| Commit message (Collapse) | Author | Age | Files | Lines |
|---|
| | |
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Now that Haddock is moving towards working entirely over `.hi` and `.hie` files,
all declarations and types are going to be synthesized via the `Convert` module.
In preparation for this change, here are a bunch of fixes to this module:
* Add kind annotations to type variables in `forall`'s whose kind is not `Type`,
unless the kind can be inferred from some later use of the variable. See
`implicitForAll` and `noKindTyVars` in particular if you wish to dive into this.
* Properly detect `HsQualTy` in `synifyType`. This is done by following suit with
what GHC's `toIfaceTypeX` does and checking the first argument of
`FunTy{} :: Type` to see if it classified as a given/wanted in the typechecker
(see `isPredTy`).
* Beef up the logic around figuring out when an explicit `forall` is needed. This
includes: observing if any of the type variables will need kind signatures, if the
inferred type variable order _without_ a forall will still match the one GHC
claims, and some other small things.
* Add some (not yet used) functionality for default levity polymorphic type
signatures. This functionality similar to `fprint-explicit-runtime-reps`.
Couple other smaller fixes only worth mentioning:
* Show the family result signature only when it isn't `Type`
* Fix rendering of implicit parameters in the LaTeX and Hoogle backends
* Better handling of the return kind of polykinded H98 data declarations
* Class decls produced by `tyThingToLHsDecl` now contain associated type
defaults and default method signatures when appropriate
* Filter out more `forall`'s in pattern synonyms
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
# Summary
This is a large architectural change to the Hyperlinker.
* extract link (and now also type) information from `.hie` instead
of doing ad-hoc SYB traversals of the `RenamedSource`. Also
adds a superb type-on-hover feature (#715).
* re-engineer the lexer to avoid needless string conversions. By going
directly through GHC's `P` monad and taking bytestring slices, we
avoid a ton of allocation and have better handling of position
pragmas and CPP.
In terms of performance, the Haddock side of things has gotten _much_
more efficient. Unfortunately, much of this is cancelled out by the
increased GHC workload for generating `.hie` files. For the full set of
boot libs (including `ghc`-the-library)
* the sum of total time went down by 9-10% overall
* the sum of total allocations went down by 6-7%
# Motivation
Haddock is moving towards working entirely over `.hi` and `.hie` files.
This change means we no longer need the `RenamedSource` from
`TypecheckedModule` (something which is _not_ in `.hi` files).
# Details
Along the way a bunch of things were fixed:
* Cross package (and other) links are now more reliable (#496)
* The lexer tries to recover from errors on every line (instead of at CPP
boundaries)
* `LINE`/`COLUMN` pragmas are taken into account
* filter out zero length tokens before rendering
* avoid recomputing the `ModuleName`-based `SrcMap`
* remove the last use of `Documentation.Haddock.Utf8` (see #998)
* restructure temporary folder logic for `.hi`/`.hie` model
|
| | |
|
| |
|
|
| |
This will make investigation of #979 easier
|
| |
|
|
|
|
|
| |
* remove redundant imports (only brought to light due to recent work for
improving redundant import detection)
* fix a bug that was casuing exports to appear in reverse order
* fix something in haddock-library that prevented compilation on old GHC's
|
| |
|
|
| |
It got introduced in ghc/ghc@ae2c9b40f5b6bf272251d1f4107c60003f541b62.
|
| |
|
|
|
|
|
|
| |
AFAICT this wasn't causing any crashes, but that's mostly because
we happen not to be forcing `pkgStr` when it would diverge. We come
dangerously close to doing that in `ppHtmlIndex`.
Fixes #569.
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
* Revert "Bump GHC version to 8.6"
This was applied to the wrong branch; there's now a `ghc-8.6` branch;
ghc-head is always supposed to point to GHC HEAD, i.e. an odd major version.
The next version bump to `ghc-head` is supposed to go from e.g. 8.5 to 8.7
This reverts commit 5e3cf5d8868323079ff5494a8225b0467404a5d1.
* README updates (#856)
* README: Remove mentions of master branch
* README: Add instructions for using html-test
* README: Change command to run _all_ the testsuites
* README: Add project overview section
(cherry picked from commit 61d6f935da97eb96685f07bf385102c2dbc2a33c)
* Export more fixities for Hoogle (#871)
This exports fixities for more things, including class methods and
type-level operators.
(cherry picked from commit 88316b972e3d47197b1019111bae0f7f87275fce)
* Avoid line breaks due to line length in Hoogle (#868)
* Avoid line breaks due to line length in Hoogle
Hoogle operates in a line-oriented fashion, so we should avoid ever
breaking due to long lines.
One way of doing this non-intrusively is to modify the 'DynFlags' that
are threaded through the 'Hoogle' module (note this is anyways only
passed through for use in the various 'showSDoc' functions).
* Amend test case
(cherry picked from commit 657b1b3d519545f8d4ca048c06210d6cbf0f0da0)
* tyThingToLHsDecl: Preserve type synonyms that contain a forall (#880)
* tyThingToLHsDecls: Preserve type synonyms that contain a forall
Fixes #879.
* Add Note [Invariant: Never expand type synonyms]
* Clarify Note [Invariant: Never expand type synonyms]
(cherry picked from commit c3eb3f0581f69e816f9453b1747a9f2a3ba02bb9)
* Fix HEAD html-test (#860)
* Update tests for 'StarIsType'
* Accept tests
* Revert "Update tests for 'StarIsType'"
This reverts commit 7f0c01383bbba6dc5af554ee82988d2cf44e407a.
* Refactor handling of parens in types (#874)
* Fix type parenthesization in Hoogle backend
Ported the logic in the HTML and LaTeX backends for adding in parens
into something top-level in 'GhcUtil'. Calling that from the Hoogle
backend fixes #873.
* Remove parenthesizing logic from LaTeX and XHTML backends
Now, the only times that parenthesis in types are added in any backend
is through the explicit 'HsParTy' constructor. Precedence is also
represented as its own datatype.
* List out cases explicitly vs. catch-all
* Fix printing of parens for QuantifiedConstraints
The priority of printing 'forall' types was just one too high.
Fixes #877.
* Accept HTML output for quantified contexts test
* Preserve docs on type family instances (#867)
* Preserve docs on type family instances
The only problem was that the instance location was slightly off
for type family instances.
* Accept output
(cherry picked from commit 133e9c2c168db19c1135479f7ab144c4e33af2a4)
* Fix broken instance source links (#869)
The problem manifests itself in instances that are defined in
modules other than the module where the class is defined. The fix
is just to thread through the 'Module' of the instance further
along.
Since orphan instances appear to already have been working, I didn't
do anything there.
(cherry picked from commit 2de7c2acf9b1ec85b09027a8bb58bf8512e91c05)
* Add some more unicode related tests (#872)
This has been fixed for sure ever since we switched from attoparsec to
parsec. Parts of it may have been working before that, but there was a
point where this would have failed (see #191).
A regression test never hurt anyone. :)
(cherry picked from commit 5ec7715d418bfac0f26aec6039792a99a6e89370)
* Misc tests (#858)
* More tests
* spliced types
* constructor/pattern argument docs
* strictness marks on fields with argument docs
* latex test cases need seperate directory
* Accept tests
* Additional tests for the identifier parser (#816)
* Add tests for the identifier parser
* docs: Clarify how to delimit identifiers
(cherry picked from commit 0861affeca4d72938f05a2eceddfae2c19199071)
|
| | |
|
| |
|
| |
This fixes #836.
|
| | |
|
| |\ |
|
| | | |
|
| | | |
|
| | | |
|
| | | |
|
| | | |
|
| | |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
* Support Haddocks on constructor arguments
This is in conjunction with https://phabricator.haskell.org/D4094.
Adds support for rendering Haddock's on (non-record) constructor
arguments, both for regular and GADT constructors.
* Support haddocks on pattern synonym arguments
It appears that GHC already parsed these - we just weren't using them.
In the process of doing this, I tried to deduplicate some code around
handling patterns.
* Update the markup guide
Add some information about the new support for commenting constructor
arguments, and mention pattern synonyms and GADT-style constructors.
* Overhaul LaTeX support for data/pattern decls
This includes at least
* fixing several bugs that resulted in invalid LaTeX
* fixing GADT data declaration headers
* overhaul handling of record fields
* overhaul handling of GADT constructors
* overhaul handling of bundled patterns
* add support for constructor argument docs
* Support GADT record constructors
This means changes what existing HTML docs look like.
As for LaTeX, looks like GADT records were never even supported. Now they are.
* Clean up code/comments
Made code/comments consistent between the LaTeX and XHTML backend
when possible.
* Update changelog
* Patch post-rebase regressions
* Another post-rebase change
We want return values to be documentable on record GADT constructors.
|
| | |
| |
| |
| | |
This tracks the refactoring of HsDecl.ConDecl.
|
| | |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
* Metadoc stores a package name
This means that '@since' annotations can be package aware.
* Get the package name the right way
This should extract the package name for `@since` annotations the
right way. I had to move `modulePackageInfo` around to do this and,
in the process, I took the liberty to update it.
Since it appears that finding the package name is something that can
fail, I added a warning for this case.
* Silence warnings
* Hide package for local 'since' annotations
As discussed, this is still the usual case (and we should avoid being
noisy for it).
Although this commit is large, it is basically only about threading a
'Maybe Package' from 'Haddock.render' all the way to
'Haddock.Backends.Xhtml.DocMarkup.renderMeta'.
* Bump binary interface version
* Add a '--since-qual' option
This controls when to qualify since annotations with the package they
come from. The default is always, but I've left an 'external' variant
where only those annotations coming from outside of the current
package are qualified.
* Make ParserSpec work
* Make Fixtures work
* Use package name even if package version is not available
The @since stuff needs only the package name passed in, so it
makes sense to not be forced to pass in a version too.
|
| | |
| |
| |
| |
| |
| | |
* extractDecl: Allow extraction of data family instance constructors
* extractDecl: extract data family instance constructors
|
| | |
| |
| |
| | |
Brought back some mistakenly deleted code for handling encoding and eager
reading of files from e0ada1743cb722d2f82498a95b201f3ffb303137.
|
| | | |
|
| | |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
* Add 'show' option to complement 'hide'
The behaviour is for flags passed in the command line to override
flags in file headers. In the command line, later flags override
earlier ones.
Fixes #751 and #266.
* Add a '--show-all' option
|
| | |
| |
| |
| |
| |
| |
| |
| | |
In https://github.com/haskell/haddock/issues/287 we found that
haddock-2.19.0 would miss documentation on class methods with
multiples names.
This patch uses expandSigDecls in a more sensible place.
|
| | | |
|
| | | |
|
| | |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
* Start changing to use GHC lexer
* better cpp
* Change SrcSpan to RealSrcSpan
* Remove error
* Try to stop too many open files
* wip
* wip
* Revert "wip"
This reverts commit b605510a195f26315e3d8ca90e6d95a6737553e1.
Conflicts:
haddock-api/haddock-api.cabal
haddock-api/src/Haddock/Interface.hs
* Remove pointless 'caching'
* Use dlist rather than lists when finding vars
* Use a map rather than list
* Delete bogus comment
* Rebase followup
Things now run using the GHC lexer. There are still
- stray debug statements
- unnecessary changes w.r.t. master
* Cleaned up differences w.r.t. current Haddock HEAD
Things are looking good. quasiquotes in particular look beautiful: the
TH ones (with Haskell source inside) colour/link their contents too!
Haven't yet begun to check for possible performance problems.
* Support CPP and top-level pragmas
The support for these is hackier - but no more hacky than the existing
support.
* Tests pass, CPP is better recognized
The tests were in some cases altered: I consider the new output to be more
correct than the old one....
* Fix shrinking of source without tabs in test
* Replace 'Position'/'Span' with GHC counterparts
Replaces 'Position' -> 'GHC.RealSrcLoc' and 'Span' -> 'GHC.RealSrcSpan'.
* Nits
* Forgot entry in .cabal
* Update changelog
|
| | | |
|
| |/ |
|
| |
|
|
| |
This reverts commit 01eeeb048acd2dd05ff6471ae148a97cf0720547.
|
| |
|
|
| |
This reverts commit 73a26af844ac50b8bec39de11d64452a6286b00c.
|
| |
|
|
| |
This reverts commit 134a7bb054ea730b13c8629a76232d73e3ace049.
|
| | |
|
| | |
|
| | |
|
| |
|
|
|
|
|
| |
... event if they have no documentation (e.g. noDocForDecl)
By using the information in the AvailInfo we don't need additional
export checks.
|
| | |
|
| | |
|
| | |
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
* Use Avails for export resolution
* Support reexported modules
* Factor out availExportItem
* Use avails for fullModuleExports
* Don't use subMap in attachInstances
* lookupDocs without subMap
* Completely remove subMap
* Only calculate unqualified modules when explicit export list is given
* Refactor
* Refine comment
* return
* Fix
* Refactoring
* Split avail if declaration is not exported itself
* Move avail splitting
|
| |
|
|
|
|
| |
As of GHC commit f609374a55bdcf3b79f3a299104767aae2ffbf21 GHC retains the
AvailInfo associated with each IE. @alexbiehl has a patch making proper use of
this change, but this is just to keep things building.
|
| | |
|
| |
|
|
| |
This reverts commit a850ba86d88a4fb9c0bd175453a2580e544e3def.
|
| | |
|
| |\ |
|
| | |
| |
| |
| | |
This allows us to later throw warnings if can't find an identifier
|
| | |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
* Fix pretty-printing of pattern signatures
Pattern synonyms can have up to two contexts, both having a
different semantic meaning: The first holds the constraints
required to perform the matching, the second contains the
constraints provided by a successful pattern match. When the
first context is empty but the second is not it is necessary
to render the first, empty context.
* Generate pattern synonym signatures for ctors exported as patterns
This fixes #653.
* Simplify extractPatternSyn
It is not necessary to generate the simplest type signature since
it will be simplified when pretty-printed.
* Add changelog entries for PR #663
* Fix extractPatternSyn error message
|
| | | |
|
