GitLab

* Some preprocessors don't like the C99/C++ '//' comments after a
  directive, so use '/* */' instead. For consistency, a lot of '//' in
  the include files were converted, too.

* UnDOSified libraries/base/cbits/runProcess.c.

* My favourite sport: Killed $Id$s.

SRC_HC_OPTS: only add -ignore-package if supported by underlying GHC.
             (reqd to be able to compile greencard/lib, hdirect/(com)lib with
	     non-cutting edge GHCen.)

Merge to STABLE.

Cleanup the upsweep strategy in GHC.load.

Now it's hopefully clearer how we decide what modules to recompile,
and which are "stable" (not even looked at) during a reload.  See the
comments for details.

Also, I've taken some trouble to explicitly prune out things that
aren't required before a reload, which should reduce the memory
requirements for :reload in GHCi.  Currently I believe it keeps most
of the old program until the reload is complete, now it shouldn't
require any extra memory.

tweaks to a (commented-out) trace message

mkStgAltType: try to find the non-abstract TyCon if the one in the
type of the case binder turns out to be an HiBootTyCon.  Fixes
possible code generation bugs when compiling programs with recursive
modules.

Don't build GHC in all of GhcLibWays, like we do for other libraries

omit __hscore_getFolderPath if __GLASGOW_HASKELL__ < 604

__hscore_getFolderPath is ccall, not stdcall.

Disable SplitObjs when building stage2 (which is also package now, so
was getting the splitting treatment by default).

FPTOOLS_HAPPY, FPTOOLS_ALEX: cope with version.mk:ProjectVersion 'x.y.z' formats (ignoring the 'z' patchlevel). Merge to STABLE (I suppose)

Rename the filenames used for the 'win32' package to {lib}HSwin_32{1.o,2.o,.a}
so as to avoid conflicts with the 'Win32' package when installing to a common
library directory on a case-insensitive filesystem (as we do when building
binary dists under Windows.) This fixes the off-the-bat brokenness experienced
by people trying to load the 'Win32' package with ghci-6.4.

Note: package creators may want to make a note of this if they haven't already,
i.e., while the package infrastructure is case-sensitive, it's unsafe to assume
that the underlying filesystem is.

Merge to STABLE.

Build GHC package as part of stage 2, and install it.

The following changes will affect those building the GHC package:

  - BuildPackageGHC=YES is no longer required in build.mk
  - You must build stage 2 in order to get package ghc.
  - 'make install-inplace-pkg' is not required (nor does it work)
  - -package ghc can be used with the local stage1 or stage2 compiler
    in the current build tree, and it will be available after a
    'make install'.

The GHC package is no longer optional, but it doesn't add much to the
build time.

Fix XML (sorry, Sven)

Windows wibbles

A start on the GHC API:

Flesh out the GHC module so that it can replace CompManager.  Now, the
clients that used CompManager consume the GHC API instead (namely
Main, DriverMkDepend, and InteractiveUI).  Main is significantly
cleaner as a result.

The interface needs more work: in particular, getInfo returns results
in the form of IfaceDecls but we want to use full HsSyn and
Id/DataCon/Class across the boundary instead.

The interfaces for inspecting loaded modules are not yet implemented.

wibble

update list of files to be touched during bootstrapping

remove redundant dependency

Mac OS X: Make bindist work with dynamic libraries

MERGE TO STABLE

__hscore_getFolderPath(): Don't limit ourselves to shell32.dll, look up
  shfolder.dll too.

Merge to STABLE.

Update the -fno-cse flags

Complete the transition of -split-objs into a dynamic flag (looks like I
half-finished it in the last commit).

Also: complete the transition of -tmpdir into a dynamic flag, which
involves some rearrangement of code from SysTools into DynFlags.

Someday, initSysTools should move wholesale into initDynFlags, because
most of the state that it initialises is now part of the DynFlags
structure, and the rest could be moved in easily.

[Windows only]
for System.Directory / Compat.Directory functionality that probes the OS
for local details re: misc user directories, perform late binding of
SHGetFolderPath() from shell32.dll, as it may not be present.
(cf. ghc-6.4's failure to operate on Win9x / NT boxes.) If the API isn't
there, fail with UnsupportedOperation.
Packages.readPackageConfigs: gracefully handle excns from getAppUserDataDirectory.

Merge to STABLE.

Import wibble

Fix stage-2 build

Undo rev. 1.6, these changes shouldn't have been in this tree

Flags cleanup.

Basically the purpose of this commit is to move more of the compiler's
global state into DynFlags, which is moving in the direction we need
to go for the GHC API which can have multiple active sessions
supported by a single GHC instance.

Before:

$ grep 'global_var' */*hs | wc -l
     78

After:

$ grep 'global_var' */*hs | wc -l
     27

Well, it's an improvement.  Most of what's left won't really affect
our ability to host multiple sessions.

Lots of static flags have become dynamic flags (yay!).  Notably lots
of flags that we used to think of as "driver" flags, like -I and -L,
are now dynamic.  The most notable static flags left behind are the
"way" flags, eg. -prof.  It would be nice to fix this, but it isn't
urgent.

On the way, lots of cleanup has happened.  Everything related to
static and dynamic flags lives in StaticFlags and DynFlags
respectively, and they share a common command-line parser library in
CmdLineParser.  The flags related to modes (--makde, --interactive
etc.) are now private to the front end: in fact private to Main
itself, for now.

merge rev. 1.6.2.1, simplified slightly:

  Initialise a CostCentreStack by generating SIZEOF_CostCentreStack
  (gotten from the C compiler) zeros, padded to the nearest word.
  Improves on the previous fixes for unpredictable padding (see comment).

Add SIZEOF_CostCentreStack

Clarify -main-is documentation

Fix a few CCCS that should be W_[CCCS] that I just noticed.  I suspect
that heap-profiling STM code might run into difficulties in 6.4.

Fix +RTS -xc, which is broken in 6.4.

Rearrange the fields of CostCentreStack to leave no (or less) room for
arbitrary padding decisions by C compilers.

Re-clarify hs-boot docs

Follow-up fix to this:
  ghc --make should check for a module graph that
  contains a hs-boot file with no corresponding parent module

Namely, instead of complaining, just add the parent module.

	Merge to STABLE

Clarify hs-boot docs

minor to fix to (still probably broken) stglint

ghc --make should check for a module graph that
  contains a hs-boot file with no corresponding parent module

	Merge to STABLE

Buglet in compiling hs-boot files
	We should make GlobalIds not LocalIds

	Merge to STABLE

Comments