Andres Löh authored 9 years ago and Mikhail Glushenkov committed 9 years ago

(cherry picked from commit 2854bcb3)

kristenk authored 9 years ago and Mikhail Glushenkov committed 9 years ago

This commit removes references to the solver log that prevented it from being
garbage collected.  It also forces evaluation of the current level and variable
stack in 'Message.showMessages'.

(cherry picked from commit 37f28f23)

(cherry picked from commit d137c93e)

Andres Löh authored 9 years ago and Mikhail Glushenkov committed 9 years ago

Previously, a setting of `--max-backjumps=-1` was reported in
the debug output as `0`. But `0` has a very different meaning,
and that is confusing.

(cherry picked from commit 8614859d)

[ci skip]

(cherry picked from commit 28b08d6b)

Andres Löh authored 9 years ago and Mikhail Glushenkov committed 9 years ago

This change primarily does two things:

1. For `--reorder-goals`, we use a dedicated datatype `Degree`
   rather than an `Int` to compute the approximate branching
   degree. We map 0 and 1 to the same value. We then use a
   lazy ordering and a shortcutting minimum function to look
   for the "best" goal.

   The motivation here is that we do not want to spend
   unnecessary work. Following any goal that has 0 or 1 as degree
   cannot really be "wrong", so we should not look at any others
   and waste time.

   This will still not always make the use of `--reorder-goals`
   better than not using it, but it will reduce the overhead
   introduced by it.

2. We use partitioning rather than sorting for most of the other
   goal reordering heuristics that are active in all situations.
   I think this is slightly more straightforward and also slightly
   more efficient, whether `--reorder-goals` is used or not.

I have run some preliminary performance comparisons and they
seem to confirm that in both cases separately (with or without
`--reorder-goals`), these changes are a relative improvement
over the status quo. I will run additional tests before
merging this into master.

(cherry picked from commit 150d6ca9)

Herbert Valerio Riedel authored 9 years ago and Mikhail Glushenkov committed 9 years ago

In GHC 8.0 -Wall implies -Wredundant-constraints, so we have to
address those warings in one way or another.

(cherry picked from commit fa8364f6)

Herbert Valerio Riedel authored 9 years ago and Mikhail Glushenkov committed 9 years ago

(cherry picked from commit 880ad15e)

(cherry picked from commit 28e4d38a)

[ci skip]

Andres Löh authored 9 years ago and Mikhail Glushenkov committed 9 years ago

A number of small changes:

- Some comment typos fixed.
- The main function for cycle detection is now called
  `cycleDetectionPhase`, in analogy with all the other
  phases.

I've run a superficial performance test trying to install
all of Hackage on a clean db with ghc-7.10.3. This is not
likely to trigger any situations where cycle detection
actually kicks in, but it confirms in general that there is
no negative performance (or correctness) impact for the
common case.

I've also considered moving the cycle detection phase to
"earlier" in the solver, but after performance testing, decided
against it, and documented the decision and the reasons in
the code.

(cherry picked from commit 64a014ec)

Edsko de Vries authored 9 years ago and Mikhail Glushenkov committed 9 years ago

(cherry picked from commit a0a80420)

Edsko de Vries authored 9 years ago and Mikhail Glushenkov committed 9 years ago

and fix bug in qualityDeps

(cherry picked from commit 927ec763)

kristenk authored 9 years ago and Mikhail Glushenkov committed 9 years ago

(cherry picked from commit b36f0a44)

kristenk authored 9 years ago and Mikhail Glushenkov committed 9 years ago

(cherry picked from commit d75e4ab1)

kristenk authored 9 years ago and Mikhail Glushenkov committed 9 years ago

Fixes #3059.

cabal's handling of non-existent sources depends on the behavior of the
directory package. 'canonicalizePath' can fail on non-existent paths before
directory-1.2.3.0. This commit updates the test 'fail_on_nonexistent_source' to
allow 'cabal sandbox delete-source' to fail or succeed.  It also changes
'fail_removing_source_thats_not_registered' so that it only tests existing
sources.

(cherry picked from commit 4a06b1d4)

(cherry picked from commit 05dccd62)

See https://github.com/haskell/cabal/issues/3198#issuecomment-190978842.

(cherry picked from commit f78ba4bb)

See https://github.com/haskell/cabal/pull/3197#issuecomment-189839042.

(cherry picked from commit d3d781b5)

(cherry picked from commit 3f0e75da)

Herbert Valerio Riedel authored 9 years ago and Mikhail Glushenkov committed 9 years ago

(cherry picked from commit 9c3e6091)

Herbert Valerio Riedel authored 9 years ago and Mikhail Glushenkov committed 9 years ago

This increases compile-time (until GHC becomes more clever) but the
generated code is expected to be at least as good (if not better) than
the manually generated code.

While at it, this removes -XCPP usage from all modules touched.

This addresses #3169

(cherry picked from commit 3bcae47a)

Herbert Valerio Riedel authored 9 years ago and Mikhail Glushenkov committed 9 years ago

This increases compile-time (until GHC becomes more clever) but the
generated code is expected to be at least as good (if not better) than
the manually generated code.

This addresses #3169

(cherry picked from commit ec6dd74d)

Herbert Valerio Riedel authored 9 years ago and Mikhail Glushenkov committed 9 years ago

This is preparatory work for implementing #3169
it's kept in a different commit in order to facilitate
comparing code-generation.

(cherry picked from commit 9b38b38b)

Herbert Valerio Riedel authored 9 years ago and Mikhail Glushenkov committed 9 years ago

This is preparatory work for implementing #3169
it's kept in a different commit in order to facilitate
comparing code-generation.

(cherry picked from commit dd5fe69a)

(cherry picked from commit 3113eb8b)

Herbert Valerio Riedel authored 9 years ago and Mikhail Glushenkov committed 9 years ago

This implements the suggestion from
https://github.com/haskell/cabal/pull/3196#issuecomment-190285328

(cherry picked from commit 0077e2c4)

Herbert Valerio Riedel authored 9 years ago and Mikhail Glushenkov committed 9 years ago

This implements the suggestions mentioned at
https://github.com/haskell/cabal/issues/3169#issuecomment-189281916

The main benefit of this change is turning 'ConfigFlags' into a uniform
product-type suitable for generic derivation of pointwise
`Semigroup`/`Monoid` instances.

NB: This changes the `Binary` serialisation of `ConfigFlags` since there's
now an additional `Maybe` inserted in `configPrograms`'s type

(cherry picked from commit 62c3aa62)

Herbert Valerio Riedel authored 9 years ago and Mikhail Glushenkov committed 9 years ago

Since Cabal makes use of Maps and Sets we need them in place in our
Semigroups compat-layer.

(cherry picked from commit a03031e7)

Herbert Valerio Riedel authored 9 years ago and Mikhail Glushenkov committed 9 years ago

(cherry picked from commit 61047bd4)

mistydemeo authored 9 years ago and Mikhail Glushenkov committed 9 years ago

Fixes #3191.

(cherry picked from commit d46e71d3)

kristenk authored 9 years ago and Mikhail Glushenkov committed 9 years ago

This reverts commit 727f4479.

This comment is unnecessary after #3180. [ci skip]

(cherry picked from commit 89c24c93)

kristenk authored 9 years ago and Mikhail Glushenkov committed 9 years ago

(cherry picked from commit 970f5b27)

lspitzner authored 9 years ago and Mikhail Glushenkov committed 9 years ago

(cherry picked from commit 7875b104)

Skip Cabal package tests that cannot run in the current environment

The functions 'testWhen' and 'testUnless' filter tests in the 'TestTreeM' monad.
The option '--enable-all-tests' ignores filtering.  This commit also applies
filtering to the existing package tests that cannot run on Windows.