GitLab

Darwin/PPC:
Make StgRunIsImplementedInAssembler non-static; gcc has recently acquired
a habit of dead-stripping it.

MERGE TO STABLE

Default signal handlers weren't being installed; amazing that this has
been broken ever since I rearranged the signal handling code.

- get from the same repo as the main GHC repo, if that was a local filesystem
- allow darcs whatsnew
- use --repodir if possible

The ./darcs-all script at the top level is an easier way to do darcs
pull/push/get on the whole tree (it should probably allow more
commands; I'll fix that later).

libraries/default-packages is a list of darcs repositories with which
to populate the libraries tree.

Adding files from libraries that aren't in the other
packages sub-repos.  I haven't bothered to try to keep
history for these files, for history go back to the CVS
repo.

GHC.runStmt: run the statement in a new thread to insulate the
environment from bad things that the user code might do, such as fork
a thread to send an exception back at a later time.  In order to do
this, we had to keep track of which thread the ^C exception should go
to in a global variable.

Also, bullet-proof the top-level exception handler in GHCi a bit;
there was a small window where an exception could get through, so if
you lean on ^C for a while then press enter you could cause GHCi to
exit.

+RTS -S: replace "collected" with "copied", which is more useful.
+RTS -Dg: print size of mutable list, and breakdown by type of closure
(MUT_VAR, MUT_ARR, others).

fix calculation of copied bytes, we had a words/bytes mismatch when
adding the size of the mutable list

time_str:
  - use ctime_r if available
  - avoid use of strcpy on overlapping regions

Ticket #480 (patch modified by me)

check for ctime_r

Document that source files are ASCII or UTF-8

put unicode keywords under -fglasgow-exts, they aren't Hasell98

MAYBE_GC: we should check alloc_blocks in addition to CurrentNursery,
since some allocateLocal calls don't allocate from the nursery.

fix string desugaring: we can only use the ASCII unpackCString# if all
the chars are <= 0x7F, not <= 0xFF.

(fixes recent breakage in nofib/real/compress2)

Add -threaded when building stage2+ again

understand Unicode lambda as a synonym for \

Fix a comment

Char primops: the Char# rep is wordRep, not I32 (fixed -dcmm-lint
problems on x86_64)

prevChar: don't back up over decoding errors

Add a TODO

reportLexError: don't back up one character, the buffer returned by
Alex is the one *before* the erroneous lexeme started.

Z-encode cost centre symbols when printing them out.

Fix compilation with GHC 6.2.x, hopefully

add file argument to docs for +RTS -t

Initialise part of a structure to NULL to avoid a gcc warning
  I haven't checked the code on this one to be sure this isn't just a bug.

Put entry_is_read_only and tvar_is_locked

Avoid "dereferencing type-punned pointer will break strict-aliasing rules" warning

Avoid "dereferencing type-punned pointer will break strict-aliasing rules" warnings

Remove duplicate imports

Remove dead panic

Put a dummy record initialisation in to fix a warning

Use correct format specifier for a size_t value

Include string.h in unlit, fixing gcc "implicit declaration" warnings

Include stdlib in lndir, fixing gcc "implicit declaration" warnings

Avoid desugaring bug in HEAD (see test ds057).

Fix up to compile with GHC 5.04.x again.

Also includes a fix for a memory error I discovered along the way:
should fix the "scavenge_one" crash in the stage2 build of recent
HEADs.

ord# and chr# should be no-ops, not conversions between wordRep and I32.

Add support for UTF-8 source files

GHC finally has support for full Unicode in source files.  Source
files are now assumed to be UTF-8 encoded, and the full range of
Unicode characters can be used, with classifications recognised using
the implementation from Data.Char.  This incedentally means that only
the stage2 compiler will recognise Unicode in source files, because I
was too lazy to port the unicode classifier code into libcompat.

Additionally, the following synonyms for keywords are now recognised:

  forall symbol 	(U+2200)	forall
  right arrow   	(U+2192)	->
  left arrow   		(U+2190)	<-
  horizontal ellipsis 	(U+22EF)	..

there are probably more things we could add here.

This will break some source files if Latin-1 characters are being used.
In most cases this should result in a UTF-8 decoding error.  Later on
if we want to support more encodings (perhaps with a pragma to specify
the encoding), I plan to do it by recoding into UTF-8 before parsing.

Internally, there were some pretty big changes:

  - FastStrings are now stored in UTF-8

  - Z-encoding has been moved right to the back end.  Previously we
    used to Z-encode every identifier on the way in for simplicity,
    and only decode when we needed to show something to the user.
    Instead, we now keep every string in its UTF-8 encoding, and
    Z-encode right before printing it out.  To avoid Z-encoding the
    same string multiple times, the Z-encoding is cached inside the
    FastString the first time it is requested.

    This speeds up the compiler - I've measured some definite
    improvement in parsing at least, and I expect compilations overall
    to be faster too.  It also cleans up a lot of cruft from the
    OccName interface.  Z-encoding is nicely hidden inside the
    Outputable instance for Names & OccNames now.

  - StringBuffers are UTF-8 too, and are now represented as
    ForeignPtrs.

  - I've put together some test cases, not by any means exhaustive,
    but there are some interesting UTF-8 decoding error cases that
    aren't obvious.  Also, take a look at unicode001.hs for a demo.

Document -Rghc-timing