Background: haskell packages extensively use upper bounds for haskell depends. It's a widespread conscious upstream decision (no need to read, but for curious: https://wiki.haskell.org/Package_versioning_policy) Recent example is bug #554768 where stable pandoc-1.12.1 had the following RDEPEND=">=dev-haskell/text-0.11 <dev-haskell/text-0.12" text-0.11 was stable for a while and all was well Then text-1.1.1.3 got stabilised as a bug #550400 and users started reporting various dependency blockers. I'll need to tweak/fast-stabilise pandoc after the fact. The request itself: Teach repoman to find "outdated" dependencies and warn about them. Examples: package A: - SLOT=0, VERSION=1.0, RDEPEND="", KEYWORDS=~amd64 - SLOT=0, VERSION=2.0, RDEPEND="", KEYWORDS=~amd64 package B: - SLOT=0, VERSION=1.0, RDEPEND="=A-1.0*", KEYWORDS=~amd64 - SLOT=0, VERSION=1.5, RDEPEND=">=A-1.0 <>=A-1.9", KEYWORDS=~amd64 - SLOT=0, VERSION=2.0, RDEPEND="=A-2.0*", KEYWORDS=~amd64 repoman should warn about: B-1.0.ebuild: not matching latest A-2.0 in SLOT=0 B-1.5.ebuild: not matching latest A-2.0 in SLOT=0 Thanks!
Another example: https://bugs.gentoo.org/554002
What is preventing this issue from being solved?
(In reply to crocket from comment #2) > What is preventing this issue from being solved? Last of a patch in my understanding.
(In reply to Sergei Trofimovich from comment #3) > (In reply to crocket from comment #2) > > What is preventing this issue from being solved? > > Last of a patch in my understanding. s/Last/Lack/
Do you mean lack of interest or manpower rather than rejection of upper bound checks?
It would be a useful enhancement. Also, this could check for things like USE or subslot dependencies that only match older versions.
(In reply to crocket from comment #5) > Do you mean lack of interest or manpower rather than rejection of upper > bound checks? Don't think anyone can answer this question for project open to outside contribution.
I think this issue is more urgent than other haskell issues since it blocks "emerge -quDN @world". Is there a way to work around this issue other than to wait for fixes to come to haskell overlay?
repoman support has been removed per bug 835013. Please file a new bug (or, I suppose, reopen this one) if you feel this check is still applicable to pkgcheck and doesn't already exist.
I think this might still be relevant for pkgcheck.