(cherry picked from commit 05860310)

Show import tree provenance (backport #9578)

With this change to the solver message rendering, I also fix some bugs around project imports, adding tests for those cases. Reviewers asked that the Y-shaped import checks (using IORef) be made on a separate pull request. Removing those lead to cascading deletions.

- Regenerate expected .out files
- Show tree provenance of import constraint
- Add trimmed down PackageTests/VersionPriority
- Add changelog entry
- Use NonEmpty
- Fix check for cyclical import
- Use primes for next iteration
- Remove unused LANGUAGE pragmas
- Rename to projectConfigPathRoot
- Docs for ProjectConfigPath and showProjectConfigPath
- Renaming
- Add cyclical import tests with 1 and 2 hops in cycle
- Use full path for cyclical error message
- Expected output has project with full project path
- Add fullPath local function
- Project directory as FilePath, not Maybe FilePath
- Use (_, projectFileName) binding splitFileName
- Need full path to project parsing legacy
- Inline seenImports conversion
- Add cyclical checks with same file names and hops
- Add noncyclical tests that hop over folders
- Add a project testing skipping in and out of a folder
- Update expectations of cyclical tests
- Use canonicalizePath for collapsing .. when possible
- Capture trace for later
- Add module for ProjectConfigPath
- Move functions for ProjectConfigPath to its module
- Fetch URI is not prefixed with ./https://etc
- Document normaliseConfigPath
- Add doctests for normaliseConfigPath
- Add doctest of canonicalizeConfigPath
- Show an example of canonical paths
- Use importer and importee in canonicalizeConfigPaths
- Add logging

Conform BSD-2-Clause and BSD-3-Clause text to SPDX (backport #9813)

`cabal init` text for BSD-2-Clause and BSD-3-Clause licence differed
slightly from the one published at SPDX. [1] [2]

This caused some problems to users when dealing with licence-recognition
software. [3]

[1] https://spdx.org/licenses/BSD-2-Clause.html
[2] https://spdx.org/licenses/BSD-3-Clause.html
[3] https://discourse.haskell.org/t/non-standard-license-generated-by-stack-new/9059

(cherry picked from commit 1e86730a)

Bump time upper bound to acccomodate 1.14 (backport #9848)

(cherry picked from commit e7d9b84a)

CI validate-old-ghcs: pin to haskell-actions/setup@v2.6 (backport #9859)

This provides temporary life support to the `validate-old-ghcs` action
until node16 will be finally axed by GHA in May 2024.

Workaround for:
- https://github.com/haskell/cabal/issues/9858

(cherry picked from commit 2da8b2fa)

* Bump version numbers for 3.12.0.0

Version fields, constraints, licences, and bootstrap.

☞ N.B.:
- This commit removes 8.10.7 from bootstrap list.
- This release only deals with Cabal, not cabal-install.

* Update index state in cabal.project.release

This is needed since we need the new (0.6.2.5) release of
hackage-security.

* Regenerate bootstrap .json files

Update licence list

Per https://github.com/haskell/cabal/wiki/Updating-the-license-list
to version 3.23 (2024-02-08) of SPDX License List.

Make `check` recognise `main-is` in conditional branches (backport #9768)

* Add tests for #9742

`main-is` not picked up when inside a multibranch CondNode.

* Fix comments

* Add simplifyBranch to Distribution.Types.CondTree

Goes hand in hand with with simplifyCondTree.

* Make `check` deal correctly with multiple branches

`cabal check` had a problem recognising fields in presence of
multiple branches. This patch fixes the problem and does not
meaningfully increases CI time of particularly taxing tests
(like “duplicate flagged dependencies” from MemoryUsage).

---------

Co-authored-by: mergify[bot] <37929162+mergify[bot]@users.noreply.github.com>
(cherry picked from commit 74b1f215)

Add 3.12 Cabal changelog

- generated with
     git log --pretty=oneline --no-color cabal-install-v3.10.1.0..HEAD > 3.12.prlog
- deduped
- stored WIP “release notes” file for future `cabal-install` release

Update .cabal files (backport #9761)

`extra-doc-files` and `no-autogen-paths` warnings.

(cherry picked from commit 28daab12)

Find build-tool installed programs before programs in path (backport #9762)

We must consider the path to the installed build-tool before the path to
existing versions of the build tool in paths such as `extra-prog-path`
or in the system path.

This was previously fixed by #8972 but undone by #9527.

This also renames `appendProgramSearchPath` to
`prependProgramSearchPath` to describe correctly what that function
does.

Fixes #9756

(cherry picked from commit 84c30c26)

update copyright for 2024

testsuite: Run tests in temporary system directories

1. Working directory is not contaminated with results of the testsuite.
   Tests can be written in a more liberal way (creating/writing files
   etc).
2. The tests are more hermetic.. for example it was basically impossible
   to write a test which run outside of a project context.
3. The overall path lengths will be much shorter, which has been a
   consistent issue on windows. We can remove hacks like
   `withShorterPathForNewBuildStore`.

Fixes #9711

Download stackage.org/.../cabal.config locally

cabal-install: Clarify the semantics of package-db flag

In this PR we make the `--package-db` flag apply only to the default
immutable initial package stack rather than also applying to the store
package database.

There are two special package databases which cabal install knows about
and treats specially.

* The store directory, a global cache of installed packages where cabal
  builds and installs packages into.
* The inplace directory, a local build folder for packages, where cabal
  builds local packages in.

It is very important that cabal registers packages it builds into one of
these two locations as there are many assumptions that packages will end
up here.

With the current design of the `--package-db` flag, you are apparently
allowed to override the store location which should have the effect of
making the last package database in the package stack the "store"
directory. Perhaps this works out in theory but practically this
behaviour was broken and things were always registered into the store
directory that cabal knew about. (The assertion in
`ProjectBuilding.UnpackedPackage` was failing (see added test)).

With `--package-db` not being able to modify the location of the store
directory then the interaction of `--package-db`, `--store-dir` and
`--dist-dir` flags become easy to understand.

* `--package-db` modify the initial package database stack, these
  package database will not be mutated by cabal and provide the initial
  package environment which is used by cabal.
* `--store-dir` modify the location of the store directory
* `--dist-dir` modify the location of the dist directory (and hence
  inplace package database)

Treating the flags in this way also fix an assertion failure when
building packages.

Fixes #9678

CI updates (drop 8.4.4, update old-ghcs, update bootstrap plans)

It fails in the old-ghcs 8.4.4 configuration, and debugging why seems to
have very little pay-off for this niche feature in a very old
configuration.

And add the relevant CI jobs for these plans

We were missing coverage from running tests with these ghc versions.

We no longer support 8.4.4 as it is older than three years.
cabal-install can
still be used with 8.4.4, but just not built with it, and this is tested
by the old-ghcs jobs.

add the extension ListTuplePuns (GHC proposal 475)

Fix #9534 Add constructor `PPC64LE` to type `Arch`, for powerpc64le

Also makes consequent changes.

Fix multi-repl closure computation with invalid Custom setup

This patch deletes a lot of hairy logic to do with removing "bad" edges
from the closure. Now the solver just forces `Cabal >= 3.11` universally
when `--enable-multi-repl` is enabled.

Fixes #9710