aboutsummaryrefslogtreecommitdiff
Commit message (Collapse)AuthorAgeFilesLines
...
* Update ANNOUNCEDavid Waern2009-03-211-27/+44
|
* Update CHANGESDavid Waern2009-03-211-0/+48
|
* Define __GHC_PATCHLEVEL__=1 when using ghc-6.10.1David Waern2009-03-211-0/+4
|
* Bump .haddock file version number (due to change of format)David Waern2009-03-211-5/+11
|
* Relax base dependency to < 4.2, not < 4.1Ian Lynagh2009-03-071-1/+1
|
* Install haddock in the new GHC build systemIan Lynagh2009-03-011-5/+19
|
* Do not show package name in warning messageDavid Waern2009-02-281-1/+1
|
* Shorten warning messageDavid Waern2009-02-271-3/+2
|
* Fix error message conflictDavid Waern2009-02-271-5/+4
| | | | | | The module name is already written in the beginning of the message, as seems to be the convention in Haddock. Perhaps not so clear, but we should change it everywhere in that case. Leaving it as it is for now.
* Bug fixDavid Waern2009-02-272-8/+75
| | | | | | | | | | | | | | | | | | We tried to filter out subordinates that were already exported through their parent. This didn't work properly since we were in some cases looking at the grand-parent and not the parent. We now properly compute all the parent-child relations of a declaration, and use this information to get the parent of a subordinate. We also didn't consider record fields with multiple parents. This is now handled correctly. We don't currently support separately exported associated types. But when we do, they should be handled correctly by this process too. Also slightly improved the warning message that we give when filtering out subordinates.
* Add the module to one of haddocks warningsIan Lynagh2009-02-261-1/+1
|
* Hide fynTyConName only for recent GHC versionsDavid Waern2009-02-251-0/+6
|
* 'needsTemplateHaskell' is not defined in current stable GHC.Thomas Schilling2009-02-241-0/+3
|
* Define __GHC_PATCHLEVEL__ for recent version of GHC (stable).Thomas Schilling2009-02-241-4/+2
|
* TAG GHC 6.10.1 releaseIan Lynagh2008-11-070-0/+0
|
* Create inplace/lib/html in the new GHC build systemIan Lynagh2009-02-251-0/+6
|
* Give haddock a wrapper on unix in the new GHC build systemIan Lynagh2009-02-251-8/+14
|
* do .depend generation for haddock with the stage1 compilerIan Lynagh2009-02-131-0/+7
| | | | | | | | | This is a bit of a hack. We mkdepend with stage1 as if .depend depends on the stage2 compiler then make goes wrong: haddock's .depend gets included, which means that make won't reload until it's built, but we can't build it without the stage2 compiler. We therefore build the stage2 compiler before its .depend file is available, and so compilation fails.
* Add a ghc.mkIan Lynagh2009-02-131-0/+12
|
* Don't build the library when building in the GHC treeIan Lynagh2009-02-121-0/+4
|
* Hide funTyConName, now exported by TypeRepsimonpj2009-02-021-1/+1
|
* Fix bad error handling with newer GHCsDavid Waern2009-02-252-27/+27
| | | | | | | | When support for GHC 6.10 was added, an error handler was installed only around the typechecking phase. This had the effect that errors thrown during dependency chasing were caught in the top-level exception handler and not printed with enough detail. With this patch we wrap the error handler around all our usage of the Ghc monad.
* Fix small bugDavid Waern2009-02-241-6/+17
| | | | | | | | The rule is to prefer type constructors to other things when an identifier in a doc string can refer to multiple things. This stopped working with newer GHC versions (due to a tiny change in the GHC renamer). We implement this rule in the HTML backend for now, instead of fixing it in GHC, since we will move renaming of doc strings to Haddock in the future anyway.
* Typo in commentJoachim Breitner2009-02-201-1/+1
|
* Remove html files in the tests/output/ directory which have been ↵David Waern2009-02-0715-6111/+0
| | | | accidentally added
* Update test following change to top level subordinatesDavid Waern2009-02-073-173/+2
|
* Do not show a subordinate at the top level if its parent is also exportedDavid Waern2009-02-071-6/+44
| | | | See note in the source code for more info.
* Correct commentDavid Waern2009-01-241-2/+2
|
* Correct spelling mistake in error messageDavid Waern2009-01-241-1/+1
|
* Filter out separately exported associated types in a smarter wayDavid Waern2009-01-231-17/+19
|
* Improve quasi-quotation test and add reference outputDavid Waern2009-01-223-2/+430
|
* Add test for quasi quotation. No reference output yet.David Waern2009-01-222-0/+45
|
* Fix Trac #50: do not attach docs to pragmas or other kinds of non-declarationsDavid Waern2009-01-221-12/+10
| | | | | We now filter out everything that is not a proper Haskell declaration before collecting the docs and attaching them to declarations.
* Do not indicate that a constructor argument is unboxedDavid Waern2009-01-211-1/+2
| | | | | | We only show the strictness annotation for an unboxed constructor argument. The fact that it is unboxed is an implementation detail and should not be part of the module interface.
* Fix Trac #59: TH-generated declarations disappearingDavid Waern2009-01-201-7/+1
| | | | This patch was contributed by Joachim Breitner (nomeata).
* Document H.I.Create.collectDocs betterDavid Waern2009-01-201-1/+4
|
* Fix the location of INPLACE_PKG_CONF; fixes the buildIan Lynagh2009-01-111-1/+1
| | | | Spotted by Conal Elliott
* Only use needsTemplateHaskell when compiling with GHC 6.10.2 or aboveDavid Waern2009-01-101-0/+4
|
* Fix Trac #68: Turn on compilation via C for Template Haskell packagesDavid Waern2009-01-101-1/+14
| | | | | | | | We can't use HscNothing if we need to run code coming from modules inside the processed package during typechecking, which is the case for some packages using Template Haskell. This could be improved, to e.g. use HscInterpreted and HscNothing where possible, instead of using HscC for all modules in the package.
* Add test for Template Haskell splicingDavid Waern2009-01-104-0/+182
|
* Update tests to account for version number changeDavid Waern2009-01-1033-33/+33
|
* Change version to 2.4.2 - no need to go to 2.5.0David Waern2009-01-103-3/+3
|
* Export modules also when coming from external packagesDavid Waern2009-01-081-4/+15
| | | | | This seems to have regressed since a refactoring that was part of the 2.3.0 release.
* Do not process boot modulesDavid Waern2009-01-061-26/+38
| | | | | | | | | | We should of course not try to produce documentation for boot modules! The reason this has worked in the past is that the output of "real" modules overwrites the output of boot modules later in the process. However, this causes a subtle link environment problem. So let's get rid of this stupid behaviour. We avoid processing boot modules, but we continue to typecheck them.
* Show re-exported names from external packages againDavid Waern2009-01-027-51/+89
| | | | | | | This fixes GHC ticket 2746. In order to also link to the exported subordinate names of a declaration, we need to re-introduce the sub map in the .haddock files.
* Wrap the GHC usage with defaultCleanupHandlerIan Lynagh2008-12-111-5/+6
| | | | This fixes a bug where haddock leaves /tmp/ghc* directories uncleaned.
* Make visible names from ExportItemsDavid Waern2008-12-084-69/+29
| | | | | | | | | | | | | | Instead of a complicated calculation of visible names out of GHC's export items, we can get them straight out of the already calculated ExportItems. The ExportItems should represent exactly those items that are visible in an interface. If store all the exported sub-names in ExportDecl instead of only those with documentation, the calculation becomes very simple. So we do this change as well (should perhaps have been a separate patch). This should fix the problem with names from ghc-prim not appearing in the link environment.
* Try and find a better package name than "main" for Hoogle, goes wrong when ↵Neil Mitchell2008-12-181-1/+2
| | | | working on an executable rather than a library
* Make forall's in constructors explicit, i.e. data Foo = Foo {foo :: Eq a => a}Neil Mitchell2008-12-151-2/+8
|
* Resolve conflict properlyDavid Waern2008-12-071-0/+1
|