ocGetNames: sort the acquired symbol table by symbol name, so that
ocLookupSym can then use a fast binary search to find symbols.

Retagged <Literal>s into <Filename>, <Option>, <Command>, &c.

fix more breakage

ordering matters when you're using ':='...

type.c: implement typechecker(POSTPREL), so that initialisation
    of the typechecker is completed correctly in combined mode.
storage.c(addTupleTycon): create a name table entry for () so that
    nameUnit in the above can be bound to something.

import Float# from PrelGHC

set $(HC)=$(GHC_INPLACE)

- set directories using ':=' because the value of $(TOP) can change
- remove some old cruft

A few fixups:

	- correct some addresses

	- update the mail archive pointers to
	  point to the new archives on mail-archive.com

	- use <email> tag for email addresses.

storage.c:   unZcode tuple types (eg Z4T) correctly (off by one)
interface.c(startGHCClass):
              remember to do dictapsToQualtype on class member types
           (processInterfaces): return a Bool if Prelude.hi was
              processed, so we can know when to do everybody(POSTPREL)

Add $(SGML_DOC).tex to $(CLEAN_FILES).

finishGHCModule: correctly create export list entries for non-Names
importEntity: allow TUPLE as a kind of TYCON (am dubious about this)

Add a few more symbols to EXTERN_SYMS to track recent GHC Prelude changes.

Remove _fast from macro name gcdIntegerIntzh_fast.

Update Hugs' knowledge of the GHC Prelude to track recent Prelude changes.

Reinstate debug printing, but only only for combined mode.

Remove #ifdef COMPILER, since the contents of this file is now needed
by the combined system.

Change NUM_TUPLES to 37, to match GHC.

Fix a bug in inlining that gave unresolved references
whenever you compile without -O.  Silly me.

Commenting out the debugging messages for in the new interface code.

Fix more interface-reading bugs:
-- Qualtypes were not being generated from DICTAPs for overloaded
   non-class values.  New function dictapsToQualtype to handle this.
-- Incorrect construction of constructor result type in startGHCDataDecl
   for parameterised types eg   Just :: a -> Maybe a.

Changed meaning of DICTAP, so that the construction is
  ap(DICTAP, (QConId, Type))  rather than
  ap(DICTAP, (QConId, [Type])).  Will have to undo this if we want
                                 to support multiparam type classes.

Add test for unterminated ``.

Emit a reasonable error message instead of crashing when there's an
unterminated literal-liberal in the source file.

Remember to add entities to module(m).names/.tycons/.classes as well as
to module(m).exports.  Otherwise the 'eval environment' will be wrong.

Add assertions in storage.c for addName/addTycon/addClass.

Fix some serious errors in the handling of instances in interfaces.

Changed rules to make SGML documentation using DocBook tools rather than
LinuxDoc.

VSGML files removed because of conversion from LinuxDoc to DocBook (now
have just SGML files).

VSGML files changed back to plain SGML during change from LinuxDoc to DocBook.

Documentation changed from LinuxDoc to DocBook format, and license added to
the user guide.

This commit arranges that literal strings will fuse
nicely, by expressing them as an application of build.

* NoRepStr is now completely redundant, though I havn't removed it yet.

* The unpackStr stuff moves from PrelPack to PrelBase.

* There's a new form of Rule, a BuiltinRule, for rules that
  can't be expressed in Haskell.  The string-fusion rule is one
  such.  It's defined in prelude/PrelRules.lhs.

* PrelRules.lhs also contains a great deal of code that
  implements constant folding.  In due course this will replace
  ConFold.lhs, but for the moment it simply duplicates it.

Many tweaks to expected output

Set MKDEPENDHS to $(HC) instead of $(GHC).
This means that dependencies will be computed with the same
compiler as compilations, which is presumably the right thing.

Clarification on output files

Improve rn029, remove dup rn033

Fix a renamer bug that rejected

	import M hiding( C )

where C is a constructor.

Add rn037

Remove more vestiges of IntAbsOp, and now unused absIntCode.

Add comment about TMPDIR.

remove old unused $RTS_PREFIX I just noticed.

Hopefully banish bogus "Happy version 1.6 or later is required to
compile GHC" messages.  The new rule is: if you don't have an
installed copy of Happy, but have a Happy source tree, we point the
build system at the latter (and pull the version number out of
happy/mk/version.mk).