Summary: | dev-haskell/cryptohash-0.11.9 potential slot interaction with dev-haskell/cryptonite-0.21 | ||
---|---|---|---|
Product: | Gentoo Linux | Reporter: | Kent Fredric (IRC: kent\n) (RETIRED) <kentnl> |
Component: | Current packages | Assignee: | Gentoo's Haskell Language team <haskell> |
Status: | RESOLVED DUPLICATE | ||
Severity: | normal | ||
Priority: | Normal | ||
Version: | unspecified | ||
Hardware: | All | ||
OS: | Linux | ||
Whiteboard: | |||
Package list: | Runtime testing required: | --- |
(In reply to Kent Fredric (IRC: kent\n) from comment #0) > Just in general, my understanding is this tool is a stop-gap to handle > problems that aren't fixed by slot-op rebuilds. > > This may not be a bug. *shrug*. slot op is not expressive enough to define the following properties: - if package ABI depends on ABI of not just immediate DEPENDS but on ABI of transitive dependencies you can't express it: bug #449094 - if package ABI depends on package USE flags you can't express it There is no easy workaround for it and requires PMS extension. *** This bug has been marked as a duplicate of bug 449094 *** |
I've hit upon an odd situation, after many @world upgrades, I still have this in @preserved-rebuild, and I usually take this as a sign something is wrong: >>> package: dev-haskell/cryptonite-0.21 * - /usr/lib64/x86_64-linux-ghc-8.0.2/libHScryptonite-0.21-2Ew6DXO9nl8IZIwltTAY3M-ghc8.0.2.so * used by /usr/lib64/x86_64-linux-ghc-8.0.2/libHScryptohash-0.11.9-DGrJmUFHDPW76lxYxTylPv-ghc8.0.2.so (dev-haskell/cryptohash-0.11.9) Relevant qlop history: 2016-03-18T14:48:35 >>> dev-lang/ghc-7.10.3: 2 hours, 18 minutes, 25 seconds 2016-03-18T17:34:53 >>> dev-haskell/cryptohash-0.11.6: 40 seconds 2017-01-28T04:14:15 >>> dev-lang/ghc-7.10.3: 4 hours, 49 minutes, 9 seconds 2017-02-18T07:43:03 >>> dev-lang/ghc-8.0.2: 5 hours, 57 minutes, 1 second 2017-02-18T18:14:48 >>> dev-haskell/cryptohash-0.11.6: 1 minute, 1 second 2017-06-07T02:13:37 >>> dev-haskell/cryptonite-0.21: 1 minute, 18 seconds 2017-06-07T02:14:55 >>> dev-haskell/cryptohash-0.11.9: 31 seconds 2017-12-13T22:23:34 >>> dev-lang/ghc-8.0.2: 2 hours, 59 minutes, 18 seconds 2017-12-14T01:47:44 >>> dev-haskell/cryptonite-0.21: 1 minute, 52 seconds 2017-12-14T01:55:07 >>> dev-haskell/cryptohash-0.11.9: 39 seconds 2018-09-08T00:42:51 >>> dev-lang/ghc-8.0.2: 3 hours, 19 minutes, 23 seconds 2018-10-12T06:28:02 >>> dev-lang/ghc-8.0.2: 4 hours, 10 minutes, 7 seconds 2019-09-12T10:22:05 >>> dev-haskell/cryptonite-0.21: 1 minute, 30 seconds 2019-09-12T10:32:28 >>> dev-haskell/cryptohash-0.11.9: 40 seconds 2019-09-18T14:48:44 >>> dev-lang/ghc-8.0.2: 4 hours, 23 minutes, 41 seconds 2019-10-13T23:08:30 >>> dev-haskell/cryptonite-0.21: 1 minute, 56 seconds 2019-11-25T08:03:42 >>> dev-haskell/cryptonite-0.21: 1 minute, 44 seconds As it is, portage isn't even considering that cryptohash needs rebuilding: emerge -uvatDN1 --backtrack 10 cryptohash These are the packages that would be merged: Calculating dependencies... done! Total: 0 packages, Size of downloads: 0 KiB Nothing to merge; quitting (Usually that invocation is enough for subslot-op rebuilds ) haskell-updater does indeed however detect that work needs done: Found the following broken packages: * dev-haskell/cryptohash:0 * dev-vcs/darcs:0 Just in general, my understanding is this tool is a stop-gap to handle problems that aren't fixed by slot-op rebuilds. This may not be a bug. *shrug*.