barrucadu/dejafu
1
1
Fork 0
mirror of https://github.com/barrucadu/dejafu.git synced 2024-12-18 19:11:37 +03:00
Code Issues Projects Releases Wiki Activity
636 Commits 2 Branches 194 Tags 3.4 MiB
6bc2fd044f
Commit Graph

636 Commits

This Branch
This Branch
All Branches
Author SHA1 Message Date
Michael Walker
6bc2fd044f Remove unused imports 2017-06-07 16:59:24 +01:00
Michael Walker
f590f233a8 Implement all nonsense checks directly in the sct* functions 2017-06-07 16:50:56 +01:00
Michael Walker
040580d4f2 Remove bound-specific SCT functions 2017-06-07 16:50:56 +01:00
Michael Walker
6b1fd17024 Implement uniform random scheduling 
Adds a new `uniformly` smart constructor and `sctUniformRandom`
function.

Also renames `sctRandom` to `sctWeightedRandom`.
 2017-06-07 16:50:56 +01:00
Michael Walker
1146ce9b38 Add a smart constructor for constructing swarmy executions 
The `randomly` constructor now corresponds exactly to the old
`Randomly`.

Also refactor tests a bit.
 2017-06-07 16:50:56 +01:00
Michael Walker
fba26e6c73 Make Way abstract and expose smart constructors 2017-06-07 16:50:56 +01:00
Michael Walker
d65b8359a9 Allow re-using the weights for multiple executions in sctRandom 2017-06-07 16:45:43 +01:00
Michael Walker
a0698bf575 Merge branch 'propcheck' 
Closes #82
 2017-06-07 14:32:19 +01:00
Michael Walker
748a55d813 Re-export Test.DejaFu.Refinement from Test.DejaFu 2017-06-07 14:25:55 +01:00
Michael Walker
94e22a765e Add tests for refinement checking 
New coverage report, as there have been a bunch of new tests since the
one in the CONTRIBUTING file was generated:

 54% expressions used (4311/7948)
 51% boolean coverage (67/131)
      47% guards (50/106), 31 always True, 6 always False, 19 unevaluated
      57% 'if' conditions (11/19), 2 always True, 1 always False, 5 unevaluated
     100% qualifiers (6/6)
 61% alternatives used (413/671)
 83% local declarations used (212/254)
 28% top-level declarations used (313/1099)
 2017-06-07 14:25:55 +01:00
Michael Walker
e785e51ddc Expose refinement properties in hunit-dejafu and tasty-dejafu 2017-06-07 14:20:04 +01:00
Michael Walker
a62350dd5c Switch to new refinement operators used in CoCo 
Also include some explanation of this apparently backwards notation.
 2017-06-07 14:20:00 +01:00
Michael Walker
3af7d4206d Implement refinement property testing 
This lets dejafu test CoCo-style properties, and introduces a
dependency on [leancheck][].

For example, consider this statement about MVars: "using readMVar is
better than a takeMVar followed by a putMVar because the former is
atomic but the latter is not."

This module can test properties like that:

```
sig e = Sig
  { initialise = maybe newEmptyMVar newMVar
  , observe    = \v _ -> tryReadMVar v
  , interfere  = \v s -> tryTakeMVar v >>
                         maybe (pure ()) (void . tryPutMVar v) s
  , expr = e
  }

> check $ sig (void . readMVar) `equivalentTo`
          sig (\v -> takeMVar v >>= putMVar v)
*** Failure: (seed Just ())
    left:  [(Nothing,Just ())]
    right: [(Nothing,Just ()),(Just Deadlock,Just ())]
```

The two expressions are not equivalent, and we get given the
counterexample!

In addition to "equivalentTo", there is also "refines" and
"strictlyRefines" (and negated variants)

[leancheck]: https://hackage.haskell.org/package/leancheck
 2017-06-07 14:18:50 +01:00
Michael Walker
4ffd7587ad Define default text execution parameters in Test.DejaFu.Defaults 2017-05-31 15:26:27 +01:00
Michael Walker
44c31f879c Update test-framework-hunit git location 2017-05-31 15:08:06 +01:00
Michael Walker
b6713505c5 Prevent re-use of MVar, CRef, and TVar IDs inside subconcurrency. 
Fixes #81.
 2017-05-03 23:12:26 +01:00
Michael Walker
c11f1d92db Test case for #81. 
This is a large diff because it pulls in a variant of QSemN.
 2017-05-03 23:09:56 +01:00
Michael Walker
2d97030abe tasty-dejafu-0.5.0.0 2017-04-08 21:18:20 +01:00
Michael Walker
9af76cc4f2 hunit-dejafu-0.5.0.0 2017-04-08 21:15:32 +01:00
Michael Walker
dd5748e3e6 dejafu-0.6.0.0 2017-04-08 21:08:03 +01:00
Michael Walker
efaf353920 Rename Conc to ConcT and turn into a MonadTrans. 
Closes #70.
 2017-04-08 20:59:05 +01:00
Michael Walker
d3062234fa Make Way a GADT. 
Closes #65.
 2017-04-08 20:57:25 +01:00
Michael Walker
94c4afe090 CHANGELOG errata: dejafu-0.5.1.2 should have been 0.5.2.0. 
Introducing new non-orphan typeclass instances is a minor bump.
 2017-04-08 20:21:22 +01:00
Michael Walker
9c33eff1f7 Add Haddock @since annotations to everything in dejafu/hunit/tasty 2017-04-08 20:21:22 +01:00
Michael Walker
98af53bbc3 Consider Test.DejaFu.Common part of the public API. 2017-04-08 20:21:22 +01:00
Michael Walker
ab599b637d Add guidance on documenting changes to CONTRIBUTING. 2017-04-08 20:21:22 +01:00
Michael Walker
6b71e9a76f Remove the internal SCT graphviz functions. 2017-04-08 10:38:06 +01:00
Michael Walker
da09c2268a Inline the findSchedulePrefix predicate. 
The function is only used in one place, with a constant
predicate. There was no point in it being a parameter.
 2017-04-08 10:34:46 +01:00
Michael Walker
5f9ad29b1b Include READMEs in cabal sdists. 
Closes #79.
 2017-04-08 10:30:58 +01:00
Michael Walker
93d8cc64e3 Add a separate CONTRIBUTING file. 2017-04-08 05:55:42 +01:00
Michael Walker
d4df9eed0c Add section on code style to README + scripts to run tools. 
- `lint.sh` runs HLint, returning 1 if HLint is NOT version 1 (as the
  .hlint.yaml configuration file is only supported in version 2 and
  later) AND issues are found.

- `style.sh` runs stylish-haskell. I would like it to return 1 if it
  reformatted anything, but the tool doesn't appear to support
  reporting that :(
 2017-04-08 05:42:26 +01:00
Michael Walker
f90ecb2125 Change "Code Coverage" to "Test Coverage" in README. 2017-04-08 05:42:25 +01:00
Michael Walker
cc27e03d6d Use stylish-haskell to format import lists. 2017-04-08 05:42:25 +01:00
Michael Walker
d934222fe7 Implement isEmptyMVar in terms of tryRead, not tryTake. 2017-04-08 05:42:25 +01:00
Michael Walker
5cd55a1921 Add an HLint2 config file & fix warnings. 2017-04-08 05:42:25 +01:00
Michael Walker
0080678a65 Add a repo-wide changelog. 
Closes #78.
 2017-04-07 20:22:22 +01:00
Michael Walker
de8ced69a6 Fiddle with changelog format. 2017-04-07 20:05:11 +01:00
Michael Walker
60f771d50e Write changelogs. 2017-04-07 18:16:30 +01:00
Michael Walker
8d23a2715e Bump concurrency to 1.1.2.0 and dejafu to 0.5.1.3. 2017-04-05 21:41:22 +01:00
Michael Walker
7cd0c665fa Add Haddock "@since" comments to everything in concurrency. 2017-04-05 21:25:02 +01:00
Michael Walker
da9f622ca9 Add Stackage LTS-8 to the travis config. 2017-04-05 21:06:25 +01:00
Michael Walker
a4ab84e132 Only define Semigroup Async instance for base >= 4.9.0.0. 2017-04-05 21:01:23 +01:00
Michael Walker
1532a75f78 Add replicateConcurrently{,_} to Async. 2017-04-05 20:57:27 +01:00
Michael Walker
92de856ed0 Add concurrently_, mapConcurrently_, and forConcurrently_. 2017-04-05 20:56:20 +01:00
Michael Walker
2e7aeac8b3 Add Semigroup and Monoid instances for Concurrently. 2017-04-05 20:54:21 +01:00
Michael Walker
71beaed0e5 Add uninterruptibleCancel to Async. 2017-04-05 20:46:00 +01:00
Michael Walker
15af09a406 Rexport mask_ and uninterruptibleMask_ from Conc.Class 2017-04-05 20:45:40 +01:00
Michael Walker
9a7a883ab9 Update cancel docs to mention blocking. 2017-04-05 20:36:53 +01:00
Michael Walker
dcddfa42a5 Merge pull request #77 from gamazeps/cancel 
Make cancel blocking
 2017-04-05 20:33:23 +01:00
Felix Raimundo
ed08975e3f Make cancel blocking 
In the documentation of `cancel`, it is said to be blocking, waiting on
the cancelled `async` to die. It seems to me that is not the case.

This patch simply applies the code written by Simon Marlow in order
to wait for the termination of the `async`.

See:
https://hackage.haskell.org/package/async-2.1.1/docs/src/Control.Concurrent.Async.html#cancel
 2017-04-05 16:44:08 +02:00