Previously many declarations (e.g. `Int`) were declared to have a "home"
in `Prelude`. However, now Haddock instead chooses to put these in
more specific homes (e.g. `Data.Int`). Given that the "home" decision is
driven by heuristics and in general these changes seem
quite reasonable I am accepting them:

 * `Int` moved from `Prelude` to `Data.Int`
 * `(~)` moved from `Prelude` to `Data.Type.Equality`
 * `Type` moved from `GHC.Types` to `Data.Kind`
 * `Maybe` moved from `Prelude` to `Data.Maybe`
 * `Bool` moved from `Prelude` to `Data.Bool`
 * `Ordering` moved from `Prelude` to `Data.Ord`

As well, more identifiers are now hyperlinked; it's not immediately
clear *why*, but it is an improvement nevertheless.

Here we extend Haddock to admit `@since` annotations in export lists.
These can be attached to most export list items (although not
subordinate lists). These annotations supercede the declaration's
`@since` annotation in produced Haddocks.

Match GHC

- EPA: Comments in AnchorOperation
- EPA: Remove EpaEofComment

This prevents `data FUN` from being shown at the top of the Prelude docs. Fixes
\#23920 on GHC.

Matthew Pickering authored 1 year ago and sheaf committed 1 year ago

When a class method was reexported, it's default methods were also
showing up in the generated html page.

The simplest and most non-invasive fix is to not look for the default
method if we are just exporting the class method.. because the backends
are just showing default methods when the whole class is exported.

In general it would be worthwhile to rewrite this bit of code I think as
the logic and what gets included is split over `lookupDocs` and
`availExportDecl` it would be clearer to combine the two. The result of
lookupDocs is always just passed to availExportDecl so it seems simpler
and more obvious to just write the function directly.

Vladislav Zavialov authored 1 year ago and sheaf committed 1 year ago

The old reasoning no longer applies, nowadays those names can be
mentioned in export lists.

For GHC 9.8.

This commit avoids running into an incomplete record update warning
in the Hoogle backend of Haddock.

This was only noticed now, because incomplete record updates were broken
in GHC 9.6. Now that they are fixed, we have to avoid running into them!

Bartłomiej Cieślar authored 1 year ago and Ben Gamari committed 1 year ago

MR !10283 changes the alternatives for WarningTxt pass. This MR reflects those changes in the haddock codebase.

Signed-off-by: Bartłomiej Cieślar <bcieslar2001@gmail.com>

Associated data families were being given their own export DocStructureItems,
which resulted in them being documented separately from their classes in
haddocks. This commit fixes it.

Finley McIlwaine authored 1 year ago and Ben Gamari committed 1 year ago

Due to ghc!10469.

Finley McIlwaine authored 1 year ago and Ben Gamari committed 1 year ago

The NFData SourceText instance is now available in GHC

Handle mod_iface mi_warns now being IfaceWarnings