GitLab

Comments only

oops, back out rev. 1.31

awaitEvent(): for Integrity, use socketpair() instead of pipe() to implement socket()-aborting mechanism

Add a (rather vague) FAQ entry about the readline build problem
reported in bug #988585.

back out rev 1.61; to hardwire path to gcc, use --with-gcc=<foo>

tidy up gcc warnings re: %-related format specifiers, mainly rewriting \%%\% to %%%%

Workaround for hbc: It compiles matching against 0-ary constructors in
list comprehensions into wrong code, e.g.

   [ () | True <- [False] ]

results in something like

   Error: No match in I4680Pinteractive

when issued in hbi.

sec-makefile-dependencies: -o option is no longer supported.

Merge to STABLE.

Small syntactic changes to please hbc's interpretation of the layout rule and records

Correction to the construction of .hi filenames in dependency
generation.

Fixes sourceforge bug #978543

Merge to STABLE

Tiny perf improvement

Add FAQ item: how to compile a program with and without profiling in
the same directory.

Remove "EXOTICA" from the description of -osuf/-hisuf

arrow notation fixes (problems reported bu John Hughes):

* allow an infixexp (exp0) to the left of -<.  This adds 4 more
  shift/reduce conflicts, because it makes if/lambda/let/case/proc
  before -< ambiguous.  This is the same sort of thing as
  "if x then y else z + 1", and as there shifting does the right thing.

* described the grammar more precisely

* fixed an example

merge to STABLE

rearrange documentation of SPECIALIZE.

Typo about TH

Clarify words about the -package flag

-----------------------------------------------
       Improve reporting of TH reify out-of-scope errors
	-----------------------------------------------

No change to functionality, just better error reports.

Improve pretty printing slightly

Renamed Opt_WarnMisc to Opt_WarnDodgyImports

Have loadObj ignore requests to load an object more than once, instead
of actually loading the object, then dying.

This change lets user-land loaders safely work inside GHCi, when they
would have no way to know what packages had already been loaded.

Help from SimonM and Andre Pang.

Mention the splitter in the section on porting.

Add linker symbols into linker symbol table.
This lets things that bind to the linker run in GHCi.
Tested on OSX and OpenBSD.

Thanks to Andre Pang.

Tell the linker that we barf in lockFile now. Nuked an evil $Id$ on the way.

Synched "Overflown relocation field" message with reality

Oops, accidentally deleted this.  Reinstate.

Remove hsc2hs files prior to re-grafting the entire directory.

Move Makefiles for nhc98's build here, prior to grafting the directory.

* Fixed ANNOUNCE, LICENSE and README handling
* Made binaries explicit
* Small reformatting

Added empty "install" and "install-docs" targets

* Moved snapshot handling to configure.ac (well, sort of...)
* Updated Requires and BuildRequires
* Synched documentation building/packaging with reality

Advance HEAD version to 6.3.

Added empty "all" and "boot" targets

"make dvi ps html" in $(TOP) builds *all* documentation now

* Added list of current committers
* Slight change of description

More rpm spec file cleanup, including:
* added some BuildRequires
* changed packager to me, so people can complain at the right place :-]
* consistently refer to haskell.org instead of www.haskell.org

ffi-art is dead, Jim...

ANSIfication (from Ralf Wildenhues)

Extremely strange: Some backend emit an ugly trailing ">" when "xref" is an
empty element, but none when only a start tag is used. *sigh* So let's revert
this part of the changes for now...

* Fixed some "screen" vs. "programlisting"

* Don't use leading spaces in verbatim elements just for formatting purposes

* Be extremely cautious with whitespace in verbatim elements to avoid strange
  empty leading/trailing lines. More investigation needed what's really going on
  here, this shouldn't be necessary...