In order not to break downstream expressions that (manually) introduced
dependencies on now renamed attributes, add aliases (with proper
warnings about the renames) for all attributes that were previously part
of Nixpkgs, but have been renamed with this change.
Even though I couldn't find any obvious differences between our setup
here and the upstream CI (if python-dbusmock and upower are addded),
running the dbusmock tests doesn't work. TAFFYBAR_{TEST_,}VERBOSE
doesn't give many insights unfortunately.
This is useful so that you can call `Setup build` on other targets e.g.
for code generation while still having the same flags.
The context that led me to write this patch is that we have HIE file
generation in an overlay at work and I was packaging Glean, which
requires a code gen component get built, and it was building twice due
to the hie flags not getting propagated to the Setup build commands I
was running manually.
This was implemented when we were using 9.8 in order to ease the
transition to 9.10, but we don't really need to be able to build
git-annex on 9.8 anymore.
By default text-icu sets homebrew=true. This can cause a nix build
of text-icu to accidentally link the homebrew icu. It can also
cause "doesn't work on my machine" issues when an old homebrew version
with icu is installed on a mac and is accidentally picked up.
all-cabal-hashes: 2025-10-22T21:20:29Z -> 2025-11-11T08:13:55Z
(generated by maintainers/scripts/haskell/update-package-set.sh)
haskellPackages.cabal2nix-unstable: 2025-10-31 -> 2025-11-11
This eliminates the need for 2f1f72d5a2.
Without the change `haskell.packages.ghc912.extensions` eval fails as:
$ nix build --no-link -f. haskell.packages.ghc912.extensions
error:
… in the left operand of the update (//) operator
at pkgs/development/haskell-modules/lib/compose.nix:41:6:
40| f: drv:
41| (drv.override (
| ^
42| args:
error: attribute 'extensions_0_1_0_3' missing
at pkgs/development/haskell-modules/configuration-ghc-9.12.x.nix:75:29:
74|
75| extensions = doDistribute self.extensions_0_1_0_3;
| ^
76| ghc-exactprint = doDistribute self.ghc-exactprint_1_12_0_0;
Did you mean one of extensions_0_1_0_1 or extensions_0_1_1_0?
This is correctly added to libraryPkgconfigDepends where it needs to go.
We could drop this in the future if we add a corresponding mapping to
cabal2nix.
It was complaining about different versions of some regex dependency in
the closure, the differing one introduced by hls-plugin-api.
GHC 9.6 didn't fail that way, but I figured such a change can't hurt
either, for consistency.
(cherry picked from commit 7f8ab22df9)
This dependency is only needed for the "plugin-tutorial" executable.
This is only built with GHC 9.6 and 9.8, but the expression is created
with GHC 9.10. Thus add the missing dependency.
It would be better to just disable that tutorial, but I didn't find a
quick way of doing that - toggling the related flag did not work.
This only fixed the build for GHC 9.6 for me, GHC 9.8 failed with a
different error right after.
(cherry picked from commit a36931ccf2)
It was complaining about different versions of some regex dependency in
the closure, the differing one introduced by hls-plugin-api.
GHC 9.6 didn't fail that way, but I figured such a change can't hurt
either, for consistency.
This dependency is only needed for the "plugin-tutorial" executable.
This is only built with GHC 9.6 and 9.8, but the expression is created
with GHC 9.10. Thus add the missing dependency.
It would be better to just disable that tutorial, but I didn't find a
quick way of doing that - toggling the related flag did not work.
This only fixed the build for GHC 9.6 for me, GHC 9.8 failed with a
different error right after.
