Tobias C. Rittweiler [Wed, 16 Dec 2009 09:58:54 +0000 (09:58 +0000)]
1.0.33.13: Catch errors during compile-time-too processing.
* Errors signaled within an (EVAL-WHEN (:COMPILE-TOPLEVEL) ...),
i.e. during ``compile-time-too' processing, are now caught and
reported just like errors at macroexpansion time.
Previously, we landed in the debugger from within the compiler and
thus provided pretty much no information about the actual source
context to the user.
Now, we won't land in the debugger, and the appropriate source
context is reported along the error.
* Some slight generalization in the test suite.
* So we can add tests for errors in EVAL-WHEN and during
macroexpansion.
Rudi Schlatte [Tue, 15 Dec 2009 16:09:02 +0000 (16:09 +0000)]
1.0.33.12: Less ugly html manuals
... add some light css styling
... also create single-file manual for easier grepping
... unilaterally deciding on light blue as our new corporate colour
Gabor Melis [Tue, 15 Dec 2009 15:22:22 +0000 (15:22 +0000)]
1.0.33.11: fix detection of tls exhaustion
... that was added in 1.0.16.21 and whose ways were shown to be wrong
by the additional pseudo atomic asserts added later.
With this fix the user gets to see the tls exhausted message instead
of a random complaint about pa.
Alastair Bridgewater [Sat, 12 Dec 2009 20:55:15 +0000 (20:55 +0000)]
1.0.33.10: Make slam.sh usable to obtain a trace-file post-build.
* Have slam.lisp re-load build-order.lisp-expr in order to obtain new
flags as needed.
* Also check for outdated/non-extant trace file for stems flagged as
requiring one.
Alastair Bridgewater [Sat, 12 Dec 2009 20:52:13 +0000 (20:52 +0000)]
1.0.33.9: LDB stability improvements.
* Make sure that a struct object is actually in the lisp heap space
before attempting to access its memory.
* We have a perfectly good is_lisp_pointer() function, use it in
print_obj().
Alastair Bridgewater [Sat, 12 Dec 2009 20:49:55 +0000 (20:49 +0000)]
1.0.33.8: Fix x86-64 interrupt context display in ldb.
* Half of the register names were missing, causing random failures
from using words past the end of the register name array as string
pointers.
Tobias C. Rittweiler [Thu, 10 Dec 2009 23:32:18 +0000 (23:32 +0000)]
1.0.33.7: Indicate *DEBUGGER-HOOK* / SB-EXT:*INVOKE-DEBUGGER-HOOK* in backtrace.
Gabor Melis [Fri, 4 Dec 2009 17:59:48 +0000 (17:59 +0000)]
1.0.33.6: prettier PRINT-OBJECT default method
If *PRINT-PRETTY*, the default method of PRINT-OBJECT now establishes
a logical block around PRINT-UNREADABLE-OBJECT that, in turn, adds a
few PPRINT-NEWLINES to allow for the stuff between #< and > be broken
into multiple lines as the pretty printer sees fit. This allows #<...>
to be wrapped properly.
https://bugs.launchpad.net/sbcl/+bug/488979
Gabor Melis [Fri, 4 Dec 2009 16:07:38 +0000 (16:07 +0000)]
1.0.33.5: improve DESCRIBE (#lp488976)
Instead of printing the object type and address in an around method on
DESCRIBE-OBJECT, make it a normal method that all other methods are
supposed to call via CALL-NEXT-METHOD if they wish. This makes it
possible to completely change the output of DESCRIBE for one class
while leaving other classes alone.
Also, kill the unused *IN-DESCRIBE* var in the process.
Christophe Rhodes [Wed, 2 Dec 2009 21:07:14 +0000 (21:07 +0000)]
1.0.33.4: rewriting history, the regexp way
The html generation in sbcl-page now supports automatic linking to launchpad
bug IDs mentioned in NEWS, given a certain format. Convert existing
references to launchpad bugs to that format.
Christophe Rhodes [Wed, 2 Dec 2009 08:54:43 +0000 (08:54 +0000)]
1.0.33.3: restore buildability from CMUCL
- Add extra keyword in the ftype declaration for %DEFKNOWN (oh, the irony);
- move the THREAD structure into the cross-compiler, so that MUTEX slots
can be declared to have type (OR NULL THREAD);
- IGNORE on non-binding in saptran
(build failure reported by xme@gmx.net)
Alastair Bridgewater [Sat, 28 Nov 2009 22:53:45 +0000 (22:53 +0000)]
1.0.33.2: Fix bug 489698 (reading #p"\\\\" on windows).
* Actual bug was a missed range check prior to calling POSITION in
EXTRACT-DEVICE in src;code;win32-pathname. Fixed.
* Added test case.
Tobias C. Rittweiler [Sat, 28 Nov 2009 13:25:21 +0000 (13:25 +0000)]
1.0.33.1: DEFTYPE-LAMBDA-LIST for builtin types.
Make SB-INTROSPECT:DEFTYPE-LAMBDA-LIST also work on most builtin
types.
Christophe Rhodes [Fri, 27 Nov 2009 17:59:15 +0000 (17:59 +0000)]
1.0.33: will be tagged as sbcl_1_0_33
Alastair Bridgewater [Fri, 27 Nov 2009 16:59:28 +0000 (16:59 +0000)]
1.0.32.39: Win32 build fix
* Because both sbcl and <rpcndr.h> define a boolean type,
grovel-headers was failing to build on win32. Copied the kludge for
working around this issue from src/runtime/runtime.h to
tools-for-build/grovel-headers.c.
trittweiler [Tue, 24 Nov 2009 16:31:19 +0000 (16:31 +0000)]
1.0.32.38: liberalize WHO-SPECIALIZES-DIRECTLY/GENERALLY
Other Xref functions (WHO-CALLS etc.) are very liberal at what they
accept as input. Adjust the two new Xref functions
WHO-SPECIALIZES-DIRECTLY and WHO-SEPCIALIZES-GENERALLY accordingly.
Christophe Rhodes [Fri, 20 Nov 2009 21:40:04 +0000 (21:40 +0000)]
1.0.32.37: declarations and symbol-macros in methods
A stray backquote caused compilation warnings and run-time errors if, within
a method, a symbol macro with a type declaration was given a new value with
SETQ.
Fixes bug #485019, reported by Iban Hatchondo
Christophe Rhodes [Fri, 20 Nov 2009 21:33:51 +0000 (21:33 +0000)]
1.0.32.36: (subtypep 'symbol 'keyword) must be NIL, T
Special-case the hairy type (SATISFIES KEYWORDP) and its interaction with the
SYMBOL type. (We could potentially be cleverer at this point and additionally
tell the system that all non-symbols are non-(SATISFIES KEYWORDP) types, but
we're somewhat late in the development cycle now)
Fixes bug #485972
Nikodemus Siivola [Fri, 20 Nov 2009 11:48:12 +0000 (11:48 +0000)]
1.0.32.35: SB-POSIX:FORK guards against forking with multiple threads
* It wasn't going to work right in any case, so better to refuse up
front.
* Remove invalid FIXME from globaldb.lisp: DO-INFO is also used by
SB-INTROSPECT.
Nikodemus Siivola [Thu, 19 Nov 2009 16:35:50 +0000 (16:35 +0000)]
1.0.32.34: remove curious-hacker-cruft from constraints.lisp
...that was not supposed to go in!
*blush*
Nikodemus Siivola [Thu, 19 Nov 2009 16:08:37 +0000 (16:08 +0000)]
1.0.32.33: GENCGC tuning parameters
* Define +HIGHEST-NORMAL-GENERATION+ and +PSEUDO-STATIC-GENERATION+
in lisp, and let genesis tell C about them.
* Make various generations[gen].foo tweakable from lisp, give them
nice interface functions and document the same.
* Bits of manual prettification in related parts.
Christophe Rhodes [Thu, 19 Nov 2009 15:38:50 +0000 (15:38 +0000)]
1.0.32.32: add support for x86-64/NetBSD
Patch from Aymeric Vincent.
Christophe Rhodes [Thu, 19 Nov 2009 11:50:42 +0000 (11:50 +0000)]
1.0.32.31: type system now understands (and <array-type> (not simple-array))
Taken to mean the <array-type> with COMPLEXP T (rather than :MAYBE).
Adjust the type test transform to use the old technique for testing for
complex arrays (using an explicit (NOT SIMPLE-ARRAY) test rather than a
full call to %TYPEP, as you would otherwise get; this is a KLUDGE, but
no worse than before).
Include a test case for bug #309129, which this fixes.
Alastair Bridgewater [Thu, 12 Nov 2009 16:07:49 +0000 (16:07 +0000)]
1.0.32.30: Fix breakage from 1.0.32.26.
* Use FIND-CONSTANT instead of MAYBE-EMIT-MAKE-LOAD-FORMS to prepare
arglists for dumping, thus unbreaking compilation to core.
* Add a test case for in-core case.
trittweiler [Thu, 12 Nov 2009 15:10:04 +0000 (15:10 +0000)]
1.0.32.29: Add build flag :sb-xref-for-internals.
Enabling :sb-xref-for-internals in customize-target-features.lisp,
will make Sbcl collect Xref data about itself during the build. This
increases the core size drastically by about 5-6mb, but it's useful
for SBCL developers because they can now use M-? (slime-edit-uses) to
get a list of call/expansion/reference sites for internal stuff.
It may be interesting to Lisp advocacy who can now show off with
finding the use sites of standardized functions like CONS, etc. :-)
Additionally -- regardless of :sb-xref-for-internals --, we now also
collect xref data for keywords because they're "fine" names for
functions and macros, and I know of people who use MACROLET on
keywords for their DSLs.
Christophe Rhodes [Thu, 12 Nov 2009 11:29:23 +0000 (11:29 +0000)]
1.0.32.28: fix listen / read-char-no-hang
Whoops. The external-format changes broke read-char-no-hang after an
unread-char: very noticeable indeed just as soon as an error occurs in
slime. In my defence, this was already broken on bivalent streams; I just
propagated that brokenness to all fd-streams.
Include a run-program-based test, because that produces the kind of streams
that can have no data and yet not be at EOF. In the slime context, they're
sockets, so it's a bit difficult to test the slime case exactly; ideas for
good simulations welcome.
trittweiler [Thu, 12 Nov 2009 10:42:50 +0000 (10:42 +0000)]
1.0.32.27: Add test case for 1.0.32.26.
Alastair Bridgewater [Thu, 12 Nov 2009 00:10:21 +0000 (00:10 +0000)]
1.0.32.26: Fix fasdumping of function arglists with instances as default values
Fasdumping of a function arglist was breaking if there was a structure
or other instance as a default value because the arglist wasn't
processed for dumping as a literal constant.
* Process the arglist for dumping as a literal constant when creating
entry-info for external entry points.
This is https://bugs.launchpad.net/sbcl/+bug/310132 .
Nikodemus Siivola [Wed, 11 Nov 2009 19:01:20 +0000 (19:01 +0000)]
1.0.32.25: save page table allocation information into core files
* Allows unboxed objects in saved cores to be written to without
write-protection & later scavenging.
Christophe Rhodes [Wed, 11 Nov 2009 18:20:34 +0000 (18:20 +0000)]
1.0.32.24: document stream external-formats and :replacement option
Christophe Rhodes [Wed, 11 Nov 2009 18:08:31 +0000 (18:08 +0000)]
1.0.32.23: use :replacement in the external format for standard IO streams
For *terminal-io*, a bidirectional stream, we have to make an arbitrary
choice on Windows, where in theory the input and output code pages can
differ. We arbitrarily choose the output format; I have no idea whether
this matters.
Christophe Rhodes [Wed, 11 Nov 2009 17:52:39 +0000 (17:52 +0000)]
1.0.32.22: implement output restarts for UCS2
Most of the work was done; we just had to actually turn the catch tag
on, and write tests that are correct rather than wrong.
Christophe Rhodes [Wed, 11 Nov 2009 17:34:09 +0000 (17:34 +0000)]
1.0.32.21: compress most unibyte-external-format definitions
All the unibyte-mapper-based external-formats had huge amounts of
cut-and-pasted code, differing only in names of functions. This is,
oddly enough, a clear case for abstracting away the repeated code into a
macro.
In the process, convert them to the multibyte apparatus, which has support
for the nice restarts, and remove the too-simple unibyte
DEFINE-EXTERNAL-FORMAT (and EXTERNAL-FORMAT-DECODING-ERROR) which are now
unused.
Include a far-from-comprehensive set of tests, which are mostly for
iso-8859-x formats
Christophe Rhodes [Wed, 11 Nov 2009 17:21:05 +0000 (17:21 +0000)]
1.0.32.20: bug fixes in unibyte external formats
* handling undefined codepoints:
There's a difference between "unassigned codepoint", represented as
(<code> nil), and "codepoint mapping to character with that codepoint",
represented by the absence of an entry in the exceptions list. Simply
testing for trueness of (cadr (assoc <code> exceptions)) isn't good
enough; test for trueness of (cdr (assoc ...)) and use the car if so.
* undefined codepoints as the first exception:
If the first exception to code<->byte is an undefined character, as for
example in iso-8859-8, the lowest-code exception was computed wrongly,
leading to incorrect encoding.
Christophe Rhodes [Wed, 11 Nov 2009 17:10:40 +0000 (17:10 +0000)]
1.0.32.19: Support for :REPLACEMENT external-format / encoding modifier
For a given keyword :foo naming an encoding, allow an external-format
(:foo :replacement <character>) such that any stream or octet coding
errors are automatically treated by using the <character> as replacement
instead.
To do this, wrap each of the functions in the external-format object
named by the keyword with a function establishing handlers for the
exceptional conditions. At the moment, the output restarts for c-string
external format conversion are not implemented (so handle specific condition
types like STREAM-FOOCODING-ERROR, not general FOOCODING-ERROR).
Nikodemus Siivola [Wed, 11 Nov 2009 15:06:29 +0000 (15:06 +0000)]
1.0.32.18: additional allocation information
* SB-INTROSPECT:ALLOCATION-INFORMATION also reports on boxedness of
the page the object resides on.
* When heap statistics are printed, always print the details of the
pseudo-static generation. Make print_generation_stats extern so it
can be called from lisp.
Christophe Rhodes [Wed, 11 Nov 2009 13:52:19 +0000 (13:52 +0000)]
1.0.32.17: make the utf-8 external format more robust
Detect all malformed sequences, including attempts to decode or encode
Unicode surrogate codepoints (disallowed by the Unicode definition of
UTF-8). Some error tests change behaviour, and some (unexported)
condition classes are not triggered under the same circumstances any
more.
Also, handle null-termination on a successful conversion of an empty range of
a nil array.
Christophe Rhodes [Wed, 11 Nov 2009 13:21:37 +0000 (13:21 +0000)]
1.0.32.16: external-format restart enhancements
* OUTPUT-REPLACEMENT restart for fd-stream external-formats, taking a
string designator argument and attempting to encode that instead of the
erroneous output;
* fixes for the FORCE-END-OF-FILE fd-stream external-format restart, using
a somewhat involved call / return protocol for communicating out-of-band
information between output routines and drivers;
* INPUT-REPLACEMENT restart for fd-stream external-formats, again with
complicated out-of-band information communication. This also interacts
with UNREAD-CHAR;
* fix the ATTEMPT-RESYNC restart (and similar) at or near the end of file,
ensuring that there is always a valid CATCH tag to be THROWN to;
* fix a double-error case in the USE-VALUE restart for unibyte octet
conversions;
* bandage fix for mb-util decoding-error USE-VALUE restart -- there's more
factoring to be done, but this fixes lp #314939
Christophe Rhodes [Wed, 11 Nov 2009 12:56:15 +0000 (12:56 +0000)]
1.0.32.15: update Unicode data files to Unicode 5.2
We do still need also to update a small bit of code, but at least the
explanatory comment now makes it obvious which bits.
Christophe Rhodes [Sat, 7 Nov 2009 10:03:22 +0000 (10:03 +0000)]
1.0.32.14: fix build under clisp
It already worked with 'clisp -on-error -abort', but since there's only one
type declaration in the whole system that clisp doesn't understand, we might
as well fix it. Report and patch by Josh Elasser, lp #396597.
trittweiler [Fri, 6 Nov 2009 22:07:57 +0000 (22:07 +0000)]
1.0.32.13: WITH-STANDARD-IO-SYNTAX must also bind *PRINT-PPRINT-DISPATCH*...
...'cuz CLHS says so. We bind it to the standard pprint dispatch
table, and guard against its modification in SET-PPRINT-DISPATCH,
mimicking the guard against modification of the standard readtable
introduced in 1.0.24.
Christophe Rhodes [Fri, 6 Nov 2009 21:12:16 +0000 (21:12 +0000)]
1.0.32.12: Fix slot-value on specialized parameters in SVUC methods
The slot accessors' functions were being computed too early, requiring a
wrapper (for the PV optimization) when none was available. The fix
delays the computation by the usual trick of installing a closure that
will perform the slot accessor computation on demand.
Include a correct MOP test for this (slightly hard because of
constraints about order of instantiation: see lp #473699, reported by
Lars Rune Nodstal).
Christophe Rhodes [Tue, 3 Nov 2009 10:42:18 +0000 (10:42 +0000)]
1.0.32.11: fix octets portion of latin-2/iso-8859-2 external-format
Typo in the external format definition. Include a test case; reported
by Attila Lendvai (in lp #471689)
Gabor Melis [Mon, 2 Nov 2009 10:56:13 +0000 (10:56 +0000)]
1.0.32.10: fix timer starvation caused by setting the system clock back
Reported by Leslie P. Polzer.
https://bugs.launchpad.net/sbcl/+bug/460283
Nathan Froyd [Fri, 30 Oct 2009 18:44:34 +0000 (18:44 +0000)]
1.0.32.9: fix thinko in DEFINE-EXTERNAL-FORMAT/VARIABLE-WIDTH
Users of this OUTPUT-BYTES/FOO function must have been perfect.
Nathan Froyd [Fri, 30 Oct 2009 18:42:02 +0000 (18:42 +0000)]
1.0.32.8: macroize x86-64 {8,16,32}-bit array accessors
One wonders if the profusion of array access VOPs are merely bloat...
Nathan Froyd [Fri, 30 Oct 2009 18:22:10 +0000 (18:22 +0000)]
1.0.32.7: fix open-coding of EQL in the cross-compiler
EQ-COMPARABLE-TYPE wasn't being defined properly, so things like
(EQL FOO :KEYWORD) were going through GENERIC-EQL.
Thanks to Christophe for pointing out the fix.
trittweiler [Fri, 30 Oct 2009 18:10:38 +0000 (18:10 +0000)]
1.0.32.6: WHO-SPECIALIZES-DIRECTLY & WHO-SPECIALIZES-GENERALLY
* Add two new Xref functions to SB-INTROSPECT:
WHO-SPECIALIZES-DIRECTLY returns method definitions which
specialize one the designated class itself.
WHO-SPECIALIZES-GENERALLY returns methods definitions which
specialize on the designated class, or subclasses of it.
Both take CLASS-EQ, and EQL specializers into account.
* Refactor xref tests, and add tests for the two new functions.
* Fix sb-introspect.asd to perform the test-op with a *D-P-D* bound
to contrib/sb-introspect/.
Christophe Rhodes [Fri, 30 Oct 2009 15:23:11 +0000 (15:23 +0000)]
1.0.32.5: defend against full MAKE-ARRAY before the type system is ready
This is basically defence against bugs of the form lp #316323, where the
wrong answer comes out of a make-array before SUBTYPEP is working
properly. We should be able to arrange the build so that we never have
to do a full MAKE-ARRAY before we're ready for it.
Slight modifications of a couple of macros: VOP-related and multibyte
encoding-related. There should be no user-visible change.
Christophe Rhodes [Thu, 29 Oct 2009 17:53:21 +0000 (17:53 +0000)]
1.0.32.4: fix INSPECT on closures
(really fix %closure-values, but the inspector was the only user of this)
Christophe Rhodes [Thu, 29 Oct 2009 16:39:01 +0000 (16:39 +0000)]
1.0.32.3: O_LARGEFILE support on x86-64/linux
Reported by Daniel Janus (and patch) as lp bug #453080.
Christophe Rhodes [Thu, 29 Oct 2009 15:36:15 +0000 (15:36 +0000)]
1.0.32.2: minor portability fix
Clozure CL issues full warnings at compile time if individual clauses in
typecases are shadowed by others; this makes writing portable
warning-free typecases for character types a little bit tricky. In the
case of the FORMAT directive finders, though, the workaround is
straightforward because the whole of the CHARACTER branch in the
typecases can be removed.
While we're at it, fix the theoretical portability bug that we were
calling the host's CHAR-CODE on BASE-CHARs to fill data structures later
to be accessed using the target's CHAR-CODE. (Theoretical because all
implementations at present agree on CHAR-CODE for BASE-CHARs).
(This commit does not suffice to allow Clozure CL to build SBCL)
CVS: ----------------------------------------------------------------------
CVS: Enter Log. Lines beginning with `CVS:' are removed automatically
CVS:
CVS: Committing in .
CVS:
CVS: Modified Files:
CVS: NEWS version.lisp-expr src/code/late-format.lisp
CVS: src/code/target-format.lisp
CVS: ----------------------------------------------------------------------
Christophe Rhodes [Mon, 26 Oct 2009 08:24:56 +0000 (08:24 +0000)]
1.0.32.1: fix MIPS build
Regression from 1.0.30.38, when %unary-truncate handling was changed.
(Actually, this might not fix the build completely; we have reports of
SIGBUS while compiling PCL, but it gets noticeably further.
Christophe Rhodes [Sun, 25 Oct 2009 15:49:10 +0000 (15:49 +0000)]
1.0.32: will be tagged as sbcl_1_0_32
Nikodemus Siivola [Thu, 8 Oct 2009 11:48:41 +0000 (11:48 +0000)]
1.0.31.32: missing --no-userinit in core.test.sh
Reported by Robert Brown.
https://bugs.launchpad.net/sbcl/+bug/440052
Nikodemus Siivola [Wed, 7 Oct 2009 18:06:35 +0000 (18:06 +0000)]
1.0.31.31: SATISFIES cannot refer to local functions
* Fix misoptimization: use SB-C::GLOBAL-FUNCTION instead of
CL:FUNCTION. (Reported by Stanislaw Halik)
* Also fix a typo in the COMPARE-AND-SWAP docstring. (Thanks to Larry
Valkama.)
Nikodemus Siivola [Wed, 7 Oct 2009 09:56:48 +0000 (09:56 +0000)]
1.0.31.30: fix IF docstring
https://bugs.launchpad.net/sbcl/+bug/439264
...good thing no-one trusts a docstring, eh. A bug report to the
effect "IF is broken, executes THEN when condition true" would have
been mortifying...
Juho Snellman [Sun, 4 Oct 2009 16:15:58 +0000 (16:15 +0000)]
1.0.31.29: Fix x86-64/FreeBSD/thread build
* Remove a forwards incompatible FreeBSD 5 kludge (patch by Hiroyuki
Komatsu)
Juho Snellman [Sat, 3 Oct 2009 23:03:29 +0000 (23:03 +0000)]
1.0.31.28: Add :EMIT-CFASL to the COMPILE-FILE defknown
Juho Snellman [Sat, 3 Oct 2009 22:36:59 +0000 (22:36 +0000)]
1.0.31.27: RUN-PROGRAM process group change
* Have RUN-PROGRAM with :INPUT T only run the subprocess in a
new process group if it doesn't need to share stdin with the
sbcl process. (patch by Leslie Polzer)
Nathan Froyd [Tue, 29 Sep 2009 13:35:37 +0000 (13:35 +0000)]
1.0.31.26: improved DATA-VECTOR-{REF,SET} on x86-64
Use the low registers directly rather than going through RAX.
Nathan Froyd [Tue, 29 Sep 2009 01:27:33 +0000 (01:27 +0000)]
1.0.31.25: fix thinko in ECHO-N-BIN
I happened to look at some STYLE-WARNINGs in octets.lisp and found this
gem. Sometimes it's nice having a chatty compiler...
Nathan Froyd [Tue, 29 Sep 2009 01:24:23 +0000 (01:24 +0000)]
1.0.31.24: add BSWAP instruction for x86
Kinda weird, lifting code for x86 from x86-64...
Nathan Froyd [Tue, 29 Sep 2009 01:02:20 +0000 (01:02 +0000)]
1.0.31.23: OAOOize external-format support
fd-streams and octets support independently kept records of
external-format->function maps, suitable for the purposes of each. This
revision stores all the relevant information for an external format in a
single place--a new EXTERNAL-FORMAT structure--and has both clients
reference things in that single place.
Doing so offers opportunities for other cleanups and speedups.
fd-streams external-format storage was an alist of lists, which was bad
for client code, since everything referred to fields with NTH or
SECOND/FOURTH/FIFTH. A proper DEFSTRUCT helps here and should be
slightly more space-efficient, as we're replacing a list with
(effectively) a vector. Also, since clients had to scan through an
alist to find an external-format, this design was hurting performance in
streams code, most notably OPEN. Replacing the alist with a hash table
(which the octets code was already using) should make things a lot
snappier.
Nathan Froyd [Sun, 27 Sep 2009 01:18:47 +0000 (01:18 +0000)]
1.0.31.22: improved SAP-{REF,SET}-* for x86-64
- Don't go through RAX in reffers; zero- or sign-extend the memory
referenced into the result register directly. Specialized array
references already do this.
- Don't go through RAX in setters; use the appropriately-sized low part
of the value register directly.
Nathan Froyd [Sat, 26 Sep 2009 23:53:26 +0000 (23:53 +0000)]
1.0.31.21: OAOOize SAP-REF-64 and friends
Move SAP-REF-64-related transforms to saptran.lisp, since there's no reason
current and future 32-bit platforms should have to define these
individually.
Nathan Froyd [Sat, 26 Sep 2009 21:34:45 +0000 (21:34 +0000)]
1.0.31.20: smaller code for failing ECASE/ETYPECASE
(ERROR 'CASE-FAILURE ...) generates a lot of code. Commonize the code in
a CASE-FAILURE function and have the macros call that function instead.
This change results in fewer entries in the constant vector and smaller
code since fewer arguments have to be loaded. This makes the error case
slightly slower, but that's not a problem.
Shrinks core size by ~160K on x86-64.
Brian Mastenbrook [Tue, 22 Sep 2009 02:22:41 +0000 (02:22 +0000)]
1.0.31.19: build x86-64 on OS X when available by default
* For consistency with other platforms, build x86-64 by default on OS
X when the OS and hardware support it.
* ... also clean up some stragglers needing .cvsignore entries
Nikodemus Siivola [Mon, 21 Sep 2009 14:37:05 +0000 (14:37 +0000)]
1.0.31.18: better SXHASH on pathnames
Patch by Walter C. Pelissero, based on CMUCL code.
Also fix stale LOGICAL-PATHNAME test, left out from 1.0.31.17.
Nikodemus Siivola [Fri, 18 Sep 2009 11:31:23 +0000 (11:31 +0000)]
1.0.31.17: LOGICAL-PATHNAME signals a TYPE-ERROR
* LOGICAL-PATHNAME is specified to signal a TYPE-ERROR if pathspec is
incorrect.
Nikodemus Siivola [Fri, 18 Sep 2009 10:40:04 +0000 (10:40 +0000)]
1.0.31.16: fix DELETE-FILE on logical pathnames
(This is really the commit message for 1.0.31.15 -- which
accidentally got the message from 1.0.31.14 again. Sorry.)
Not real content in this commit.
Nikodemus Siivola [Fri, 18 Sep 2009 10:36:57 +0000 (10:36 +0000)]
1.0.31.14: removing :DEFAULT-INITARGS via DEFCLASS
* If a DEFCLASS is executed without :DEFAULT-INITARGS,
pass :DIRECT-DEFAULT-INITARGS NIL to ENSURE-CLASS.
AMOP specifies defaulting direct default initargs for
reinitialization using existing CLASS-DIRECT-DEFAULT-INITARGS, but
there is no requirement for DEFCLASS to preserve them, and doing
that seems quite undesirable -- though I cannot find a requirement
from CLHS for removing them either.
Nikodemus Siivola [Fri, 18 Sep 2009 09:42:38 +0000 (09:42 +0000)]
1.0.31.14: removing :DEFAULT-INITARGS via DEFCLASS
* If a DEFCLASS is executed without :DEFAULT-INITARGS,
pass :DIRECT-DEFAULT-INITARGS NIL to ENSURE-CLASS.
AMOP specifies defaulting direct default initargs for
reinitialization using existing CLASS-DIRECT-DEFAULT-INITARGS, but
there is no requirement for DEFCLASS to preserve them, and doing
that seems quite undesirable -- though I cannot find a requirement
from CLHS for removing them either.
Nikodemus Siivola [Thu, 17 Sep 2009 15:35:47 +0000 (15:35 +0000)]
1.0.31.13: working XREF for inlined lambda with hairy lambda-lists
Reported by Peter Seibel.
* When a function with a hairy lambda-list is converted, the
functional we get back is an OPTIONAL-DISPATCH, which the XREF code
never sees: we need to mark the entry-points as resulting from the
inlining to have things work.
* While at it, address a FIXME by annotating the CLAMBDAs with the
original GLOBAL-VAR, so that we don't need to make guesses based on
names.
Nikodemus Siivola [Wed, 16 Sep 2009 11:46:39 +0000 (11:46 +0000)]
1.0.31.12: use global lexicals for world-lock and finalizers
*WORLD-LOCK* -> **WORLD-LOCK**.
*FINALIZER-STORE* -> **FINALIZER-STORE**.
*FINALIZER-STORE-LOCK* -> **FINALIZER-STORE-LOCK**.
...eating your own dogfood is good for you.
Nikodemus Siivola [Tue, 15 Sep 2009 21:40:05 +0000 (21:40 +0000)]
1.0.31.11: better handling of vector types in LOOP
This is really 1.0.26.12 in a fixed form.
* LOOP-DECLARE-VAR calls LOOP-TYPED-INIT iff there is no explicit
initialization form, and when it does call it, the type is
constructed from the result using TYPE-OF.
* LOOP-TYPED-INIT knows how to construct zero-length vectors for all
reasonable vector types (ones expressible with an ARRAY-TYPE.)
* LOOP-MAKE-VAR informs LOOP-DECLARE-VAR about user-provided
initialization.
Nikodemus Siivola [Tue, 15 Sep 2009 20:23:40 +0000 (20:23 +0000)]
1.0.31.10: run-sbcl.sh to support --core
Thanks to Attila Lendvai.
Nikodemus Siivola [Tue, 15 Sep 2009 11:07:38 +0000 (11:07 +0000)]
1.0.31.9: some PCL micro-optimizations
* Make *BOOT-STATE* a global variable and rename it **BOOT-STATE**.
* Make various *S?-FOO-INDEX* variables constants, and rename them
+S?-FOO-INDEX+.
* Special love for SAFE-METHOD-FOO functions: store standard method
class list in a global variable, use EQ for membership testing, and
use STD-INSTANCE-SLOTS instead of GET-SLOTS (if the method has one
of the standard classes, we know it is a standard instance.)
Low-lying Nutrient Poor Fruit 'R Us.
Nikodemus Siivola [Mon, 14 Sep 2009 09:32:52 +0000 (09:32 +0000)]
1.0.31.8: specialized out-of-line CONCATENATE for strings
https://bugs.launchpad.net/sbcl/+bug/417229
CONCATENATE 'STRING was already decent when SPEED > SPACE thanks to
open coding by the deftransform.
Deal with low-speed policies by adding %CONCATENATE-TO-STRING and
%CONCATENATE-TO-BASE-STRING and transforming to them when
appropriate.
Nikodemus Siivola [Sun, 13 Sep 2009 09:53:58 +0000 (09:53 +0000)]
1.0.31.7: transform %FIND-POSITION for strings
* Based on patch by Karol Swietlicki.
https://bugs.launchpad.net/sbcl/+bug/410122
Nikodemus Siivola [Sat, 12 Sep 2009 06:40:03 +0000 (06:40 +0000)]
1.0.31.6: better error reporting for bogus parameter specializer names
https://bugs.launchpad.net/sbcl/+bug/414788
Nikodemus Siivola [Fri, 11 Sep 2009 15:35:56 +0000 (15:35 +0000)]
1.0.31.5: WITH-SLOTS on THE forms
Thanks to David Tolpin.
Nikodemus Siivola [Fri, 11 Sep 2009 14:49:34 +0000 (14:49 +0000)]
1.0.31.4: misuse of ABORT-IR1-TRANSFORM
Should have been GIVE-UP-IR1-TRANSFORM.
Nikodemus Siivola [Fri, 11 Sep 2009 13:19:59 +0000 (13:19 +0000)]
1.0.31.3: fix saving runtime options from executable cores
Runtime options used to be clobbered from the executable core even
if there were none saved there.
Patch by Zach Beane.
Fixes https://bugs.launchpad.net/sbcl/+bug/411925
Cyrus Harmon [Mon, 31 Aug 2009 06:11:12 +0000 (06:11 +0000)]
1.0.31.2: snow leopard (macos 10.6) build fixes
* on x86/darwin use -arch i386 flags in sb-grovel C compilation
* comment out failing sb-posix tests
* add -arch i386 flags to Config.x86-darwin CFLAGS and LINKFLAGS
* remove #include <ucontext.h> instances
Christophe Rhodes [Thu, 27 Aug 2009 20:50:29 +0000 (20:50 +0000)]
1.0.31.1: document new new new SourceForge File Release procedure
Christophe Rhodes [Thu, 27 Aug 2009 20:06:02 +0000 (20:06 +0000)]
1.0.31: release, to be tagged as sbcl_1_0_31
Christophe Rhodes [Wed, 26 Aug 2009 20:57:49 +0000 (20:57 +0000)]
1.0.30.54: Fix a cross-compiler leak in specialized %unary-truncate
Don't use host values of most-fooative fixnum. (Also expand on the
comment about using an exclusive test for floats within the fixnum
range)
Christophe Rhodes [Wed, 26 Aug 2009 20:25:41 +0000 (20:25 +0000)]
1.0.30.53: fix occasional SIGBUS in single-float complex arithmetic
The alignment restrictions on movaps for moving a pair of single floats
are too strict; use movq instead. (From Paul Khuong)
Christophe Rhodes [Wed, 26 Aug 2009 17:01:49 +0000 (17:01 +0000)]
1.0.30.52: fix for multiple-value TRUNCATE
Regression from 1.0.30 (in 1.0.30.28); Noted by Lars Nostdal; fix by
Paul Khuong
Christophe Rhodes [Sun, 23 Aug 2009 21:36:13 +0000 (21:36 +0000)]
1.0.30.51: fix for COERCE compilation regression
Code of the form
(defun foo (x)
(declare (type simple-vector x))
(coerce x '(vector (unsigned-byte 8))))
should not cause a full WARNING, but with the new COERCE transforms,
expanded into one of those IFs where one branch is dead, but the
compiler couldn't prove it.
Define a whole heap of new backend type predicates for all specialized
vectors, generalizing VECTOR-T-P. (Some specialized vectors are
implemented using widetags, and so are excluded from these new
definitions).
Christophe Rhodes [Sun, 23 Aug 2009 15:06:39 +0000 (15:06 +0000)]
1.0.30.50: fix SAVE-LISP-AND-DIE docstring
From Leslie Polzer sbcl-devel 2009-08-07
Christophe Rhodes [Sun, 23 Aug 2009 09:09:52 +0000 (09:09 +0000)]
1.0.30.49: delete-file behaviour reversal
In the 1.0.29.x series, DELETE-FILE started deleting the truename of the
given pathname designator (rather than the pathname itself), meaning
that symlinks could no longer be deleted. Revert back to the historical
behaviour. (Noted on sbcl-devel by Luis Oliveira 2009-08-18.)
Although the behaviour hasn't changed recently, there's a similar issue
in RENAME-FILE, but that operator has such bizarre behaviour anyway it
might be worth just letting sleeping pathname functions lie.
Christophe Rhodes [Sat, 22 Aug 2009 15:54:46 +0000 (15:54 +0000)]
1.0.30.48: utf-8 simple-array-nil correctness
If an empty range of a simple-array-nil is requested for conversion, we
should return an empty octet sequence, not blow up.
Christophe Rhodes [Sat, 22 Aug 2009 15:43:20 +0000 (15:43 +0000)]
x86 sap fixes
The sap-ref-with-offset stuff was wrong in that the displacement was
multiplied by a notional element size, rather than being uniformly
treated as a number of bytes. Mostly this codepath wasn't exposed at
all (other than with sap-ref-8, which worked by "accident"), but
attempts to implement UTF-16, which requires (sap-ref-16 sap (+ offset
2)), showed up the problem.
Nikodemus Siivola [Wed, 12 Aug 2009 12:22:19 +0000 (12:22 +0000)]
1.0.30.46: SB-INTROSPECT test adjustment
* Fix ALLOCATION-INFORMATION test on PPC, and hopefully on SPARC as
well. Thanks to Bruce O'Neel.
Nikodemus Siivola [Wed, 12 Aug 2009 11:59:04 +0000 (11:59 +0000)]
1.0.30.45: various pretty-printing improvements
Patch by Tobias Rittweiler:
* Add a PPRINT-DECLARE which a) makes sure that (DECLARE (FUNCTION
F)) is not printed as (DECLARE #'F), and b) places each declaration
specifier on its own line. Also used for DECLAIM.
* Better pprint SETQ forms which assign to multiple variables. At the
moment it's printed like
(SETQ FOO
(FROB-FOO 0 1 2 3 4 5 6 7 8 9)
QUUX
(FROB-QUUX 9 8 7 6 5 4 3 2 1 0))
With the patch it's indented like
(SETQ FOO (FROB-FOO 0 1 2 3 4 5 6 7 8 9)
QUUX (FROB-QUUX 9 8 7 6 5 4 3 2 1 0))
It uses the former indentation style if the value (e.g. the
"(FROB-FOO ...)") does not fit on a single line.
This also affects PSETQ, SETF, PSETF.
* Add pprint entry for SB-INT:DX-FLET because there are CL macros
which expand to that.
* Fix typo in *LOOP-SEPARATING-CLAUSES*; I mistakenly put WHERE
instead of WITH in it.
* Fix PPRINT-IF to make sure that the predicate is always printed
right after the IF. The current definition may occassionally print
an IF form like
(IF
(PREDICATE)
(THEN)
(ELSE))
* Some small refactoring work:
- Use PPRINT-LINEAR, and PPRINT-FILL instead of equivalent, but
hairy FORMAT calls.
- Add PPRINT-SPREAD-FUN-CALL which is the common subtrate of
pretty-printing simple LOOP forms, and DECLARE forms.
Nikodemus Siivola [Wed, 12 Aug 2009 11:14:15 +0000 (11:14 +0000)]
1.0.30.44: DECODE-FLOAT and INTEGER-DECODE-FOAT are not flushable
...as they should signal an error for non-floats in safe code,
and hence UNSAFELY-FLUSHABLE.
Ditto for FLOAT-SIGN, FLOAT-RADIX, FLOAT-DIGITS, and FLOAT-PRECISION,
though for some reason Python seemed unwilling to delete these calls
despite them being marked as flushable as well.
See: https://bugs.launchpad.net/sbcl/+bug/412416
Nikodemus Siivola [Tue, 11 Aug 2009 11:22:17 +0000 (11:22 +0000)]
1.0.30.43: LVAR-MATCHES needs to deal with unnamed leaves
...by passing ERRORP=NIL to COMBINATION-FUN-SOURCE-NAME. Also smooth
the return value convention of C-F-S-N by adding a secondary value:
NIL is a valid name for a local function.
See: https://bugs.launchpad.net/sbcl/+bug/411563
Nikodemus Siivola [Mon, 10 Aug 2009 09:58:45 +0000 (09:58 +0000)]
1.0.30.42: missing array predicate definitions
* Not all specialized array predicates had an out-of-line predicate.
Generate them from the SAETP vector. Reported by Stelian Ionescu.
* Test case.