Alexey Dejneka [Wed, 9 Apr 2003 12:42:21 +0000 (12:42 +0000)]
0.pre8.51:
TYPE=-SET uses 3-values logic.
Daniel Barlow [Wed, 9 Apr 2003 01:03:39 +0000 (01:03 +0000)]
0.pre8.50
Chopped out unused (largely unimplemented) stream-command mechanism.
Experiemntal "sb-grovel" contrib is a turbo-charged
grovel_headers replacement, which has seen use in
sb-bsd-sockets and is now being made available separately to
see if other packages find it useful too. See sbcl-devel
message "sb-grovel contrib FFI helper code" for more info
William Harold Newman [Tue, 8 Apr 2003 23:56:27 +0000 (23:56 +0000)]
0.pre8.49:
merged antonio martinez' INTERACTIVE-STREAM-P and
*STANDARD-INPUT* patches (sbcl-devel ca. 2003-04-07)
Kevin Rosenberg [Tue, 8 Apr 2003 19:39:47 +0000 (19:39 +0000)]
contrib/sb-aclrepl/repl.lisp [0.pre8.48]
- use prompt function hook rather than calling sb-aclrepl's prompt function
- use reader conditionals for #+sb-thread and #+aclrepl-debug to control
which top-level commands are usable
contrib/sb-aclrepl/inspect.lisp
- Have inspect length now better mimic ACL's inspect length
- Print suspension points and last value of a long component
Kevin Rosenberg [Tue, 8 Apr 2003 10:51:04 +0000 (10:51 +0000)]
More repl/inspector improvements [0.pre8.47]:
Added :processes & :kill commands
Improve :help and :alias output
Add keywords to (setf alias) for description and abbreviation length
William Harold Newman [Tue, 8 Apr 2003 00:36:04 +0000 (00:36 +0000)]
0.pre8.46:
merged Tony Martinez' protection-from-#.*STREAM* patch
(sbcl-devel 2003-04-07)
William Harold Newman [Mon, 7 Apr 2003 23:17:30 +0000 (23:17 +0000)]
0.pre8.45:
new declarations-are-assertions bug in the compiler...
...new BUGS entry
...workaround for symptoms in OPEN and LOAD
Daniel Barlow [Mon, 7 Apr 2003 17:56:16 +0000 (17:56 +0000)]
0.pre8.43
=== Threads merge, 0.1953125 metres ===
Fix embarrassing Linux bug pointed out by azu on #lisp
Daniel Barlow [Mon, 7 Apr 2003 13:16:52 +0000 (13:16 +0000)]
0.pre8.40
=== Threads merge, 0.390625 metres ===
Unithread fixes for all remaining architectures, although I'm
not 100% that BSD ports are all working.
Various 32/64 bit fixes for Alpha
Removed the allocate-16Mb-at-a-time hacks from the Linux port,
substitute with MAP_NORESERVE: CMUCL has been doing this for a
long time already without user complaint. Also the retryable
mmap errors: as there's no logic in the caller to actually
retry, we'd be better off losing if mmap fails.
Kevin Rosenberg [Mon, 7 Apr 2003 05:26:17 +0000 (05:26 +0000)]
Add framework for setting component values in the inspector. Add the ability to destroy threads when using :exit command
Kevin Rosenberg [Sun, 6 Apr 2003 20:19:40 +0000 (20:19 +0000)]
sb-aclrepl update: use *repl-output*,*repl-input* to avoid passing output-stream to user defined aliases. Start of thread support for :exit command
Kevin Rosenberg [Sat, 5 Apr 2003 23:15:38 +0000 (23:15 +0000)]
sb-aclrepl updates
Kevin Rosenberg [Sat, 5 Apr 2003 23:15:12 +0000 (23:15 +0000)]
fix Makefile, improve test-op
Kevin Rosenberg [Sat, 5 Apr 2003 22:10:03 +0000 (22:10 +0000)]
add basic test-op
Kevin Rosenberg [Sat, 5 Apr 2003 20:51:43 +0000 (20:51 +0000)]
Add inspector
Kevin Rosenberg [Sat, 5 Apr 2003 20:51:20 +0000 (20:51 +0000)]
update for sb-aclrepl
Kevin Rosenberg [Sat, 5 Apr 2003 20:50:09 +0000 (20:50 +0000)]
Add sb-aclrepl news
Kevin Rosenberg [Sat, 5 Apr 2003 20:49:44 +0000 (20:49 +0000)]
Add hook for external inspector
Christophe Rhodes [Sat, 5 Apr 2003 13:04:14 +0000 (13:04 +0000)]
0.pre8.39:
=== Threads merge, 0.78125 metres ===
Merge patch (CSR sbcl-devel 2003-04-04) supporting SPARC/SunOS. This version
builds on SPARC/SunOS and x86/Linux in #!-sb-thread mode; it has not been
tested on x86/Linux with threads enabled (and will obviously not run on
SPARC/SunOS with threads, because of the minor matter of various implementation
details such as a garbage collector and multiprocessing primitives).
Daniel Barlow [Fri, 4 Apr 2003 14:46:18 +0000 (14:46 +0000)]
0.pre8.34
=== Threads merge, 1.5625 metres ===
Reimplement CURRENT-THREAD-OFFSET-SAP using aliens on non-x86
targets - should make other non-x86 targets easier to fix.
Immediate effect is that control stack exhaustion checking now
works on PPC.
Christophe Rhodes [Fri, 4 Apr 2003 14:24:51 +0000 (14:24 +0000)]
0.pre8.37:
Fix forward-referenced DEFCLASS bug (reported by Paul Dietz on
cmucl-imp 2003-04-02)
... pass initargs to CHANGE-CLASS (thanks to Gerd Moellmann)
Christophe Rhodes [Fri, 4 Apr 2003 13:08:07 +0000 (13:08 +0000)]
0.pre8.36:
=== Threads merge, 3.125 metres ===
Reenable control stack guard page checking.
Define control_stack_end as an exclusive bound.
Christophe Rhodes [Fri, 4 Apr 2003 10:23:45 +0000 (10:23 +0000)]
0.pre8.35:
Just sync my BUGS file
Daniel Barlow [Thu, 3 Apr 2003 23:48:32 +0000 (23:48 +0000)]
0.pre8.34
=== Threads merge, 6.25 metres ===
*CONTROL-STACK-END* should be declared as special in early-impl
Daniel Barlow [Thu, 3 Apr 2003 18:26:59 +0000 (18:26 +0000)]
0.pre8.33
=== Threads merge, 12.5 metres ===
Added ppc vop for CURRENT-THREAD-OFFSET-SAP : note that other
ports (alpha, sparc, etc) wil need this too
Fix the PPC symbol hashing problem with symbol names < 4
characters long (thanks CSR)
Replace CONTROL_STACK_FOO with thread-> or SymbolValue as
appropriate in various non-x86 places (cheneygc, ldb backtrace
etc)
get_spinlock defn moved into $arch-arch.c, as unlikely to vary
much between one OS and another on a given arch. Other arches
need to add this too, but for non-threaded ports the dummy
version in ppc-arch.c will do fine.
Stub arch_os_get_current_thread, arch_os_thread_{init,cleanup}
added for Linux/PPC: needs adding to other ports
Add missing UNIX-SETSID definition, used in MAKE-LISTENER-THREAD
Daniel Barlow [Thu, 3 Apr 2003 15:33:21 +0000 (15:33 +0000)]
0.pre8.32
=== Threads merge, 25 metres ===
struct thread now contains a control_stack_end element,
eventually allowing for different threads to have different
stack sizes
Alexey Dejneka [Thu, 3 Apr 2003 09:09:39 +0000 (09:09 +0000)]
0.pre8.31:
When EXTRACT-FUN-ARGS replaces inner function with the dummy
LIST, mark the inner combination as known.
Christophe Rhodes [Wed, 2 Apr 2003 16:08:25 +0000 (16:08 +0000)]
0.pre8.30:
Fix the PCL fix, *sigh*
... don't error if we can't find the class at macroexpand time,
but emit a diagnostic instead. (caught by McCLIM code)
Also enhancement to make-target-contrib.sh
... allow contribs to depend asdfwise on each other
... (side effect: now ASDF-hooked REQUIRE works on fresh-built
uninstalled SBCL with the command line:
SBCL_HOME=`pwd`/contrib ./src/runtime/sbcl --core output/sbcl.core
)
Daniel Barlow [Wed, 2 Apr 2003 14:01:38 +0000 (14:01 +0000)]
new
Daniel Barlow [Wed, 2 Apr 2003 13:50:00 +0000 (13:50 +0000)]
new
Daniel Barlow [Wed, 2 Apr 2003 12:46:19 +0000 (12:46 +0000)]
0.pre8.29
=== Threads merge, 50 metres ===
Thread-related cleanups: now builds threaded SBCL.
... use %fs not %gs throughout, in the vain hope that NPTL will be less
likely to break on it
... various thinkos
Daniel Barlow [Wed, 2 Apr 2003 11:15:10 +0000 (11:15 +0000)]
0.pre8.28
=== Threads merge, 100 metres ===
This is the first commit of experimental native threads for
SBCL. Note that thread support is by default not compiled in
- you need to add :sb-thread to target features. Note also
that non-x86 probably doesn't build in this version - that
will be fixed imminently
See log messages for dan_native_threads_branch,
dan_native_threads_2_branch, dan_native_threads_3_branch for
more information. I'm not going to type it all in again
Christophe Rhodes [Wed, 2 Apr 2003 11:06:48 +0000 (11:06 +0000)]
0.pre8.27:
Fix bug 20, cause it's easy now
... mostly done anyway by SB-PCL:CLASS -> CL:CLASS
... minor adjustment in logic for type declarations in methods
Alexey Dejneka [Wed, 2 Apr 2003 10:20:21 +0000 (10:20 +0000)]
0.pre8.26:
* Fixed bug reported by Robert E. Brown on sbcl-devel
2003-01-21 (optimization issue 10)
** DERIVE-NODE-TYPE replaces references to lambda variables
of type (EQL x) with a constant reference to x;
** UNLINK-NODE checks for a possibility of IF optimization;
Christophe Rhodes [Tue, 1 Apr 2003 14:14:09 +0000 (14:14 +0000)]
0.pre8.25:
Quieten the cross-compiler (and incidentally the target) some
more:
... in DEFTRANSFORM %DATA-VECTOR-AND-INDEX, name the variables
%ARRAY and %INDEX to avoid naming collisions that prompt
spurious code deletion notes
... in NUMBER-PSXHASH, don't do SHORT-FLOAT case, and LONG-FLOAT
only for #!+LONG-FLOAT
Christophe Rhodes [Tue, 1 Apr 2003 13:18:18 +0000 (13:18 +0000)]
0.pre8.24:
Quieten the cross-compiler a little
... SB!XC:PROCLAIM the forthcoming existence of a bunch of
internal defining functions;
... move PROPER-LIST-OF-LENGTH-P earlier, both for quietness
reasons and for use within early-extensions;
... delete some outdated commentary;
... declare *TYPE-SYSTEM-INITIALIZED* special in target-type,
since that's its first use.
Bye bye BUG 6 (and a little more tidying up of BUGS)
Alexey Dejneka [Tue, 1 Apr 2003 08:47:24 +0000 (08:47 +0000)]
0.pre8.22:
Add call for reoptimization after MV-BIND -> LET conversion.
Christophe Rhodes [Mon, 31 Mar 2003 15:20:41 +0000 (15:20 +0000)]
0.pre8.21:
Merge patch from Gerd Moellmann (sbcl-devel 2003-03-29) to make
FIND-CLASS (SB-PCL:FIND-CLASS as was) available earlier
in the MOP methods (requested by KMR sbcl-devel
2003-03-28)
... (SETF FIND-CLASS) in UPDATE-CLASS.
Christophe Rhodes [Mon, 31 Mar 2003 14:20:40 +0000 (14:20 +0000)]
0.pre8.20:
Delete long-standing duplicate code in debug-info (which was
copied, not moved, in sparc port way back when)
Christophe Rhodes [Mon, 31 Mar 2003 12:19:25 +0000 (12:19 +0000)]
0.pre8.19:
Export those classes which are present in our implementation of
the MOP from SB-MOP
... also document absence of METAOBJECT and disjointness of
STANDARD-OBJECT and FUNCALLABLE-STANDARD-OBJECT
Christophe Rhodes [Mon, 31 Mar 2003 10:34:21 +0000 (10:34 +0000)]
0.pre8.18:
Fix a couple of condition system bugs from the test suite
... slot access in the presence of multiple initargs;
... USE-VALUE and friends in the presence of multiple restarts
of the same name where some are associated with other
conditions.
Christophe Rhodes [Sun, 30 Mar 2003 13:54:31 +0000 (13:54 +0000)]
0.pre8.17:
Really actually fix self-build problems
... paper over a bug in our own FIND-CLASS (SB-PCL::FIND-CLASS
as was) by reordering conditions in CROSS-TYPEP. A
search for a proper fix is ongoing
Now we can build from ourselves, absolutely definitely. I've
tested it.
Christophe Rhodes [Sat, 29 Mar 2003 18:51:51 +0000 (18:51 +0000)]
0.pre8.16:
Fix nasty bug introduced in 0.pre8.1 or so, wherein #S no longer
worked. (mea maxima culpa)
... export STRUCTURE-CLASSOID
... write a test for it, so it doesn't happen again
Now we can build from ourselves again, probably :-/
Christophe Rhodes [Sat, 29 Mar 2003 17:10:32 +0000 (17:10 +0000)]
0.pre8.15:
Merge LOAD cleanup from tonyms (sbcl-devel 2003-03-24)
... always load to dynamic space
Christophe Rhodes [Sat, 29 Mar 2003 15:48:26 +0000 (15:48 +0000)]
0.pre8.14:
Merge x86-vm cleanup patch (NJF sbcl-devel 2003-03-23)
... slightly fewer bare constants in the code has to be a good
thing
Christophe Rhodes [Sat, 29 Mar 2003 15:14:28 +0000 (15:14 +0000)]
0.pre8.13:
Make ENSURE-CLASS-USING-CLASS's arguments go the AMOP-specified
way round (noted sbcl-devel 2003-29-03 by CSR)
Christophe Rhodes [Sat, 29 Mar 2003 13:03:39 +0000 (13:03 +0000)]
0.pre8.12:
Merge preliminary Unicode patchlet (Teemu Kalvas "Unicode
preparation" sbcl-devel 2003-03-29)
... SIMPLE-STRING won't necessarily have 8-bit elements in
future, so make explicit SIMPLE-BASE-STRINGs in (some)
places where this matters
Alexey Dejneka [Thu, 27 Mar 2003 07:28:59 +0000 (07:28 +0000)]
0.pre8.11:
Fixed evaluation order in optional entries (reported by
Gilbert Baumann in #lisp 2003-03-26).
Daniel Barlow [Wed, 26 Mar 2003 01:04:41 +0000 (01:04 +0000)]
0.pre8.10
Add some files needed by the soon-to-be-merged thread support
Some non-invasive (supposed not to break unithread builds)
changes to existing files
... add SB-THREAD package
... fs-segment-prefix "instruction" definition
... some documentation
(OK, so it will break McCLIM. That's easily fixable, though)
Christophe Rhodes [Tue, 25 Mar 2003 16:30:56 +0000 (16:30 +0000)]
0.pre8.9:
The other half of the forward-referenced-classes-in-methods
issue (again, thanks to Gerd Moellmann)
... define and use CPL-OR-NIL rather than CLASS-PRECEDENCE-LIST,
to accommodate the fact that some specializers might not
be finalized
Christophe Rhodes [Tue, 25 Mar 2003 15:39:41 +0000 (15:39 +0000)]
0.pre8.8:
Partial bugfix for defining methods where some of the
specializers are forward-referenced-classes
... noted by James Anderson, fixed by Gerd Moellmann
Daniel Barlow [Tue, 25 Mar 2003 13:40:09 +0000 (13:40 +0000)]
0.pre8.7
Implement WITH-TIMEOUT macro, using the SIGALRM handler.
Write each output from TRACE in a single write() call, to make
it much much easier (as in, possible) to see what's happening
when multiple threads are calling TRACEd code at once
Delete some dead code in compiler/x86/macros.lisp
Christophe Rhodes [Tue, 25 Mar 2003 12:34:32 +0000 (12:34 +0000)]
0.pre8.6:
Fix COPY-ALIST bug
... required to signal a type error on dotted lists.
Alexey Dejneka [Tue, 25 Mar 2003 03:45:10 +0000 (03:45 +0000)]
0.pre8.5:
Known functions, which cannot be open coded by backend, are
considered to be able to check types of their
arguments. (reported by Nathan Froyd)
Daniel Barlow [Tue, 25 Mar 2003 01:49:35 +0000 (01:49 +0000)]
0.pre8.4
Change *gc-inhibit* into a counter which increments every time
without-gcing is called. Now we can call without-gcing
recursively or from >1 thread without bad things happening
Daniel Barlow [Tue, 25 Mar 2003 00:29:54 +0000 (00:29 +0000)]
0.pre8.3
contrib fix: sb-bsd-sockets ought to without-gcing in some
places that it wasn't doing. Bug unlikely to be exposed
in #-sb-thread, but showed up testing threaded araneida
Daniel Barlow [Tue, 25 Mar 2003 00:13:10 +0000 (00:13 +0000)]
0.8pre.2
Time to slip my PCL changes in under cover of CSR's patch landing:
Apply patch from Gerd Moellmann to remove cacheing of GF
caches. This makes lots of WITHOUT-INTERRUPTSs in PCL code
go away - a bonus for threading
http://entomotomy.cliki.net/pcl-gf-cache-cacheing-not-needed
Replaced remaining (two) WITHOUT-INTERRUPTSs with WITH-PCL-LOCK
- a new macro that spinlocks around the approriate section.
On unithread SBCL it actually does nothing at all, which I'm
not sure about - we're assuming that WITHOUT-INTERRUPTS has been
used as a substitute for WITHOUT-SCHEDULING. This is likely
but not certain
Christophe Rhodes [Mon, 24 Mar 2003 18:38:56 +0000 (18:38 +0000)]
0.pre8.1
Merge pcl_class_defrobulation_branch
... CL:CLASS is conforming!
... still maybe some breakage around the edges (e.g.
DESCRIBE, DOCUMENTATION)
Christophe Rhodes [Mon, 24 Mar 2003 17:30:09 +0000 (17:30 +0000)]
0.7.13.pcl-class.10
Replace a FIXME with a KLUDGE
... not-quite-parallel code for dealing with (setf
find-classoid) redefinition on host and target
Christophe Rhodes [Mon, 24 Mar 2003 15:11:23 +0000 (15:11 +0000)]
0.7.13.pcl-class.9
Remove one FIXME
... because it's fixed
William Harold Newman [Mon, 24 Mar 2003 14:42:50 +0000 (14:42 +0000)]
0.7.14:
release, tagged as sbcl_0_7_14
Christophe Rhodes [Mon, 24 Mar 2003 09:18:08 +0000 (09:18 +0000)]
0.7.13.35:
Laster minute contrib fixes
... use LENGTH not STRING-LESSP in SB-ACLREPL (from KMR)
... add newline at end of defpackage.lisp
(plus one new BUG)
Christophe Rhodes [Sun, 23 Mar 2003 17:25:06 +0000 (17:25 +0000)]
0.7.13.pcl-class.8
Let CLASS objects be dumpable
... MAKE-LOAD-FORM method from Gerd Moellmann
Daniel Barlow [Sun, 23 Mar 2003 14:21:46 +0000 (14:21 +0000)]
0.7.13.34
Runtime build fixes for Alpha, PPC
Daniel Barlow [Sun, 23 Mar 2003 13:48:28 +0000 (13:48 +0000)]
0.7.13.33
Last-minute contrib fixes ...
patch from KMR to fix history numbers in sb-aclrepl
... and cosmetic patch to sb-aclrepl so that :help command says
how to use the history list anyway
sb-bsd-sockets gets support for the experimental asdf hyperdoc
protocol
Christophe Rhodes [Sat, 22 Mar 2003 12:25:21 +0000 (12:25 +0000)]
0.7.13.pcl-class.7
New SB-MOP package
... use SB-PCL and reexport MOP-defined symbols
... to do that, we need to cope with dependencies in the
package-data structures
... adjust MOP test to suit
Christophe Rhodes [Fri, 21 Mar 2003 18:15:14 +0000 (18:15 +0000)]
0.7.13.pcl-class.6:
Phew, found it. Fix embarrassing thinko in ENSURE-CLASS-VALUES
... continue to pass initargs around
... now we can initialize strucutures again
CONDITIONS can have SB-KERNEL:INSTANCE in their CPL
... so don't remove it
Fix typo in seemingly-dead code in defs.lisp
... s/STRUCTURE/STANDARD/
... probably don't need it any more because a CLASSOID isn't a
type specifier any more, just a type
_Now_ there are no known regressions; tested both against our test suite
and GCL's ansi-tests.
Christophe Rhodes [Fri, 21 Mar 2003 15:31:00 +0000 (15:31 +0000)]
0.7.13.pcl-class.5
Fix up SB-PCL exports
... go through AMOP and make sure they agree, and add basic
consistency check.
Aargh, did I say "no regressions"? It looks like the CONDITION-CLASS
stuff broke defining classes with STRUCTURE-CLASS as a metaclass :-(
Hunting now...
Christophe Rhodes [Fri, 21 Mar 2003 12:09:43 +0000 (12:09 +0000)]
0.7.13.pcl-class.4
Fix CONDITION-CLASS regression from ansi-tests
... many thanks to Gerd Moellmann
... delete dead *FIND-STRUCTURE-CLASS* variable
... some parallel code for CONDITION-CLASS(OID) stuff
... frob ENSURE-CLASS-VALUES slightly to ensure it doesn't
automatically add on :DIRECT-SLOTS
Go back to not printing IDENTITY for named objects
... i.e. #<STANDARD-CLASS FOO> is enough, because there will
only ever (we hope) be one standard-class named FOO.
Christophe Rhodes [Fri, 21 Mar 2003 09:56:11 +0000 (09:56 +0000)]
0.7.13.pcl-class.3:
(CLASS-PROTOTYPE (FIND-CLASS 'NULL)) now returns NIL
... defer fix for other built-in-classes (e.g. BIGNUM) for
mainline post 0.7.14
Christophe Rhodes [Fri, 21 Mar 2003 09:27:12 +0000 (09:27 +0000)]
0.7.13.pcl-class.2
SB-PCL can now USE SB-KERNEL
... and so we can delete a bunch of package prefixes in the
PCL code. Yay!
Christophe Rhodes [Thu, 20 Mar 2003 16:03:39 +0000 (16:03 +0000)]
0.7.13.pcl-class.1
Turn SB-PCL::CLASS into CL:CLASS
... and to do that, turn CL:CLASS into SB-KERNEL:CLASSOID
Well, there's a little more to it than that. This commit causes no
regressions against our own test suite (once the necessary
s/SB-PCL:FIND-CLASS/FIND-CLASS/ changes have been made) but, along with
several new passes in the gcl suite, causes one new failure to do with
condition classes.
There have been some code deletions, too, as some methods that were
necessary to paper over the cracks between the two different CLASSes are
now no longer necessary, as the CLASSOID structure is now viewed as
internal. The major code addition is probably
SB-PCL::SET-CLASS-TYPE-TRANSLATOR, which communicates the necessary
information to the type engine (with extra hair to get BUILT-IN-CLASSES
right).
This branch is expected to last during the freeze period, and land
shortly after 0.7.14 is released.
Daniel Barlow [Wed, 19 Mar 2003 13:15:44 +0000 (13:15 +0000)]
0.7.13.32
Update contrib/asdf to newer upstream version
... fix loopiness
... add test-op definitions
... quote evaluation of *central-registry* components to
make it more useful when a core is dumped containing asdf
Rewrite asdf-using contrib makefiles to use common asdf-module.mk
... and their .asd files to define test-op
Christophe Rhodes [Wed, 19 Mar 2003 09:31:56 +0000 (09:31 +0000)]
0.7.13.31:
Minor build fix for Solaris
... All diffs are not GNU, sadly :-(
Christophe Rhodes [Mon, 17 Mar 2003 17:44:36 +0000 (17:44 +0000)]
0.7.13.30:
Install faster EQUAL on simple-bit-vectors
... word-at-a-time, not bit-at-a-time
Frobs for correctness
... much like one that was solved for 0.7.3.5, we must be
careful about identifying the last word of the bit vector,
particularly for bit-vectors whose length is divisible by
32^Wn-word-bits. Less critical in this case, but we could
still be reading into random space, even if not writing.
Frobs for yet more speed
... allow CMUCL to optimize ASH, as long as none of the values
are in the danger zone. Also reported the bug to CMUCL
people, and it is now fixed, so when all traces of 18d
are removed from this earth, the conditional in
ASH-DERIVE-TYPE-AUX can go too.
Daniel Barlow [Sun, 16 Mar 2003 22:59:18 +0000 (22:59 +0000)]
0.7.13.29
Update contrib/asdf to newer upstream version
Add a 'make up' target to its Makefile so that this can be an
automatic operation in future
Christophe Rhodes [Sat, 15 Mar 2003 19:01:30 +0000 (19:01 +0000)]
0.7.13.28:
Merge SXHASH improvements
... distribute (SXHASH <fixnum>) a little more widely over the
available space;
... make (SXHASH <bit-vector>) consider rather more than just
the first four bits.
Miscellaneous cleanups
... don't delete contrib's html documentation
... make vanilla modules depend on the (newly-built) sbcl.core
... some cleanups in snapshot/SB-SHOW logic
... quit from the low-level debugger now exits the process with
an error code (because, um, even getting to ldb is a
pretty serious error)
Christophe Rhodes [Fri, 14 Mar 2003 16:09:39 +0000 (16:09 +0000)]
0.7.13.27:
Fix typo in new BOUNDP optimization in PCL (thanks to Tony
Martinez for spotting it)
... can't find a test case :-/
Alexey Dejneka [Thu, 13 Mar 2003 10:38:07 +0000 (10:38 +0000)]
0.7.13.26:
SBCL does not ignore type declarations for special
variables. (reported by rif on c.l.l 2003-03-05)
Alexey Dejneka [Wed, 12 Mar 2003 12:44:03 +0000 (12:44 +0000)]
0.7.13.25:
Fixed bug reported by Rolf Wester on comp.lang.lisp:
DEFPARAMETER and DEFVAR now always assign a dynamic variable.
Alexey Dejneka [Tue, 11 Mar 2003 18:24:28 +0000 (18:24 +0000)]
0.7.13.24:
* [N]REVERSE and NRECONC check properness of list arguments;
* [P]SXHASH always consider NIL to be a symbol.
Christophe Rhodes [Tue, 11 Mar 2003 10:14:12 +0000 (10:14 +0000)]
0.7.13.23:
Add forgotten generic/array.lisp holding definitions of (ARRAY
NIL) handling (thanks to APD for noting my oversight)
Christophe Rhodes [Mon, 10 Mar 2003 17:25:11 +0000 (17:25 +0000)]
0.7.13.22:
Minor error fix, from Paul Dietz' suite
... if GETF and GET-PROPERTIES signal a error on malformed
property lists, then make it a type-error, not just a
simple-error
Christophe Rhodes [Mon, 10 Mar 2003 14:54:16 +0000 (14:54 +0000)]
0.7.13.21:
The (ARRAY NIL) has landed.
... implement a SIMPLE-ARRAY-NIL primitive type, including in
garbage collection and (ROOM) logic;
... adjust implementation of array creation and reference to
deal with arrays that can't contain anything;
... (HAIRY-)DATA-VECTOR-REF are can no longer be FLUSHABLE;
Enjoy!
Alexey Dejneka [Sat, 8 Mar 2003 12:02:10 +0000 (12:02 +0000)]
0.7.13.20:
Fixed APPEND.ERROR.1.
Christophe Rhodes [Fri, 7 Mar 2003 15:40:10 +0000 (15:40 +0000)]
0.7.13.19:
Fix build problems for mips(el)
... include some of the structure files from genesis/*.h
... .cvsignore determine-endianness
Christophe Rhodes [Fri, 7 Mar 2003 15:14:51 +0000 (15:14 +0000)]
0.7.13.18:
contrib/ frobs
... install newer asdf, which fixes my stupid thinko for
.sbcl/systems (though in point of fact there would be
an argument, I think, for resolving
user-homedir-pathname/SBCL_HOME later than we are.
... apply KMR patches (sbcl-devel 2003-03-04) for sb-aclrepl
(and turn *CMD-CHAR* into *COMMAND-CHAR* in the process)
Christophe Rhodes [Fri, 7 Mar 2003 12:15:12 +0000 (12:15 +0000)]
0.7.13.17:
Merge mini backend-refactor, motivated by APD's
*CHECK-CONSISTENCY*/non-local-exit observations
... OAOOize MAKE-DYNAMIC-SPACE-TNS VM support routine
Also add tests for documentation in the presence of generalized
function names that should have been merged before but were
forgotten (sorry).
Alexey Dejneka [Thu, 6 Mar 2003 05:47:35 +0000 (05:47 +0000)]
0.7.13.16:
Disable checking of *xxx*-like lexical variable names in
optional- and more-entries (see, e.g., bug 240).
Alexey Dejneka [Wed, 5 Mar 2003 17:30:47 +0000 (17:30 +0000)]
0.7.13.15:
On X86 fixed amount of space for saving dynamic state
corresponding to removing of eval stack saving between
0.7.4.13 and 0.7.4.22. A similar patch probably should be
applied for other platforms.
Alexey Dejneka [Wed, 5 Mar 2003 14:13:05 +0000 (14:13 +0000)]
0.7.13.14:
* New blocks are inserted into the end of component in the
direct order;
* TRANSFORM-CALL inserts new lambda immediately after the
call;
* MAKE-COMPONENT is a BOA-constructor.
Alexey Dejneka [Wed, 5 Mar 2003 06:06:42 +0000 (06:06 +0000)]
0.7.13.13:
* SIGNAL-BOUNDING-INDICES-BAD-ERROR accepts any bounding index
designators;
* fixed CEILING optimization for a divisor of form 2^k.
Christophe Rhodes [Tue, 4 Mar 2003 09:49:22 +0000 (09:49 +0000)]
0.7.13.12:
RDY is a register/register instruction, not a register/immediate
... now disassembly works better
Christophe Rhodes [Mon, 3 Mar 2003 11:16:04 +0000 (11:16 +0000)]
0.7.13.11:
Merge "type system insanity" (CSR sbcl-devel 2002-03-01)
... extend INVOKE-COMPLEX-SUBTYPEP-ARG1-METHOD to take return
values for the case that no next method is found
... define and use (once!) equivalent logic for COMPLEX-=
... be more uncertain on intersections of class types, since we
create them when we don't know enough to canonicalize
... various other cases demand more uncertainty, too (e.g.
intersections involving HAIRY-TYPEs)
No known failures inherent to the type system!
Alexey Dejneka [Mon, 3 Mar 2003 09:16:40 +0000 (09:16 +0000)]
0.7.13.10:
DEFINE-ALIEN-ROUTINE now fully declares result type.
Christophe Rhodes [Fri, 28 Feb 2003 19:26:17 +0000 (19:26 +0000)]
0.7.13.9:
Fix the sparc build
... add some headers to some runtime files
... give COLLECT-GARBAGE its argument in lispland, too
Delete src/runtime/primitive-objects.h, as it's made by
genesis now
Christophe Rhodes [Fri, 28 Feb 2003 15:50:28 +0000 (15:50 +0000)]
0.7.13.8:
Fix some more type system stuff
... the NEGATION type method should test for TYPE=, not EQ, with
*EMPTY-TYPE*
... install some hair to deal with more RATIONAL/INTEGER
confusion
(I can no longer construct an example that causes the type
system to get the answer wrong involving just INTEGER, RATIONAL,
AND, OR and NOT. That's not to say that such an example doesn't
exist, only -- as Fermat didn't quite say -- that my brain is too
small to contain it)
Christophe Rhodes [Thu, 27 Feb 2003 17:20:00 +0000 (17:20 +0000)]
0.7.13.7:
Fix really stupid bug in CONS :SIMPLE-UNION method
... don't mix the CDR type into the CAR type
While I'm there, make the CONS :SIMPLE-UNION method smarter
... canonicalize unions of (CONS A D) with (CONS A' D'), where
A subtypep A', to
(OR (CONS A (OR D D')) (CONS (AND A' (NOT A)) D'))
the point being that this is then in a form that can be further
canonicalized when more CONS types come along. This fixes about
5 bugs from pfdietz's suite.
Daniel Barlow [Thu, 27 Feb 2003 15:19:46 +0000 (15:19 +0000)]
0.7.13.6
Generate primitive-objects.h automaticslly, and move it into
genesis/ (CSR)
Fix typo(!) in gencgc.c that was stopping compilation
Remove linux-specific header files in same that were left over
from the thread branch
Delete some unused variables
Daniel Barlow [Thu, 27 Feb 2003 00:49:57 +0000 (00:49 +0000)]
0.7.13.5
Committed patch for gencgc refactoring based on work done
on threads branch.
... to make header file management a lot simpler, and allow
the use of interesting typedefs in genesis :C-TYPE slot
options, we split runtime.h into lots of smaller files that
can be (semi-)independently included.
... all GC and GCish functions now have the same interface,
so no need for (eq *internal-gc* #'collect-garbage) test in
SUB-GC
... current_region_end_addr and current_region_free_pointer
go away, eliminating potential for weird bugs when they're
not synchronized properly. Yay OAOO
... disabled (actually, removed) inline allocation, as it
depended on old current_region_* (see above) and appears
to make not a lot of actual difference to run times anyway
pseudo-atomic support is now always compiled in. I can see
no good reason for not having it
... much code in alloc() collapsed. Also alloc() no longer
attempts to drop its PA and do a collection in the middle of
allocation - instead it uses the existing maybe_gc flag to
indicate that collection should happen when the allocation is
done. Possibly this has bad effects when trying to allocate
an object bigger than available dynamic space, but that would
fit if a GC were done first. Given the (complete lack of)
error handling for out-of-memory conditions in this and all
previous SBCL versions, it would be a foolish programmer who
was depending on this anyway, though.