January 2011 - cabal-devel

cabal test and detailed-0.9 strangeness
by Daniel Peebles 10 Jan '11

10 Jan '11

Hi, I've been playing with the detailed-0.9 test-suite option on a package of mine. First, what I have: in my package root, I have my .cabal file and a tests folder containing a Tests.hs in my Tests.hs, I took the code from the online user guide (minus the odd guards with == True and == False): {-# LANGUAGE FlexibleInstances #-} module Tests ( tests ) where import Distribution.TestSuite instance TestOptions (String, Bool) where name = fst options = const [] defaultOptions _ = return (Options []) check _ _ = [] instance PureTestable (String, Bool) where run (name, True) _ = Pass run (name, False) _ = Fail (name ++ " failed!") test :: (String, Bool) -> Test test = pure -- In actual usage, the instances 'TestOptions (String, Bool)' and -- 'PureTestable (String, Bool)', as well as the function 'test', would be -- provided by the test framework. tests :: [Test] tests = [ test ("bar-1", True) , test ("bar-2", False) ] in my .cabal file: Test-Suite binutils Hs-source-dirs: tests/ Type: detailed-0.9 Test-module: Tests Build-depends: base, Cabal >= 1.9.2 I then run cabal configure --enable-tests and finally run cabal test. It gives me this: Running 1 test suites... Test suite binutils: RUNNING... Test suite binutils: PASS Test suite logged to: dist/test/charm-0.0.1-binutils.log 1 of 1 test suites (0 of 0 test cases) passed. Which seems odd, since there are 2 test cases (not 0 as the output claims) and one of them should definitely be failing (it has a False in it). Am I doing something wrong? Thanks, Daniel

3 5

patch applied (cabal): "Update the changelog for 1.8.0.4, 1.8.0.6 and 1.10.0.0" and 2 others
by Duncan Coutts 10 Jan '11

10 Jan '11

Thu Nov 18 12:37:01 PST 2010 Duncan Coutts <duncan(a)haskell.org> * Update the changelog for 1.8.0.4, 1.8.0.6 and 1.10.0.0 M ./changelog -2 +42 View patch online: http://darcs.haskell.org/cgi-bin/darcsweb.cgi?r=cabal;a=darcs_commitdiff;h=… Sat Dec 4 08:44:04 PST 2010 Thomas Tuegel <ttuegel(a)gmail.com> * Document the '--{en,dis}able-tests' options in the User Guide M ./doc/Cabal.markdown +14 View patch online: http://darcs.haskell.org/cgi-bin/darcsweb.cgi?r=cabal;a=darcs_commitdiff;h=… Sun Jan 9 10:08:11 PST 2011 Thomas Tuegel <ttuegel(a)gmail.com> * Removed documentation for obsolete options to 'setup test' from the User Guide. M ./doc/Cabal.markdown -10 View patch online: http://darcs.haskell.org/cgi-bin/darcsweb.cgi?r=cabal;a=darcs_commitdiff;h=…

1 0

Re: New test-suite functionality
by Duncan Coutts 10 Jan '11

10 Jan '11

On Fri, 2011-01-07 at 12:02 -0600, Thomas Tuegel wrote: > I actually tried to set out doing this today and realized it's not a > good idea to do it this way. I think it would be better to add a > field such as "testEnabled :: Bool" to the TestSuite object. This has > several advantages: > 1. We can distinguish between packages configured with tests disabled > and packages without test suites. > 2. Users could selectively enable and disable particular tests, which > is good from a dependency standpoint. > 3. During dependency resolution, we can actually tell if a test is > enabled or disabled, i.e., we don't have to cut out all the test > suites before resolution. (This would fix several of the TODO's that > you've indicated in Distribution.Simple.Configure.) > > This would be a little bit more involved than my previous, naive > suggestion, but I think it's worth it. Let me know if you agree; I > just don't want to waste your time by submitting a misguided patch. I'm slightly reticent to add it to the TestSuite bit itself because it's mixing two things: a static description of the package, with status tracking within the build system. That said, there's not an obvious place to put it that exists all the way through the dep resolution stage. If it were just later on then we could add something to the LocalBuildInfo. Give it a go, but add a TODO note suggesting the status info might be better to live somewhere else. We might be able to do it better when we next overhaul the GenericPackageDescription -> PackageDescription resolution stuff. Duncan

1 0

Re: [Hackage] #467: Add support for HPC profiling.
by Hackage 09 Jan '11

09 Jan '11

#467: Add support for HPC profiling. ----------------------------+----------------------------------------------- Reporter: duncan | Owner: Type: enhancement | Status: new Priority: normal | Milestone: Cabal-1.8 Component: Cabal library | Version: 1.6.0.1 Severity: normal | Keywords: Difficulty: normal | Ghcversion: Platform: | ----------------------------+----------------------------------------------- Comment(by tibbe): We also need to * set a separate `hpcdir` when building each component, and * set the `HPCTIXFILE` environment variable when running each test suite. -- Ticket URL: <http://hackage.haskell.org/trac/hackage/ticket/467#comment:5> Hackage <http://haskell.org/cabal/> Hackage: Cabal and related projects

1 0

Re: [Hackage] #467: Add support for HPC profiling.
by Hackage 09 Jan '11

09 Jan '11

#467: Add support for HPC profiling. ----------------------------+----------------------------------------------- Reporter: duncan | Owner: Type: enhancement | Status: new Priority: normal | Milestone: Cabal-1.8 Component: Cabal library | Version: 1.6.0.1 Severity: normal | Keywords: Difficulty: normal | Ghcversion: Platform: | ----------------------------+----------------------------------------------- Changes (by tibbe): * cc: johan.tibell@… (added) -- Ticket URL: <http://hackage.haskell.org/trac/hackage/ticket/467#comment:4> Hackage <http://haskell.org/cabal/> Hackage: Cabal and related projects

1 0

Re: [Hackage] #467: Add support for HPC profiling.
by Hackage 09 Jan '11

09 Jan '11

#467: Add support for HPC profiling. ----------------------------+----------------------------------------------- Reporter: duncan | Owner: Type: enhancement | Status: new Priority: normal | Milestone: Cabal-1.8 Component: Cabal library | Version: 1.6.0.1 Severity: normal | Keywords: Difficulty: normal | Ghcversion: Platform: | ----------------------------+----------------------------------------------- Comment(by tibbe): Currently the best way to get test coverage is to: {{{ cabal configure --enable-tests --ghc-option=-fhpc cabal build cabal test hpc markup <test-suite-binary.tix> }}} This dumps a bunch of HTML files in the root directory of the package. I suggest we add a configure flag `--enable-test-coverage` that: * adds `ghc-option: -fhpc` to all libraries, * runs the `cabal test` binaries in some working directory other than the root to not clutter it with .tix files, * runs `hpc <test-suite-binary.tix> --destdir=dist/SOMETHING` on all .tix files, and * prints the location of the HTML index files to stdout. We might want to opt-out the test suite's `main-is` file from the report. Perhaps using a flag. -- Ticket URL: <http://hackage.haskell.org/trac/hackage/ticket/467#comment:3> Hackage <http://haskell.org/cabal/> Hackage: Cabal and related projects

1 0

[Hackage] #786: Bad errors from getProgramInvocationOutput and friends
by Hackage 08 Jan '11

08 Jan '11

#786: Bad errors from getProgramInvocationOutput and friends ----------------------------+----------------------------------------------- Reporter: igloo | Owner: Type: defect | Status: new Priority: normal | Milestone: Component: Cabal library | Version: HEAD Severity: normal | Keywords: Difficulty: unknown | Ghcversion: Platform: | ----------------------------+----------------------------------------------- In http://hackage.haskell.org/trac/ghc/ticket/4883 `ghc-cabal` gives an unhelpful error: {{{ ghc-cabal: }}} This comes from `getProgramInvocationOutput`: {{{ (output, errors, exitCode) <- rawSystemStdInOut verbosity path args Nothing utf8 when (exitCode /= ExitSuccess) $ die errors }}} as there was nothing on `stderr`. There are several similar functions in `Cabal`. Cabal ought to also say what it was running and what the error code was, without needing a verbosity flag to be given. -- Ticket URL: <http://hackage.haskell.org/trac/hackage/ticket/786> Hackage <http://haskell.org/cabal/> Hackage: Cabal and related projects

1 0

Please apply: Patch that fixes the long standing sdist file permission bug
by ramsdell＠mitre.org 06 Jan '11

06 Jan '11

Please apply the enclosed in: http://www.haskell.org/pipermail/cabal-devel/2010-December/007091.html John

1 0