This project is mirrored from https://github.com/haskell/Cabal.
Pull mirroring updated .
- 09 Jan, 2018 3 commits
-
-
Francesco Gazzetta authored
-
Francesco Gazzetta authored
-
Francesco Gazzetta authored
This is especially useful for passing arguments to a single test suite Closes #4684.
-
- 08 Jan, 2018 1 commit
-
-
Moritz Angermann authored
At least it's not just me who is confused by the naming scheme. `project.cabal` seems much more logical to me as it's a `cabal` file. However, cabal expects the file to be called `cabal.project`, so let's make the documentation less confusing.
-
- 07 Jan, 2018 12 commits
-
-
Herbert Valerio Riedel authored
This also experiments with the use of Sphinx/Pygments ABNF syntax highlighter to express syntax/grammars in terms of ABNF code-blocks. [skip ci]
-
gershomb authored
Pass extra dirs to http transport
-
Mikhail Glushenkov authored
Disable per-component build when program coverage enabled
-
gershomb authored
-
gershomb authored
-
gershomb authored
-
Mikhail Glushenkov authored
-
Mikhail Glushenkov authored
-
Creating the test program coverage report requires the test runner to know the build directory for the library component. If per-component builds are enabled, this is not possible because the library is not in-scope at the same time as the test component. The only apparent solution that does not require rethinking per-component builds entirely is to disable it when coverage is enabled. Resolves: #4798
-
-
Mikhail Glushenkov authored
Suppress per-component for build-type:configure
-
Mikhail Glushenkov authored
Remove `UnknownBuildType` from `BuildType`
-
- 06 Jan, 2018 3 commits
-
-
Herbert Valerio Riedel authored
This is a follow-up to #4958 which opened up the opportunity to do this make-illegal-states-unrepresentable refactoring as well.
-
Herbert Valerio Riedel authored
based on a suggestion by @ezyang Workarounds #4548 for now.
-
gershomb authored
-
- 05 Jan, 2018 5 commits
-
-
Mikhail Glushenkov authored
Add cabal-doctest and idris Setup.hs as smoke tests for BC
-
Mikhail Glushenkov authored
-
Mikhail Glushenkov authored
-
Mikhail Glushenkov authored
-
Mikhail Glushenkov authored
-
- 04 Jan, 2018 8 commits
-
-
Mikhail Glushenkov authored
-
Oleg Grenrus authored
-
Oleg Grenrus authored
-
Mikhail Glushenkov authored
See https://github.com/haskell/cabal/pull/4996#issuecomment-355255426. Apparently I forgot to cherry-pick 5797013f from 2.0 to master.
-
Mikhail Glushenkov authored
Restore BC for rewriteFile
-
Mikhail Glushenkov authored
+ a whitespace fix.
-
Edward Z. Yang authored
Signed-off-by:
Edward Z. Yang <ezyang@fb.com>
-
Edward Z. Yang authored
Signed-off-by:
Edward Z. Yang <ezyang@fb.com>
-
- 03 Jan, 2018 8 commits
-
-
Mikhail Glushenkov authored
[ci skip]
-
Mikhail Glushenkov authored
Fix #4986, and extra refactoring
-
Mikhail Glushenkov authored
-
Mikhail Glushenkov authored
Target package names
-
Edward Z. Yang authored
Before: pkgStanzasEnabled is initialized to an empty list and filled in on the first pruning pass. After: pkgStanzasEnabled is initialized to the set of explicitly requested stanzas, and then we add more enabled stanzas as we do "pruning" passes (a bit of a misnomer). Why is it good? Now we always satisfy the invariant that the set of enabled stanzas is a superset of the set of requested stanzas; previously, the invariant was broken up until the point you ran pruning (which, in the case of new-configure, meant that the invariant was always broken!) and we were tripping over this when attempting to render the set of configure flags for plan display. Signed-off-by:
Edward Z. Yang <ezyang@fb.com>
-
Edward Z. Yang authored
Before: In order, the following would be built (use -v for more details): - Cabal-2.1.0.0 (lib) --enable-library-profiling (first run) After: In order, the following would be built (use -v for more details): - Cabal-2.1.0.0 (lib) (first run) Signed-off-by:
Edward Z. Yang <ezyang@fb.com>
-
Edward Z. Yang authored
Signed-off-by:
Edward Z. Yang <ezyang@fb.com>
-
Edward Z. Yang authored
Previously, if you had: build-tool-depends: happy:happy build-tools: happy This would cause an executable dependency to show up twice in exe_deps. It turns out that this made its way all the way to cabal-install's InstallPlan, and then broke an invariant in the install plan execution engine. This is the "correct" fix but the whole arrangement is a little brittle. Signed-off-by:
Edward Z. Yang <ezyang@fb.com>
-