1. 14 Aug, 2006 1 commit
    • simonpj@microsoft.com's avatar
      Inline into tail-called constructor args · 098d99aa
      simonpj@microsoft.com authored
      Consider
      	x = case y of { True -> (p,q); ... }
      
      The occurrence analyser was marking p,q as 'Many', because they args
      of a constructor in an RhsCtxt.  But actually they aren't in a RhsCtxt,
      and in this case it's better to inline.
      098d99aa
  2. 22 May, 2006 1 commit
    • simonpj@microsoft.com's avatar
      Add idHasRules · b7d8dffa
      simonpj@microsoft.com authored
      Add Id.idHasRules :: Id -> Bool, with the obvious semantics.
      This patch makes sense by itself, but it's just a tidy-up.
      
      b7d8dffa
  3. 07 Apr, 2006 1 commit
    • Simon Marlow's avatar
      Reorganisation of the source tree · 0065d5ab
      Simon Marlow authored
      Most of the other users of the fptools build system have migrated to
      Cabal, and with the move to darcs we can now flatten the source tree
      without losing history, so here goes.
      
      The main change is that the ghc/ subdir is gone, and most of what it
      contained is now at the top level.  The build system now makes no
      pretense at being multi-project, it is just the GHC build system.
      
      No doubt this will break many things, and there will be a period of
      instability while we fix the dependencies.  A straightforward build
      should work, but I haven't yet fixed binary/source distributions.
      Changes to the Building Guide will follow, too.
      0065d5ab
  4. 02 Mar, 2006 1 commit
  5. 01 Mar, 2006 1 commit
  6. 28 Feb, 2006 1 commit
    • simonpj@microsoft.com's avatar
      Simplify the IdInfo before any RHSs · 2317c27b
      simonpj@microsoft.com authored
      	~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
      Simplfy (i.e. substitute) the IdInfo of a recursive group of Ids
      before looking at the RHSs of *any* of them.  That way, the rules
      are available throughout the letrec, which means we don't have to
      be careful about function to put first.
      
      Before, we just simplified the IdInfo of f before looking at f's RHS,
      but that's not so good when f and g both have RULES, and both rules
      mention the other.
      
      This change makes things simpler, but shouldn't change performance.
      2317c27b
  7. 03 Aug, 2005 1 commit
    • simonmar's avatar
      [project @ 2005-08-03 13:53:35 by simonmar] · cfd9e9b3
      simonmar authored
      Patch from SimonPJ (slightly tweaked by me after checking performance
      results):
      
      Fix occasional O(n^2) behaviour in the simplifier.  There was a
      possibility that by inlining a binding, we could re-simplify an
      arbitrary sized expression.  This patch fixes it by moving the
      inlining of arbitrary-sized expressiong to the binding site
      (preInlineUnconditionally), so the decision to inline happens before
      simplifying the RHS.  To do this, we have to collect more information
      during the occurrence analysis phase.
      
      We still make inlining decisions at the call site, but they are always
      size-limited, so we can't get quadratic blowup.
      cfd9e9b3
  8. 19 Jul, 2005 1 commit
    • simonpj's avatar
      [project @ 2005-07-19 16:44:50 by simonpj] · a7ecdf96
      simonpj authored
      WARNING: this is a big commit.  You might want 
      	to wait a few days before updating, in case I've 
      	broken something.
      
      	However, if any of the changes are what you wanted,
      	please check it out and test!
      
      This commit does three main things:
      
      1. A re-organisation of the way that GHC handles bindings in HsSyn.
         This has been a bit of a mess for quite a while.  The key new
         types are
      
      	-- Bindings for a let or where clause
      	data HsLocalBinds id
      	  = HsValBinds (HsValBinds id)
      	  | HsIPBinds  (HsIPBinds id)
      	  | EmptyLocalBinds
      
      	-- Value bindings (not implicit parameters)
      	data HsValBinds id
      	  = ValBindsIn  -- Before typechecking
      		(LHsBinds id) [LSig id]	-- Not dependency analysed
      					-- Recursive by default
      
      	  | ValBindsOut	-- After typechecking
      		[(RecFlag, LHsBinds id)]-- Dependency analysed
      
      2. Implement Mark Jones's idea of increasing polymoprhism
         by using type signatures to cut the strongly-connected components
         of a recursive group.  As a consequence, GHC no longer insists
         on the contexts of the type signatures of a recursive group
         being identical.
      
         This drove a significant change: the renamer no longer does dependency
         analysis.  Instead, it attaches a free-variable set to each binding,
         so that the type checker can do the dep anal.  Reason: the typechecker
         needs to do *two* analyses:
      	one to find the true mutually-recursive groups
      		(which we need so we can build the right CoreSyn)
      	one to find the groups in which to typecheck, taking
      		account of type signatures
      
      3. Implement non-ground SPECIALISE pragmas, as promised, and as
         requested by Remi and Ross.  Certainly, this should fix the 
         current problem with GHC, namely that if you have
      	g :: Eq a => a -> b -> b
         then you can now specialise thus
      	SPECIALISE g :: Int -> b -> b
          (This didn't use to work.)
      
         However, it goes further than that.  For example:
      	f :: (Eq a, Ix b) => a -> b -> b
         then you can make a partial specialisation
      	SPECIALISE f :: (Eq a) => a -> Int -> Int
      
          In principle, you can specialise f to *any* type that is
          "less polymorphic" (in the sense of subsumption) than f's 
          actual type.  Such as
      	SPECIALISE f :: Eq a => [a] -> Int -> Int
          But I haven't tested that.
      
          I implemented this by doing the specialisation in the typechecker
          and desugarer, rather than leaving around the strange SpecPragmaIds,
          for the specialiser to find.  Indeed, SpecPragmaIds have vanished 
          altogether (hooray).
      
          Pragmas in general are handled more tidily.  There's a new
          data type HsBinds.Prag, which lives in an AbsBinds, and carries
          pragma info from the typechecker to the desugarer.
      
      
      Smaller things
      
      - The loop in the renamer goes via RnExpr, instead of RnSource.
        (That makes it more like the type checker.)
      
      - I fixed the thing that was causing 'check_tc' warnings to be 
        emitted.
      a7ecdf96
  9. 28 Apr, 2005 1 commit
    • simonpj's avatar
      [project @ 2005-04-28 10:09:41 by simonpj] · dd313897
      simonpj authored
      This big commit does several things at once (aeroplane hacking)
      which change the format of interface files.  
      
      	So you'll need to recompile your libraries!
      
      1. The "stupid theta" of a newtype declaration
      ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
      Retain the "stupid theta" in a newtype declaration.
      For some reason this was being discarded, and putting it
      back in meant changing TyCon and IfaceSyn slightly.
         
      
      2. Overlap flags travel with the instance
      ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
      Arrange that the ability to support overlap and incoherence
      is a property of the *instance declaration* rather than the
      module that imports the instance decl.  This allows a library
      writer to define overlapping instance decls without the
      library client having to know.  
      
      The implementation is that in an Instance we store the
      overlap flag, and preseve that across interface files
      
      
      3. Nuke the "instnce pool" and "rule pool"
      ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
      A major tidy-up and simplification of the way that instances
      and rules are sucked in from interface files.  Up till now
      an instance decl has been held in a "pool" until its "gates" 
      (a set of Names) are in play, when the instance is typechecked
      and added to the InstEnv in the ExternalPackageState.  
      This is complicated and error-prone; it's easy to suck in 
      too few (and miss an instance) or too many (and thereby be
      forced to suck in its type constructors, etc).
      
      Now, as we load an instance from an interface files, we 
      put it straight in the InstEnv... but the Instance we put in
      the InstEnv has some Names (the "rough-match" names) that 
      can be used on lookup to say "this Instance can't match".
      The detailed dfun is only read lazily, and the rough-match
      thing meansn it is'nt poked on until it has a chance of
      being needed.
      
      This simply continues the successful idea for Ids, whereby
      they are loaded straightaway into the TypeEnv, but their
      TyThing is a lazy thunk, not poked on until the thing is looked
      up.
      
      Just the same idea applies to Rules.
      
      On the way, I made CoreRule and Instance into full-blown records
      with lots of info, with the same kind of key status as TyCon or 
      DataCon or Class.  And got rid of IdCoreRule altogether.   
      It's all much more solid and uniform, but it meant touching
      a *lot* of modules.
      
      
      4. Allow instance decls in hs-boot files
      ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
      Allowing instance decls in hs-boot files is jolly useful, becuase
      in a big mutually-recursive bunch of data types, you want to give
      the instances with the data type declarations.  To achieve this
      
      * The hs-boot file makes a provisional name for the dict-fun, something
        like $fx9.
      
      * When checking the "mother module", we check that the instance
        declarations line up (by type) and generate bindings for the 
        boot dfuns, such as
      	$fx9 = $f2
        where $f2 is the dfun generated by the mother module
      
      * In doing this I decided that it's cleaner to have DFunIds get their
        final External Name at birth.  To do that they need a stable OccName,
        so I have an integer-valued dfun-name-supply in the TcM monad.
        That keeps it simple.
      
      This feature is hardly tested yet.
      
      
      5. Tidy up tidying, and Iface file generation
      ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
      main/TidyPgm now has two entry points:
      
        simpleTidyPgm is for hi-boot files, when typechecking only
        (not yet implemented), and potentially when compiling without -O.
        It ignores the bindings, and generates a nice small TypeEnv.
      
        optTidyPgm is the normal case: compiling with -O.  It generates a
        TypeEnv rich in IdInfo
      
      MkIface.mkIface now only generates a ModIface.  A separate
      procedure, MkIface.writeIfaceFile, writes the file out to disk.
      dd313897
  10. 05 Apr, 2005 1 commit
    • simonmar's avatar
      [project @ 2005-04-05 15:38:01 by simonmar] · 86460846
      simonmar authored
      Instead of gathering a set of 'candidates' in the occurrence
      	analyser, use the isLocalId predicate to identify things
      	for which occurrence information is required.  By defn
      	isLocalId is true of Ids (whether top level or not) defined
      	in this module, and that is exactly what we want.
      
      	The 'candidates set' predated the LocalId invariant, I think.
      86460846
  11. 07 Mar, 2005 1 commit
    • simonpj's avatar
      [project @ 2005-03-07 16:46:08 by simonpj] · 36d22a1c
      simonpj authored
      -----------------------------------------
             Fix a long-standing indirection-zapping bug
      	-----------------------------------------
      
      	Merge to STABLE
      
      Up to now we zap indirections as part of the occurence analyser.
      But this is bogus.  The indirection zapper does the following:
      
      	x_local = <expression>
      	...bindings...
      	x_exported = x_local
      
      where x_exported is exported, and x_local is not, then we
      replace it with this:
      
      	x_exported = <expression>
      	x_local = x_exported
      	...bindings...
      
      But this is plain wrong if x_exported has a RULE that mentions
      something (f, say) in ...bindings.., because 'f' will then die.
      
      After hacking a few solutions, I've eventually simply made the indirection
      zapping into a separate pass (which is cleaner anyway), which wraps the
      entire program back into a single Rec if the bad thing can happen.
      
      On the way I've made indirection-zapping work in Recs too, which wasn't the
      case before.
      
      * Move the zapper from OccurAnal into SimplCore
      * Tidy up the printing of pragmas (PprCore and friends)
      * Add a new function Rules.addRules
      * Merge rules in the indirection zapper (previously one set was discarded)
      36d22a1c
  12. 22 Dec, 2004 1 commit
    • simonpj's avatar
      [project @ 2004-12-22 12:06:13 by simonpj] · d7c402a3
      simonpj authored
      ----------------------------------------
           New Core invariant: keep case alternatives in sorted order
      	----------------------------------------
      
      We now keep the alternatives of a Case in the Core language in sorted
      order.  Sorted, that is,
      	by constructor tag	for DataAlt
      	by literal		for LitAlt
      
      The main reason is that it makes matching and equality testing more robust.
      But in fact some lines of code vanished from SimplUtils.mkAlts.
      
      
      WARNING: no change to interface file formats, but you'll need to recompile
      your libraries so that they generate interface files that respect the
      invariant.
      d7c402a3
  13. 30 Sep, 2004 1 commit
    • simonpj's avatar
      [project @ 2004-09-30 10:35:15 by simonpj] · 23f40f0e
      simonpj authored
      ------------------------------------
      	Add Generalised Algebraic Data Types
      	------------------------------------
      
      This rather big commit adds support for GADTs.  For example,
      
          data Term a where
       	  Lit :: Int -> Term Int
      	  App :: Term (a->b) -> Term a -> Term b
      	  If  :: Term Bool -> Term a -> Term a
      	  ..etc..
      
          eval :: Term a -> a
          eval (Lit i) = i
          eval (App a b) = eval a (eval b)
          eval (If p q r) | eval p    = eval q
          		    | otherwise = eval r
      
      
      Lots and lots of of related changes throughout the compiler to make
      this fit nicely.
      
      One important change, only loosely related to GADTs, is that skolem
      constants in the typechecker are genuinely immutable and constant, so
      we often get better error messages from the type checker.  See
      TcType.TcTyVarDetails.
      
      There's a new module types/Unify.lhs, which has purely-functional
      unification and matching for Type. This is used both in the typechecker
      (for type refinement of GADTs) and in Core Lint (also for type refinement).
      23f40f0e
  14. 02 Apr, 2004 1 commit
    • simonpj's avatar
      [project @ 2004-04-02 13:34:42 by simonpj] · 52276d81
      simonpj authored
      Add a flag -fno-state-hack, which switches off the "state hack".
      
      It's claims that every function over realWorldStatePrimTy is a
      one-shot function.  This is pretty true in practice, and makes a big
      difference.  For example, consider
      	a `thenST` \ r -> ...E...
      The early full laziness pass, if it doesn't know that r is one-shot
      will pull out E (let's say it doesn't mention r) to give
      	let lvl = E in a `thenST` \ r -> ...lvl...
      When `thenST` gets inlined, we end up with
      	let lvl = E in \s -> case a s of (r, s') -> ...lvl...
      and we don't re-inline E.
      52276d81
  15. 17 Dec, 2003 1 commit
    • simonpj's avatar
      [project @ 2003-12-17 11:29:40 by simonpj] · ca0b7c66
      simonpj authored
      -----------------------------------------------------
        Fix a subtle loop in the context-reduction machinery
        ----------------------------------------------------
      
      This bug was provoked by a recent change: when trying to prove
      a constraint C, TcSimplify.reduce now adds C to the database before
      trying to prove C, thus building recursive dictionaries.
      
      Two bugs
      a) If we add C's superclasses (which we were) we can now build a
         bogusly-recursive dictionary (see Note [SUPERCLASS-LOOP]).
         Solution: in reduce, add C only (via addIrred NoSCs) and then
         later use addWanted to add its definition plus SCs.
      
      b) Since we can have recursive definitions, the superclass-loop
         handling machinery (findAllDeps) must carry its visited-set
         with it (which it was not doing before)
      
      
      The main file is TcSimplify; but I modified a bunch of others to
      take advantage of new function extendVarSetList
      ca0b7c66
  16. 12 Feb, 2003 1 commit
    • simonpj's avatar
      [project @ 2003-02-12 15:01:31 by simonpj] · 42b63073
      simonpj authored
      -------------------------------------
        Big upheaval to the way that constructors are named
      	-------------------------------------
      
      This commit enshrines the new story for constructor names.  We could never
      really get External Core to work nicely before, but now it does.
      
      The story is laid out in detail in the Commentary
      	ghc/docs/comm/the-beast/data-types.html
      so I will not repeat it here.
      
      	[Manuel: the commentary isn't being updated, apparently.]
      
      However, the net effect is that in Core and in External Core, contructors look
      like constructors, and the way things are printed is all consistent.
      
      It is a fairly pervasive change (which is why it has been so long postponed),
      but I hope the question is now finally closed.
      
      All the libraries compile etc, and I've run many tests, but doubtless there will
      be some dark corners.
      42b63073
  17. 20 Nov, 2002 1 commit
    • simonpj's avatar
      [project @ 2002-11-20 15:39:47 by simonpj] · 916abd02
      simonpj authored
      ------------------------------------
      	Improve occurrence analysis a little
      	------------------------------------
      
      Consider
      
      		x1 = a0 : []
      		x2 = a1 : x1
      		x3 = a2 : x2
      		g  = f x3
      
      First time round, it looks as if x1 and x2 occur as an arg of a
      let(rec)-bound constructor, and hence should not be inlined. (If the
      RHS of a let is just (C a b) where C is a constructor, then we like to
      keep it that way, with atomic a,b, so that it can be inlined easily at
      a 'case'.)
      
      But in this case, x3 is inlined in g's RHS... and now x2 is not an arg
      of a let-bound constructor, so it can be inlined, and then x1.   Result:
      
      		g = f (a2 : a1 : a0 : [])
      
      Which is fine.  What is *not* fine is that it has been costing us a
      whole simplifier iteration for each element!
      
      This commit adds another little hack to get around the problem: don't treat
      constructor RHSs specially if the bound variable looks as if it occurs just
      once so it'll be inlined.  This catches the common case very nicely.
      
      It's a pain that we have the atomic-args-for-constructor-RHSs invariant.
      But I can't see  how to do without it.
      916abd02
  18. 08 Mar, 2002 1 commit
    • simonpj's avatar
      [project @ 2002-03-08 15:47:18 by simonpj] · 4593b105
      simonpj authored
      ------------------------
      	Kill Type.splitRepFunTys
      	------------------------
      
      splitRepFunTys was a Bad Function that split up a function type
      looking through even recursive newtypes.  Alas, it diverged if
      when we had a recursive newtype with a function whose result was
      the newtype itself.
      
      I've replaced it with ordinary splitFunTys, plus a new function
      Type.dropForAlls, which does what you would expect.
      4593b105
  19. 05 Feb, 2002 1 commit
  20. 29 Jan, 2002 1 commit
    • simonpj's avatar
      [project @ 2002-01-29 09:58:14 by simonpj] · f2054812
      simonpj authored
      ------------
      	Rule phasing
      	------------
      
      This commit adds a little more control to when rules are enabled.
      
        {-# RULES
             "foo" [2] forall ...
             "baz" [~2] forall ...
        #-}
      
      Rule "foo" is active in phase 2 and later.  The new thing is that the
      "~2" means that Rule "baz" is active in phase 3 and earlier.
      (Remember tha phases decrease towards zero.)
      
      All the machinery was there to implement this, it just needed the syntax.
      
      
      Why do this?  Peter Gammie (at UNSW) found that rules weren't firing
      because of bindings of the form
      
      	M.f = f
      	f = ....
      
      where the rules where on the M.f binding.  It turned out that an old
      hack (which have for some time elicited the harmless "shortMeOut" debug
      warnings) prevented this trivial construction from being correctly
      simplified.  The hack in turn derived from a trick in the way the
      foldr/build rule was implemented....and that hack is no longer necessary
      now we can switch rules *off* as well as *on*.
      
      
      There are consequential changes in the Prelude foldr/build RULE stuff.
      It's a clean-up.... Instead of strange definitions like
      	map = mapList
      which we had before, we have an ordinary recursive defn of map, together
      with rules to convert first to foldr/build form, and then (if nothing
      happens) back again.
      
      There's a fairly long comment about the general plan of attack in
      PrelBase, near the defn of map.
      f2054812
  21. 14 Dec, 2001 1 commit
    • simonpj's avatar
      [project @ 2001-12-14 17:24:03 by simonpj] · 5f087cf4
      simonpj authored
      -------------------------
      	Performance tuning things
      	-------------------------
      
      I did some nofib tests, and fixed a number of performance problems.
      
      1.  Things were getting floated to top level, and that prevented
      some useful fusion happening.
      	y = build g
      	x = foldr k z y
      
      Fixed by arranging that we only get really keen on floating to top
      level in the second run of the let-float-out pass.
      
      
      2.  Some fettling up on the let-floater itself.  It had some parameters
      that weren't even being used!  And it was stupidly floating things out
      of a one-shot lambda, and the float-in pass didn't float them back in.
      I think I fixed both of these problems.
      
      
      3.  The eta-reducer was not eta-reducing (/\a -> g a) to g.  In general
      it has to be a bit careful because "seq" means that (\x -> g x) is
      not in general the same as g ---- but it *is* the same for a type lambda.
      
      This turned out to be important in rule matching, where the foldr/build
      rule was not firing because the LHS of the rule looked like
      	foldr k z (/\ a -> g a) = ...
      which never matched!  Result, no fusion to speak of!
      
      
      4.  The simplifier was a bit too gung ho about inlining used-once
      things bound to constructor args.  The comment is with Simplify.simplNonRecX.
      5f087cf4
  22. 31 Oct, 2001 1 commit
    • simonpj's avatar
      [project @ 2001-10-31 15:22:53 by simonpj] · 61bfd5dd
      simonpj authored
      ------------------------------------------
      	Improved handling of scoped type variables
      	------------------------------------------
      
      The main effect of this commit is to allow scoped type variables
      in pattern bindings, thus
      
      	(x::a, y::b) = e
      
      This was illegal, but now it's ok.  a and b have the same scope
      as x and y.
      
      
      On the way I beefed up the info inside a type variable
      (TcType.TyVarDetails; c.f. IdInfo.GlobalIdDetails) which
      helps to improve error messages. Hence the wide ranging changes.
      Pity about the extra loop from Var to TcType, but can't be helped.
      61bfd5dd
  23. 24 Oct, 2001 1 commit
    • simonpj's avatar
      [project @ 2001-10-24 13:46:58 by simonpj] · a2e93ece
      simonpj authored
      ----------------------------------------------
      	Minor fix to occurrence analyer (don't merge)
      	----------------------------------------------
      
      Wasn't dealing right with the case
      
      	x = e
      	y = foldr x
      
      Here we don't want to inline x, because y is a PAP.
      a2e93ece
  24. 01 Oct, 2001 1 commit
  25. 26 Sep, 2001 1 commit
    • simonpj's avatar
      [project @ 2001-09-26 16:19:28 by simonpj] · 6858f7c1
      simonpj authored
      ------------------
      		Simon's big commit
      		------------------
      	[ These files seem to have been left out for some reason ]
      
      
      This commit, which I don't think I can sensibly do piecemeal, consists
      of the things I've been doing recently, mainly directed at making
      Manuel, George, and Marcin happier with RULES.
      
      
      Reogranise the simplifier
      ~~~~~~~~~~~~~~~~~~~~~~~~~
      1. The simplifier's environment is now an explicit parameter.  This
      makes it a bit easier to figure out where it is going.
      
      2. Constructor arguments can now be arbitrary expressions, except
      when the application is the RHS of a let(rec).  This makes it much
      easier to match rules like
      
      	RULES
      	    "foo"  f (h x, g y) = f' x y
      
      In the simplifier, it's Simplify.mkAtomicArgs that ANF-ises a
      constructor application where necessary.  In the occurrence analyser,
      there's a new piece of context info (OccEncl) to say whether a
      constructor app is in a place where it should be in ANF.  (Unless
      it knows this it'll give occurrence info which will inline the
      argument back into the constructor app.)
      
      3. I'm experimenting with doing the "float-past big lambda" transformation
      in the full laziness pass, rather than mixed in with the simplifier (was
      tryRhsTyLam).
      
      4.  Arrange that
      	case (coerce (S,T) (x,y)) of ...
      will simplify.  Previous it didn't.
      A local change to CoreUtils.exprIsConApp_maybe.
      
      5. Do a better job in CoreUtils.exprEtaExpandArity when there's an
      error function in one branch.
      
      
      Phase numbers, RULES, and INLINE pragmas
      ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
      1.  Phase numbers decrease from N towards zero (instead of increasing).
      This makes it easier to add new earlier phases, which is what users want
      to do.
      
      2.  RULES get their own phase number, N, and are disabled in phases before N.
      
      e.g. 	{-# RULES "foo" [2] forall x y.  f (x,y) = f' x y #-}
      
      Note the [2], which says "only active in phase 2 and later".
      
      3.  INLINE and NOINLINE pragmas have a phase number to.  This is now treated
      in just the same way as the phase number on RULE; that is, the Id is not inlined
      in phases earlier than N.  In phase N and later the Id *may* be inlined, and
      here is where INLINE and NOINLINE differ: INLNE makes the RHS look small, so
      as soon as it *may* be inlined it probably *will* be inlined.
      
      The syntax of the phase number on an INLINE/NOINLINE pragma has changed to be
      like the RULES case (i.e. in square brackets).  This should also make sure
      you examine all such phase numbers; many will need to change now the numbering
      is reversed.
      
      Inlining Ids is no longer affected at all by whether the Id appears on the
      LHS of a rule.  Now it's up to the programmer to put a suitable INLINE/NOINLINE
      pragma to stop it being inlined too early.
      
      
      Implementation notes:
      
      *  A new data type, BasicTypes.Activation says when a rule or inline pragma
      is active.   Functions isAlwaysActive, isNeverActive, isActive, do the
      obvious thing (all in BasicTypes).
      
      * Slight change in the SimplifierSwitch data type, which led to a lot of
      simplifier-specific code moving from CmdLineOpts to SimplMonad; a Good Thing.
      
      * The InlinePragma in the IdInfo of an Id is now simply an Activation saying
      when the Id can be inlined.  (It used to be a rather bizarre pair of a
      Bool and a (Maybe Phase), so this is much much easier to understand.)
      
      * The simplifier has a "mode" environment switch, replacing the old
      black list.  Unfortunately the data type decl has to be in
      CmdLineOpts, because it's an argument to the CoreDoSimplify switch
      
          data SimplifierMode = SimplGently | SimplPhase Int
      
      Here "gently" means "no rules, no inlining".   All the crucial
      inlining decisions are now collected together in SimplMonad
      (preInlineUnconditionally, postInlineUnconditionally, activeInline,
      activeRule).
      
      
      Specialisation
      ~~~~~~~~~~~~~~
      1.  Only dictionary *functions* are made INLINE, not dictionaries that
      have no parameters.  (This inline-dictionary-function thing is Marcin's
      idea and I'm still not sure whether it's a good idea.  But it's definitely
      a Bad Idea when there are no arguments.)
      
      2.  Be prepared to specialise an INLINE function: an easy fix in
      Specialise.lhs
      
      But there is still a problem, which is that the INLINE wins
      at the call site, so we don't use the specialised version anyway.
      I'm still unsure whether it makes sense to SPECIALISE something
      you want to INLINE.
      
      
      
      
      
      Random smaller things
      ~~~~~~~~~~~~~~~~~~~~~~
      
      * builtinRules (there was only one, but may be more) in PrelRules are now
        incorporated.   They were being ignored before...
      
      * OrdList.foldOL -->  OrdList.foldrOL, OrdList.foldlOL
      
      * Some tidying up of the tidyOpenTyVar, tidyTyVar functions.  I've
        forgotten exactly what!
      6858f7c1
  26. 14 Sep, 2001 1 commit
    • simonpj's avatar
      [project @ 2001-09-14 15:51:41 by simonpj] · 5ab261bb
      simonpj authored
      --------------------------
      	Add a rule-check pass
      	(special request by Manuel)
      	--------------------------
      
      	DO NOT merge with stable
      
      The flag
      
      	-frule-check foo
      
      will report all sites at which RULES whose name starts with "foo.."
      might apply, but in fact the arguments don't match so the rule
      doesn't apply.
      
      The pass is run right after all the core-to-core passes.  (Next thing
      to do: make the core-to-core script external, so you can fiddle with
      it.  Meanwhile, the core-to-core script is in
      	DriverState.builCoreToDo
      so you can move the CoreDoRuleCheck line around if you want.
      
      The format of the report is experimental: Manuel, feel free to fiddle
      with it.
      
      Most of the code is in specialise/Rules.lhs
      
      
      Incidental changes
      ~~~~~~~~~~~~~~~~~~
      Change BuiltinRule so that the rule name is accessible
      without actually successfully applying the rule.  This
      change affects quite a few files in a trivial way.
      5ab261bb
  27. 27 Aug, 2001 1 commit
  28. 17 Jul, 2001 1 commit
    • simonpj's avatar
      [project @ 2001-07-17 15:28:30 by simonpj] · fdc83001
      simonpj authored
      --------------------------------
      	First cut at the demand analyser
      	--------------------------------
      
      This demand analyser is intended to replace the strictness/absence
      analyser, and the CPR analyser.
      
      This commit adds it to the compiler, but in an entirely non-invasive
      way.
      
      	If you build the compiler without -DDEBUG,
      	you won't get it at all.
      
      	If you build the compiler with -DDEBUG,
      	you'll get the demand analyser, but the existing
      	strictness analyser etc are still there.  All the
      	demand analyser does is to compare its output with
      	the existing stuff and report differences.
      
      There's no cross-module stuff for demand info yet.
      
      The strictness/demand info is put the IdInfo as
      	newStrictnessInfo
      	newDemandInfo
      
      Eventually we'll remove the old ones.
      
      Simon
      fdc83001
  29. 03 May, 2001 1 commit
  30. 30 Jan, 2001 1 commit
  31. 29 Jan, 2001 1 commit
  32. 14 Nov, 2000 1 commit
    • simonpj's avatar
      [project @ 2000-11-14 08:07:11 by simonpj] · db95d6e8
      simonpj authored
      Changing the way we know whether
      something is exported.
      
      THIS COMMIT WON'T EVEN COMPILE
      (I'm doing it to transfer from my laptop.)
      Wait till later today before updating.
      db95d6e8
  33. 23 Oct, 2000 1 commit
  34. 12 Oct, 2000 1 commit
  35. 28 Sep, 2000 1 commit
    • simonpj's avatar
      [project @ 2000-09-28 13:04:14 by simonpj] · 861e836e
      simonpj authored
      ------------------------------------
      	   Mainly PredTypes (28 Sept 00)
      	------------------------------------
      
      Three things in this commit:
      
      	1.  Main thing: tidy up PredTypes
      	2.  Move all Keys into PrelNames
      	3.  Check for unboxed tuples in function args
      
      1. Tidy up PredTypes
      ~~~~~~~~~~~~~~~~~~~~
      The main thing in this commit is to modify the representation of Types
      so that they are a (much) better for the qualified-type world.  This
      should simplify Jeff's life as he proceeds with implicit parameters
      and functional dependencies.  In particular, PredType, introduced by
      Jeff, is now blessed and dignified with a place in TypeRep.lhs:
      
      	data PredType  = Class  Class [Type]
      		       | IParam Name  Type
      
      Consider these examples:
      	f :: (Eq a) => a -> Int
      	g :: (?x :: Int -> Int) => a -> Int
      	h :: (r\l) => {r} => {l::Int | r}
      
      Here the "Eq a" and "?x :: Int -> Int" and "r\l" are all called
      *predicates*, and are represented by a PredType.  (We don't support
      TREX records yet, but the setup is designed to expand to allow them.)
      
      In addition, Type gains an extra constructor:
      
      	data Type = .... | PredTy PredType
      
      so that PredType is injected directly into Type.  So the type
      	p => t
      is represented by
      	PredType p `FunTy` t
      
      I have deleted the hackish IPNote stuff; predicates are dealt with entirely
      through PredTys, not through NoteTy at all.
      
      
      2.  Move Keys into PrelNames
      ~~~~~~~~~~~~~~~~~~~~~~~~~~~~
      This is just a housekeeping operation. I've moved all the pre-assigned Uniques
      (aka Keys) from Unique.lhs into PrelNames.lhs.  I've also moved knowKeyRdrNames
      from PrelInfo down into PrelNames.  This localises in PrelNames lots of stuff
      about predefined names.  Previously one had to alter three files to add one,
      now only one.
      
      3.  Unboxed tuples
      ~~~~~~~~~~~~~~~~~~
      Add a static check for unboxed tuple arguments.  E.g.
      	data T = T (# Int, Int #)
      is illegal
      861e836e
  36. 14 Sep, 2000 1 commit
    • simonpj's avatar
      [project @ 2000-09-14 13:46:39 by simonpj] · cae34044
      simonpj authored
      ---------------------------------------
      	Simon's tuning changes: early Sept 2000
      	---------------------------------------
      
      Library changes
      ~~~~~~~~~~~~~~~
      * Eta expand PrelShow.showLitChar.  It's impossible to compile this well,
        and it makes a big difference to some programs (e.g. gen_regexps)
      
      * Make PrelList.concat into a good producer (in the foldr/build sense)
      
      
      Flag changes
      ~~~~~~~~~~~~
      * Add -ddump-hi-diffs to print out changes in interface files.  Useful
        when watching what the compiler is doing
      
      * Add -funfolding-update-in-place to enable the experimental optimisation
        that makes the inliner a bit keener to inline if it's in the RHS of
        a thunk that might be updated in place.  Sometimes this is a bad idea
        (one example is in spectral/sphere; see notes in nofib/Simon-nofib-notes)
      
      
      Tuning things
      ~~~~~~~~~~~~~
      * Fix a bug in SetLevels.lvlMFE.  (change ctxt_lvl to dest_level)
        I don't think this has any performance effect, but it saves making
        a redundant let-binding that is later eliminated.
      
      * Desugar.dsProgram and DsForeign
        Glom together all the bindings into a single Rec.  Previously the
        bindings generated by 'foreign' declarations were not glommed together, but
        this led to an infelicity (i.e. poorer code than necessary) in the modules
        that actually declare Float and Double (explained a bit more in Desugar.dsProgram)
      
      * OccurAnal.shortMeOut and IdInfo.shortableIdInfo
        Don't do the occurrence analyser's shorting out stuff for things which
        have rules.  Comments near IdInfo.shortableIdInfo.
        This is deeply boring, and mainly to do with making rules work well.
        Maybe rules should have phases attached too....
      
      * CprAnalyse.addIdCprInfo
        Be a bit more willing to add CPR information to thunks;
        in particular, if the strictness analyser has just discovered that this
        is a strict let, then the let-to-case transform will happen, and CPR is fine.
        This made a big difference to PrelBase.modInt, which had something like
      	modInt = \ x -> let r = ... -> I# v in
      			...body strict in r...
        r's RHS isn't a value yet; but modInt returns r in various branches, so
        if r doesn't have the CPR property then neither does modInt
      
      * MkId.mkDataConWrapId
        Arrange that vanilla constructors, like (:) and I#, get unfoldings that are
        just a simple variable $w:, $wI#.  This ensures they'll be inlined even into
        rules etc, which makes matching a bit more reliable.  The downside is that in
        situations like (map (:) xs), we'll end up with (map (\y ys. $w: y ys) xs.
        Which is tiresome but it doesn't happen much.
      
      * SaAbsInt.findStrictness
        Deal with the case where a thing with no arguments is bottom.  This is Good.
        E.g.   module M where { foo = error "help" }
        Suppose we have in another module
      	case M.foo of ...
        Then we'd like to do the case-of-error transform, without inlining foo.
      
      
      Tidying up things
      ~~~~~~~~~~~~~~~~~
      * Reorganised Simplify.completeBinding (again).
      
      * Removed the is_bot field in CoreUnfolding (is_cheap is true if is_bot is!)
        This is just a tidy up
      
      * HsDecls and others
        Remove the NewCon constructor from ConDecl.  It just added code, and nothing else.
        And it led to a bug in MkIface, which though that a newtype decl was always changing!
      
      * IdInfo and many others
        Remove all vestiges of UpdateInfo (hasn't been used for years)
      cae34044
  37. 07 Sep, 2000 1 commit
    • simonpj's avatar
      [project @ 2000-09-07 16:32:23 by simonpj] · 4e6d5798
      simonpj authored
      A list of simplifier-related stuff, triggered
      	by looking at GHC's performance.
      
      	I don't guarantee that this lot will lead to
      	a uniform improvement over 4.08, but it it should
      	be a bit better.  More work probably required.
      
      
      * Make the simplifier's Stop continuation record whether the expression being
        simplified is the RHS of a thunk, or (say) the body of a lambda or case RHS.
        In the thunk case we want to be a bit keener about inlining if the type of
        the thunk is amenable to update in place.
      
      * Fix interestingArg, which was being too liberal, and hence doing
        too much inlining.
      
      * Extended CoreUtils.exprIsCheap to make two more things cheap:
          - 	case (coerce x) of ...
          -   let x = y +# z
        This makes a bit more eta expansion happen.  It was provoked by
        a program of Marcin's.
      
      * MkIface.ifaceBinds.   Make sure that we emit rules for things
        (like class operations) that don't get a top-level binding in the
        interface file.  Previously such rules were silently forgotten.
      
      * Move transformRhs to *after* simplification, which makes it a
        little easier to do, and means that the arity it computes is
        readily available to completeBinding.  This gets much better
        arities.
      
      * Do coerce splitting in completeBinding. This gets good code for
      	newtype CInt = CInt Int
      
      	test:: CInt -> Int
      	test x = case x of
      	      	   1 -> 2
      	      	   2 -> 4
      	      	   3 -> 8
      	      	   4 -> 16
      	      	   _ -> 0
      
      * Modify the meaning of "arity" so that during compilation it means
        "if you apply this function to fewer args, it will do virtually
        no work".   So, for example
      	f = coerce t (\x -> e)
        has arity at least 1.  When a function is exported, it's arity becomes
        the number of exposed, top-level lambdas, which is subtly different.
        But that's ok.
      
        I removed CoreUtils.exprArity altogether: it looked only at the exposed
        lambdas.  Instead, we use exprEtaExpandArity exclusively.
      
        All of this makes I/O programs work much better.
      4e6d5798
  38. 01 Aug, 2000 1 commit
    • simonpj's avatar
      [project @ 2000-08-01 09:08:25 by simonpj] · fe69f3c1
      simonpj authored
      Simon's Marktoberdorf Commits
      
      1.  Tidy up the renaming story for "system binders", such as
      dictionary functions, default methods, constructor workers etc.  These
      are now documented in HsDecls.  The main effect of the change, apart
      from tidying up, is to make the *type-checker* (instead of the
      renamer) generate names for dict-funs and default-methods.  This is
      good because Sergei's generic-class stuff generates new classes at
      typecheck time.
      
      
      2.  Fix the CSE pass so it does not require the no-shadowing invariant.
      Keith discovered that the simplifier occasionally returns a result
      with shadowing.  After much fiddling around (which has improved the
      code in the simplifier a bit) I found that it is nearly impossible to
      arrange that it really does do no-shadowing.  So I gave up and fixed
      the CSE pass (which is the only one to rely on it) instead.
      
      
      3. Fix a performance bug in the simplifier.  The change is in
      SimplUtils.interestingArg.  It computes whether an argment should 
      be considered "interesting"; if a function is applied to an interesting
      argument, we are more likely to inline that function.
      Consider this case
      	let x = 3 in f x
      The 'x' argument was considered "uninteresting" for a silly reason.
      Since x only occurs once, it was unconditionally substituted, but
      interestingArg didn't take account of that case.  Now it does.
      
      I also made interestingArg a bit more liberal.  Let's see if we
      get too much inlining now.
      
      
      4.  In the occurrence analyser, we were choosing a bad loop breaker.
      Here's the comment that's now in OccurAnal.reOrderRec
      
          score ((bndr, rhs), _, _)
      	| exprIsTrivial rhs 	   = 3	-- Practically certain to be inlined
      		-- Used to have also: && not (isExportedId bndr)
      		-- But I found this sometimes cost an extra iteration when we have
      		--	rec { d = (a,b); a = ...df...; b = ...df...; df = d }
      		-- where df is the exported dictionary. Then df makes a really
      		-- bad choice for loop breaker
      
      I also increased the score for bindings with a non-functional type, so that
      dictionaries have a better chance of getting inlined early
      
      
      5. Add a hash code to the InScopeSet (and make it properly abstract)
      This should make uniqAway a lot more robust.  Simple experiments suggest
      that uniqAway no longer gets into the long iteration chains that it used
      to.
      
      
      6.  Fix a bug in the inliner that made the simplifier tend to get into
      a loop where it would keep iterating ("4 iterations, bailing out" message).
      In SimplUtils.mkRhsTyLam we float bindings out past a big lambda, thus:
      	x = /\ b -> let g = \x -> f x x
      		    in E
      becomes
      	g* = /\a -> \x -> f x x
      	x = /\ b -> let g = g* b in E
      	
      It's essential that we don't simply inling g* back into the RHS of g,
      else we will be back to square 1.  The inliner is meant not to do this
      because there's no benefit to the inlining, but the size calculation
      was a little off in CoreUnfold.
      
      
      7.  In SetLevels we were bogus-ly building a Subst with an empty in-scope
      set, so a WARNING popped up when compiling some modules.  (knights/ChessSetList
      was the example that tickled it.)  Now in fact the warning wasn't an error,
      but the Right Thing to do is to carry down a proper Subst in SetLevels, so
      that is what I have now done.  It is very little more expensive.
      fe69f3c1
  39. 14 Jul, 2000 1 commit
    • simonpj's avatar
      [project @ 2000-07-14 08:17:36 by simonpj] · 77a8c0db
      simonpj authored
      This commit completely re-does the kind-inference mechanism.
      Previously it was inter-wound with type inference, but that was
      always hard to understand, and it finally broke when we started
      checking for ambiguity when type-checking a type signature (details
      irrelevant).
      
      So now kind inference is more clearly separated, so that it never
      takes place at the same time as type inference.  The biggest change
      is in TcTyClsDecls, which does the kind inference for a group of
      type and class declarations.  It now contains comments to explain
      how it all works.
      
      There are also comments in TypeRep which describes the slightly
      tricky way in which we deal with the fact that kind 'type' (written
      '*') actually has 'boxed type' and 'unboxed type' as sub-kinds.
      The whole thing is a bit of a hack, because we don't really have 
      sub-kinding, but it's less of a hack than before.
      
      A lot of general tidying up happened at the same time.
      In particular, I removed some dead code here and there
      77a8c0db
  40. 11 Jul, 2000 1 commit