GitLab

Import pruning

Add comments

Comments and imports only

closureSatisfiesConstraints: check whether the retainer set is valid
before attempting to match it against a constraint.  It might not be
valid if the object is an ex-weak-pointer which was finalized after
the last GC.

MERGE TO STABLE

-------------------------------------
	   Complete an earlier TH commit
	-------------------------------------

In this earlier commit:

  3.  Ensure that a declaration quotation [d| ... |] does not have a
      permanent effect on the instance environment. (A TH fix.)

I had forgotten to do a stage2 compile.  This commit just completes
the earlier work.  I moved some code from TcExpr (long) to TcSplice
(shorter) which had the happy effect of bringing related code together.

-------------------------------------
   Generate correct dependencies when reading External Core
	-------------------------------------

We have to be more careful than I realised when doing strongly-connected
component analysis of type/class decls when reading External Core.

Here's the relevant new comment:

--		Building edges for SCC analysis
--
-- When building the edges, we treat the 'main name' of the declaration as the
-- key for the node, but when dealing with External Core we may come across
-- references to one of the implicit names for the declaration.  For example:
--	class Eq a where ....
--	data :TSig a = :TSig (:TEq a) ....
-- The first decl is sucked in from an interface file; the second
-- is in an External Core file, generated from a class decl for Sig.
-- We have to recognise that the reference to :TEq represents a
-- dependency on the class Eq declaration, else the SCC stuff won't work right.
--
-- This complication can only happen when consuming an External Core file
--
-- Solution: keep an "EdgeMap" (bad name) that maps :TEq -> Eq.
-- Don't worry about data constructors, because we're only building
-- SCCs for type and class declarations here.  So the tiresome mapping
-- is need only to map   [class tycon -> class]

Part 2 of fix :i in InteractiveUI

Comments only

Another datacon-naming wibble

Fix :i in InteractiveUI

Eliminate brain-dead outputC pattern-match failure

Eliminate bogus string-literal duplication

Compile binary stuff with -O always

Add -fvia-C to utils/Panic_HC_OPTS, and remove it from main/Main_HC_OPTS.

odir-ify prelude/PrimOp.o dependency

Document phase control

-------------------------------------
	 	Two minor wibbles
	-------------------------------------


1.  Make the generic toT/fromT Ids for "generic derived classes" into
    proper ImplicitIds, with their own GlobalIdDetails. This makes it
    easier to identify them.  (The lack of this showed up as a bug
    when I made an apparently-innocuous other change.)

2.  Distinguish ClassOpIds from RecordSelIds in their GlobalIdDetails.
    They are treated differently here and there, so I made this change
    as part of (1)

3.  Ensure that a declaration quotation [d| ... |] does not have a
    permanent effect on the instance environment. (A TH fix.)

Wibbles to the new datacon story; fixes ds002

Add note to description of -no-hs-main: it forces linking in --make
mode too.

If -no-hs-main is specified in --make mode, attempt linking even when
there's no Main module.

Fixes bug #686620

MERGE TO STABLE

Support strictness annotations on data declarations and support the record
and infix constructors. Also tweaked the pretty printer a bit.

Remove un-necessary case

-------------------------------------
	 	Two minor wibbles
	-------------------------------------

[These two unrelated fixes just got tangled together in my tree.]

1.  Fix a crash when a class op is used as a record selector

2.  Fix a wibble related to the new DataCon naming story.
    In tcId, treat the DataCon case entirely separately, because
    its "stupid context" doesn't show up in its type.

    On the way, remove the DataCon cases in tcLookupId and tcLookupGlobalId
    The should not be necessary.  He says hopefully.

Cosmetic wibble

Comments only

Fix build on GHC < 5.00

make use of MBLOCK_ROUND_DOWN()

arenaAlloc(): tidy up alignment calc

More notes about building on Win32

Use hasktags by default

Restore interrupt/quit signal handlers after every evaluation in GHCi,
just in case the program set its own.

Comment change only

GHC is pre-supposed in order to build GHC.  We weren't making this
nearly as explicit as it should have been.

Emit an error if we're trying to build GHC from source without GHC installed.

Not sure if this fix is correct, but at least this module compiles
again...

-------------------------------------
   Do the top-level tcSimpifyTop (to resolve monomorphic constraints)
   once for the whole program, rather than once per splice group
	-------------------------------------

This change makes the trivial program

	main = return ()

work again.  It had stopped working (emitting an error about Monad m
being unconstrained) because the 'checkMain' stuff (which knows special
things about 'main' was happening only *after* all the groups of
decls in the module had been dealt with and zonked (incl tcSimplifyTop).

Better to postpone.  A little more plumbing, but one fewer unexpected
happenings.

Comments

A bit more debug info + comments

A bit more about scoped type variables

wibble