GitLab

- Add availableSince to language/extension fields
   - Also autogen-modules once we are on that.
- Remove default-language check for cabal-version: 3.4

We cannot make the field `uniqueFieldAlaDef`, as that would require
specifying `default-language` in each subconditional.
I leave it for https://github.com/haskell/cabal/issues/6925

Note: the default language is whatever is the compiler's default.
The GHC-8.10 default is its variant of Haskell2010
with NondecreasingIndentation and without DatatypeContexts

Now it re-exports a collection of modules,
not individual symbols.

I.e. find out where we don't yet
used `Distribution.Client.Compat.Prelude`.

- If the module is small I added direct `Prelude` imports.
- Add Exception, deepseq stuff to Cabal Prelude
- Add Parsec, Pretty and Verbosity to Client Prelude
- use for, for_, traverse and traverse_ (removes need for Control.Monad)

There was three Flag's. Now they are
- `Flag` (cli parsing)
- `PackageFlag`
- `CompilerFlag`

This allows wild-imports without `hiding (Flag)`, and make tags
navigaton work!

cabal-install support is very WIP

- We had to add gpdScannedVersion to permit solver to report
  about too new packages in the index.
  https://github.com/haskell/cabal/issues/6652
- It's hard to track what happening with Setup interface selection in
  ProjectPlanning.
  https://github.com/haskell/cabal/issues/6651 would help to
  make sense of it.

Do this already in the Parsec instance parser.
This allows removing checks from `D.PD.Check`.
Later this would also allow removing non-relevant
constructors from `VersionRange`, allowing easier testing.

This change causes 10% slowdown: from

```
135768 files processed
  7350 files contained warnings
113156 files have check warnings
     0 files failed to parse

120.901201s elapsed
117.640431s elapsed
119.663620s elapsed
119.454329s elapsed
119.785214s elapsed
```

to

```
135768 files processed
 31912 files contained warnings
113109 files have check warnings
     0 files failed to parse

130.969593s elapsed
132.016403s elapsed
134.214536s elapsed
128.753382s elapsed
131.503804s elapsed
```

I hope the slowdown is acceptable, and I have an idea which may mitigate
this. I'll try out it after I done further refactorings.

Note how (parser) warnings grew by a factor. There are plenty of (old)
files on Hackage, which don't use correct cabal-version.  For that
reason we only issue warnings, and not fail.  Quirks approach won't
scale for these. In comparison, there are even more files
with check warnings

So we do with many other stuff

These are only supported properly starting with
cabal-version:3.0 (which has been enforced by
Hackage as well -- via this very patch).

See #6033 for details

Add safe alternatives to some partial functions in
Distribution.Utils.Generic, and removes the originals from
Distribution.Compat.Prelude

Uncomment versions of functions in Distribution.Compat.Parsing
that return non-empty lists
Co-authored-by: Dale Wijnand <dale.wijnand@gmail.com>

Before:

130113 files processed
7304 have lexer/parser warnings
332 build impossible
9742 build warning
49779 build dist suspicious
38666 build dist suspicious warning
11834 build dist inexcusable

After:

130113 files processed
7304 have lexer/parser warnings
332 build impossible
10063 build warning
49779 build dist suspicious
38666 build dist suspicious warning
11834 build dist inexcusable

i.e. 321 build warnings on all Hackage

Examples:

NO_DEBUG_MODE                       -- forgotten -D?
-traditional                        -- doesn't work, nor needed
-fallow-undecidable-instances       -- wrong -options?
-fno-exceptions
-Wall
-Werror
--include=include/config.h          -- doesn't work
-maes                               -- cpp is not C++ ?
-mpclmul
-mssse3

[ci skip]

autogen-includes aren't searched nor packages by `sdist`.

This is relatively small code patch, but there are
- change in file-format
- short documentation of the field
- `cabal check`
- test-suite noise due new field in `BuildInfo`

i.e. strict pair of PackageName and LibraryName
the legacy conversion is done via Pretty/Parsec instances.

Change of `Maybe UnqualComponentName` to `LibraryName` caused
a cascade of other changes, but they all seem to be good changes.
In the sense, they made many comments not-so-necessary.

Add Distribution.Types.PackageName.Magic for special package names.

Updates in cabal-install are mostly trivial type error driven changes.
I removed few (deprecated) `Text` instances: `MungedPackageId`,
`MungedPackageName` and `LibraryName`. Turns out only a `Pretty`
part was used, so it was easy to update. Note: `LibraryName`
doesn't have `Pretty` / `Parsec` instances as it's either parsed/printed
as a `ComponentName` or `UnqualComponentName`, never stand alone.

E.g. jaeger-flamegraph.cabal failed to roundtrip 'pretty . parse';

Making the list business work didn't feel right,
so I made a bit bigger refactor.

It serves an almost identical purpose as extra-library-flavours, except
that the latter is only concerned with static library flavours, while
extra-dynamic-library-flavours lets us, in the case of libHSrts, build
a dynamic, threaded flavour of it.

Create a new syntax for depending on any library of any package.
The syntax is

    build-depends: pkgname:{pkgname, sublibname} -any

where the second `pkgname` specifies a dependency on the main unnamed
library.

Closes #4206.

Previously, when running `cabal check` all the various ghc-*-options
flags were merged together, thus losing the information about the exact
place of the warning. This PR implements separate checking of
ghc-*-options, which allows us to give users more precise warnings.

Fixes #5342

This has been a problem since #5372 began expanding globs in `cabal
check`. Now the logic of running a glob is separated from the parsing,
giving the caller the opportunity to handle parsing failures flexibly.

These globs make `haddock`, `sdist` and `install` die, so they
definitely ought to be warned about!

Because of the dying behaviour, I made the checks dist-inexcusable;
starting from a clean slate I would probably have only made them
suspicious, but this isn't terrible.

This also significantly improves the error when trying to refer to
missing directories, hopefully making it clear that it's coming from
Cabal. #5318 and snowleopard/hadrian#634 are two bugs which manifested
as Cabal trying to glob in a non-existent directory and both took some
debugging because of the obscurity of the error.

Previously, we were checking the package with a hard-coded root
directory of ".". This was not a problem before, but with #5372 we
have started to expand globs while checking packages, which breaks if
the CWD is not the directory containing the `.cabal` file and causes
snowleopard/hadrian#634.

Luckily, this is an easy fix: the correct directory is easy to
determine. Writing a test and making sure it's tickling the failing
case took longer than writing the fix!

"." is hard-coded as the root directory passed to `checkPackageFiles`
in a few other places, but those are (a) non-trivial to test, and (b)
already in places that have other assumptions about their CWD, so I
have simply documented the CWD requirement for those.

This has the effect of allowing a glob `*.html` to match the file
`foo.en.html`. For compatibility, this is only allowed with
`cabal-version: 3.0` or later; for earlier spec versions, a warning
will be generated by `cabal check` if there are files affected by this
change in behaviour.

Fixes #5057. Fixes #784. Closes #5061.

These are inspired by a plan described in a comment in #2522, and only
implement a quite limited form of recursive matching: only a single **
wildcard is accepted, it must be the final directory, and, if a **
wildcard is present, the file name must include a wildcard.

Or-patterns are not implemented, for simplicity.

Closes #3178, #2030.