sbcl Log


Commit Date  
[6f095a] by Christophe Rhodes Christophe Rhodes

0.8alpha.0.28:
Fix bug 47d (DEFGENERIC must signal PROGRAM-ERROR when
attempting to create a generic function with the same name as a
special operator).
... sounds easy, huh? No.
... make COMPILER-ERROR not inherit from ERROR any more, so that
user handlers don't (wrongly) claim to handle it;
... establish a handler for COMPILER-ERROR around the evaluator
that delegates to the compiler handlers if present, but
handles them itself if not...
... by signalling an error from a new internal restart, to allow
user handlers for ERROR and friends a chance to run.

2003-05-13 13:55:29 Tree
[d76c81] by Christophe Rhodes Christophe Rhodes

0.8alpha.0.21:
The "uh, I thought we had users beta-testing for us" release:
Make SETQ on globals return a value (and IN-PACKAGE, and ...)
... the SET VOP and the %SET-SYMBOL-VALUE IR2 thingy are
different, so one shouldn't translate the other;
... instead, create an IR2 translator for %SET-SYMBOL-VALUE by
hand, mimicking the effect of :SET-TRANS in
DEFINE-PRIMITIVE-OBJECT;
... this removes the need for late-symbol.lisp, so delete it.
While we're in a deleting mood...
... delete src/pcl/fast-init.lisp, which hasn't been part of the
build for the last 5 months or so.

2003-05-08 16:17:57 Tree
[167783] by Christophe Rhodes Christophe Rhodes

0.8alpha.0.16:
Fix for SXHASH on condition objects
... was causing compilation failures when referencing explicit
constant conditions

2003-05-07 11:18:59 Tree
[a95d27] by Christophe Rhodes Christophe Rhodes

0.8alpha.0.11:
Commit laggardly test for 0.8alpha.0.4 breakage

2003-05-04 19:20:33 Tree
[21083e] by Alexey Dejneka Alexey Dejneka

0.8alpha.0.10:
* &ENVIRONMENT argument in macro lambda list is bound first
(found by Paul Dietz);
* Added checking for duplicate variables in macro lambda lists.

2003-05-04 16:52:32 Tree
[670010] by William Harold Newman William Harold Newman

0.8alpha.0.9:
defined WITH-UNIQUE-NAMES
grepped for 'gensym "', and used WITH-UNIQUE-NAMES instead
where it seemed more convenient
tweaked miscellaneous text I noticed in my greppage
added test case for just-fixed compiler bug
bugfix: one last s/layout-class/layout-classoid/

2003-05-03 18:19:43 Tree
[4d8378] by Christophe Rhodes Christophe Rhodes

0.8alpha.0.4:
Canonicalize FUNCTION and VALUES types
... MAKE-FUN-TYPE and MAKE-VALUES-TYPE wrappers
... remove MAKE-VALUES-TYPE cache on grounds of it confusing me
... FUN-TYPE-NARGS needs to check to see if WILD-ARGS is set
[e.g. for LIST, defined as (FUNCTION (&REST T) LIST),
canonicalized to (FUNCTION * LIST)]

Ensure that FUN-TYPES and subtypes of FUNCTION aren't treated as
disjoint
... new type methods for FUNCTION :COMPLEX-{INTERSECT,UN}ION
Make the MAP/%MAP transform arglist look the same as their
DEFKNOWN
... not actually sure if this is necessary any more, but at one
point I got very strange errors without it.

Handle this one with slight care. I'm not 100% convined that
it's bulletproof, though it passes tests and builds itself
without complaint.

2003-05-02 14:56:38 Tree
[2deca0] by Christophe Rhodes Christophe Rhodes

0.8alpha.0.2:
Fix for methods on classes with incomplete subclasses (detected
by KMR/McCLIM compilation)
... distilled test for it.

2003-04-30 19:37:09 Tree
[a110dc] by Christophe Rhodes Christophe Rhodes

0.pre8.116:
Fix for :DEFAULT-INITARGS with side-effects (detected by Paul
Dietz' tests, fix from Gerd Moellmann)
... ctor needed to be smarter about the separation between
locations and initarg equality
Also really add the hyperobject tests (logically part of
sbcl-0.pre8.115)

2003-04-28 12:23:44 Tree
[5d0643] by Christophe Rhodes Christophe Rhodes

0.pre8.100:
As suggested by Fufie (Stig Sandoe) on #lisp IRC, emit a
STYLE-WARNING for slots with names that are STRING= but not EQ.
... test that we can nonetheless still make them, given the
insecurity I feel about my format string wizardry.

2003-04-24 15:22:49 Tree
[fdd6c2] by Christophe Rhodes Christophe Rhodes

0.pre8.99:
One more -0.0/0.0 fix:
... numeric type ranges should be compared with EQUALP, not
EQUAL, to make (FLOAT 0.0) be the same as (FLOAT -0.0)

2003-04-24 08:55:34 Tree
[086927] by Christophe Rhodes Christophe Rhodes

0.pre8.98:
Fix for MEMBER type (as per cmucl-imp 2003-04-23)
... (MEMBER 0.0) is not the same as (SINGLE-FLOAT 0.0 0.0);
... (MEMBER 0.0 -0.0) is the same as (SINGLE-FLOAT 0.0 0.0)
... (NOT (MEMBER 0.0)) needs to be
(OR (NOT SINGLE-FLOAT)
(SINGLE-FLOAT * (0.0))
(MEMBER -0.0)
(SINGLE-FLOAT (0.0)));
... add some tests for this one.
In the process of this fix, make -0.0 and -0.0d0 dumpable by the
cross-compiler:
... more special cases in src/code/cross-float.lisp.
And also let an :SB-SHOW build proceed to the end
... don't try to print the slots of PCL objects.

2003-04-23 17:04:52 Tree
[394ea8] by Christophe Rhodes Christophe Rhodes

0.pre8.91:
Fix for ctor MAKE-INSTANCE optimization
... don't overzealously assert that a given location is unset;
users can supply duplicate initargs.

2003-04-22 15:10:57 Tree
[ee3bfc] by Christophe Rhodes Christophe Rhodes

0.pre8.90:
Make COPY-SYMBOL work on threaded builds
... FAST-SYMBOL-VALUE need not be fast, but must not signal an error
on unbound-symbol
... add a smoke test

2003-04-22 14:58:32 Tree
[08cc6e] by Daniel Barlow Daniel Barlow

0.pre8.87
Bug 247 was fixed in 0.pre8.71: update tests and BUGS file to
match

2003-04-21 23:57:34 Tree
[ce3d6d] by Alexey Dejneka Alexey Dejneka

0.pre8.84:
Fixes in RESTART-CASE, inspired by Paul Dietz' test suit:
* MACROEXPAND requires two arguments;
* DWIM module in RESTART-CASE uses search-by-identity rather
than search-by-name.

2003-04-21 04:37:20 Tree
[c1452f] by Alexey Dejneka Alexey Dejneka

0.pre8.83:
NSET-EXCLUSIVE-OR does not return extra elements when its
arguments contain duplicated elements. (reported by Paul
Dietz)

2003-04-20 16:29:19 Tree
[c3a38a] by Alexey Dejneka Alexey Dejneka

0.pre8.82:
Fixed bugs caught by Paul Dietz' test suite:
* CONVERT-MORE-CALL failed on ((LAMBDA (&KEY) 1)
:ALLOW-OTHER-KEYS T) (fixed by Gerd Moellmann);
* &WHOLE and &REST arguments in a macro lambda list may be
patterns.

2003-04-20 10:53:42 Tree
[015c86] by Christophe Rhodes Christophe Rhodes

0.pre8.74:
Couple the classes and type systems some more
... forward-referenced-classes are now valid types.

Note: this fix follows the cmucl fix perhaps slightly too closely. It
creates CLASSOIDs for forward-referenced-classes slightly eagerly, where
previously no such CLASSOID was generated. This may have some as-yet
unnoticed effect.

2003-04-19 13:14:45 Tree
[ca379a] by William Harold Newman William Harold Newman

0.pre8.68:
Compile (COERCE FOO 'SIMPLE-VECTOR) to reasonably efficient
code (no full call to %TYPEP) even when safe.
made DESCRIBE work on structures again
bug report for ROOM

2003-04-17 22:50:39 Tree
[2e4a94] by Christophe Rhodes Christophe Rhodes

0.pre8.65:
Make COMPUTE-EFFECTIVE-SLOT-DEFINITION (more) AMOP compliant, as
per KMR cmucl-imp 2003-04-12
... write a FIXME in the test, because I don't understand the
required behaviour, but it's better than it was.

2003-04-17 15:33:57 Tree
[00c93e] by Christophe Rhodes Christophe Rhodes

0.pre8.64:
Fix FIND-CLASS of compiled-but-not-loaded structure classes
... slightly sucky hook variable in %TARGET-DEFSTRUCT, to be
used by PCL
... ENSURE-NON-STANDARD-CLASS updated to cope with the
possibility that a CLASSOID exists but the accessor
functions aren't FBOUNDP.
... (side benefit: redefinitions of structures are now reflected
in PCL classes)
... test for FIND-CLASS non-breakage.

2003-04-17 15:04:22 Tree
[d65640] by William Harold Newman William Harold Newman

0.pre8.58:
Do EVAL of one --eval form before READ of the next --eval form.
(to avoid package gotchas reported by dan_b)

2003-04-14 01:42:14 Tree
[508bf1] by Alexey Dejneka Alexey Dejneka

0.pre8.51:
TYPE=-SET uses 3-values logic.

2003-04-09 12:42:21 Tree
[b387f6] by William Harold Newman William Harold Newman

0.pre8.49:
merged antonio martinez' INTERACTIVE-STREAM-P and
*STANDARD-INPUT* patches (sbcl-devel ca. 2003-04-07)

2003-04-08 23:56:27 Tree
Older >