepackagerecipe.py (+9/-16)
lib/lp/code/interfaces/sourcepackagerecipe.py (+1/-9)
lib/lp/code/interfaces/sourcepackagerecipebuild.py (+2/-8)
lib/lp/code/model/recipebuilder.py (+9/-10)
lib/lp/code/model/sourcepackagerecipe.py (+4/-12)
lib/lp/code/model/sourcepackagerecipebuild.py (+3/-9)
lib/lp/code/model/tests/test_recipebuilder.py (+9/-9)
lib/lp/code/model/tests/test_sourcepackagerecipe.py (+7/-15)
lib/lp/code/model/tests/test_sourcepackagerecipebuild.py (+1/-2)
lib/lp/registry/interfaces/person.py (+1/-4)
lib/lp/registry/model/person.py (+2/-4)
lib/lp/soyuz/browser/builder.py (+11/-1)
lib/lp/soyuz/browser/tests/test_builder_views.py (+49/-0)
lib/lp/soyuz/doc/nascentupload-announcements.txt (+3/-3)
lib/lp/soyuz/doc/package-diff.txt (+31/-31)
lib/lp/soyuz/doc/publishing.txt (+4/-4)
lib/lp/soyuz/model/packagediff.py (+1/-1)
lib/lp/soyuz/stories/soyuz/xx-package-diff.txt (+13/-13)
lib/lp/soyuz/stories/soyuz/xx-queue-pages.txt (+28/-2)
lib/lp/soyuz/templates/distroseries-queue.pt (+14/-0)
lib/lp/soyuz/templates/sourcepackagepublishinghistory-listing-archive-extra.pt (+1/-14)
lib/lp/soyuz/templates/sourcepackagerelease-diffs.pt (+1/-13)
lib/lp/testing/factory.py (+6/-20)
| Michael Hudson-Doyle (mwhudson) wrote : | #1 |
| Jonathan Lange (jml) wrote : | #2 |
What would changing mean here? That the recipe to build a package now has a new purpose and is there to build a different package, or the same package in a different series. Is that actually something we should allow? I don't really know
| Michael Hudson-Doyle (mwhudson) wrote : | #3 |
Well, what does it mean for branches? It's a bit of a vague concept. I agree. I'm not sure if we should allow it -- I should have emphasized in my first comment that I was being very paranoid on this point when I wrote the db patch.
I think the most likely situation where we might want to allow this is when the details of how the source packages are arranged changes. like a library being refactored out of an application or something. But I certainly haven't thought about it that hard.
| Michael Nelson (michael.nelson) wrote : | #4 |
As mentioned on bug 516448, building the same package in different series is probably something very desirable. That bug is about updating the schema and builder to support building the source package once, and then potentially re-building the binaries for different distroseries (ie. something that we've not done before).
But as a fall-back measure, the less efficient creation of 4 separate builds for the one recipe (mentioned by jml above) might be one good reason for having distroseries on the SPRBuild, as it seems we might want to allow it.
| Michael Nelson (michael.nelson) wrote : | #5 |
Regarding the 4 separate builds for one recipe, see Comment 3 on https:/
| Tim Penhey (thumper) wrote : | #6 |
I guess we'll re-evaluate this once we have bedded down the design a bit more.
| Changed in launchpad-code: | |
| status: | New → Triaged |
| importance: | Undecided → Medium |
| tags: | added: recipe |
| Aaron Bentley (abentley) wrote : | #7 |
Distroseries now has a a many-to-many relationship with SourcePackageRe
| Aaron Bentley (abentley) wrote : | #8 |
I think SourcePackageRe
Fixed in db-stable r9406 <http://
| Changed in launchpad-code: | |
| assignee: | nobody → Paul Hummer (rockstar) |
| milestone: | none → 10.05 |
| status: | Triaged → Fix Committed |
| tags: | added: qa-needstesting |
| Changed in launchpad-code: | |
| status: | Fix Committed → Fix Released |
Well, you've hit the nail on the head -- can you be sure the details on the SourcePackageRecipe won't change? The only reason the analogous fields for branches can't be changed is a lack of UI (you might even be able to change them through the API already).