sbcl.git
17 years ago0.9.14.5:
Juho Snellman [Wed, 28 Jun 2006 14:42:24 +0000 (14:42 +0000)]
0.9.14.5:
Add sbcl-pwd.sh to binary-distribution.sh, to make the binaries
        actually installable.

17 years ago0.9.14.4:
Juho Snellman [Wed, 28 Jun 2006 11:35:51 +0000 (11:35 +0000)]
0.9.14.4:
FILE-POSITION sometimes returned inconsistent results for multibyte
        external-format streams. (Reported by Lutz Euler on sbcl-devel,
        patch from sbcl-devel "Patch: FILE-POSITION bug" by "vbzoli")

17 years ago0.9.14.3:
Christophe Rhodes [Wed, 28 Jun 2006 10:20:06 +0000 (10:20 +0000)]
0.9.14.3:
Merge pprint (cons symbol) change, making it sensitive to the
fboundness or otherwise of the symbol in function position.

17 years ago0.9.14.2:
Christophe Rhodes [Wed, 28 Jun 2006 08:12:29 +0000 (08:12 +0000)]
0.9.14.2:
Happy birthday to me!

Fix the internals manual

17 years ago0.9.14.1
Rudi Schlatte [Tue, 27 Jun 2006 12:52:34 +0000 (12:52 +0000)]
0.9.14.1
  Fix typo for building under cygwin

17 years ago0.9.14:
William Harold Newman [Mon, 26 Jun 2006 18:40:32 +0000 (18:40 +0000)]
0.9.14:
release version, will be tagged as sbcl_0_9_14

17 years ago0.9.13.53:
Juho Snellman [Sun, 25 Jun 2006 00:56:04 +0000 (00:56 +0000)]
0.9.13.53:
        Fix unithreaded Solaris/x86 build failure.

17 years ago0.9.13.52: Windows installer tweaks
Nikodemus Siivola [Tue, 20 Jun 2006 08:29:09 +0000 (08:29 +0000)]
0.9.13.52: Windows installer tweaks
  * Name the installer sbcl-<version>.msi, not sbcl.msi
  * De-impress Xach
     <Xach> i was impressed by ID
     <Xach> it reminded me of gentle DOOM

17 years ago0.9.13.51: Ooops -- trivial fixes against previous commit
Nikodemus Siivola [Tue, 20 Jun 2006 06:45:45 +0000 (06:45 +0000)]
0.9.13.51: Ooops -- trivial fixes against previous commit
  * s/PARSE-NATIVE-PATHNAME/PARSE-NATIVE-NAMESTRING/
  * Fix garbld docstring in user-homedir-pathname.

17 years ago0.9.13.50: Windows baby-steps
Nikodemus Siivola [Tue, 20 Jun 2006 05:38:42 +0000 (05:38 +0000)]
0.9.13.50: Windows baby-steps
  * Less sucky toplevel-init: use PARSE-NATIVE-NAMESTRING to deal with
    user-supplied init-file names, and refactor the logic between Posix
    and Windows.
  * New runtime option: --debug-environment, prints out the command line
    arguments and environment before anything fancy gets done to them.
    Good for debugging startup from Windows shortcuts, etc.
  * Less magic constants, more groveled stuff.
  * SB-WIN32::GET-FOLDER-PATH renamed to SB-WIN32::GET-FOLDER-PATHNAME,
    and it now returns pathnames instead of strings.
  * Add internal function SBCL-HOMEDIR-PATHNAME, and centralize the
    SBCL_HOME stuff there.
  * Still LESS_SHOUTING.
  * Move stuff-groveled-from-headers.lisp earlier in the build-order,
    so toplevel.lisp can use it.
  * No localtime_r and gmtime_r on Windows.

17 years ago0.9.13.49:
William Harold Newman [Mon, 19 Jun 2006 13:25:00 +0000 (13:25 +0000)]
0.9.13.49:
logged bug 404

17 years ago0.9.13.48: ASDF update from upstream
Nikodemus Siivola [Mon, 19 Jun 2006 00:18:37 +0000 (00:18 +0000)]
0.9.13.48: ASDF update from upstream
  * This version at least supports (oos 'load-op :foo :force t),
    no thanks to yours truely.

17 years ago0.9.13.47: Thread safety miscellania
Nikodemus Siivola [Sun, 18 Jun 2006 23:47:57 +0000 (23:47 +0000)]
0.9.13.47: Thread safety miscellania
  * Use localtime_r and gmtime_r instead of thread-unsafe variants
    in time.c.
  * Note gethostby* reentrancy issue in sb-bsd-sockets in BUGS.
  * Trivial PCL cleanups:
    ** *ALL-PV-TABLE-LIST* was unused -- things only got pushed on it,
       but never read. Commented out with a comment, removed code that
       did the pushing.
    ** *INITIAL-PV-TABLE* was unused, deleted.
    ** *EMPTY-PV* was unused, deleted.
    ** *EMPTY-CACHE* was unused, deleted.
    ** *MFP1* & friends implemented a two-element cache outside
       a hash-table. Deleted and replaced with simple hash-table
       access. Boinkmarks will show the truth.
    ** Moved redundant toplevel SETQs to the original DEFPARAMTER
       of various *FOO-COST*s.
    How is this thread-safety related? It happened while flagging
    some PCL globals as safe and others as worthy of attention
    in doc/internals-notes/threading-specials. Also sorted the
    names, so *THE-FOO*s stand out in one place -- seeming to
    me as least-likely causes of trouble.
  * Whitespace.

17 years ago0.9.13.46: GET-INTERNAL-RUN-TIME on Windows, + Windows cleanups
Nikodemus Siivola [Sun, 18 Jun 2006 19:11:31 +0000 (19:11 +0000)]
0.9.13.46: GET-INTERNAL-RUN-TIME on Windows, + Windows cleanups
  * Merge the patch by Frank Buss, that implements
    get-internal-run-time for Windows, modulo stylistic changes.
    (Restores buildability on Windows.)
  * LESS UPCASE.
  * Also slightly less #!+/-win32 conditionalization.
  * Grovel more stuff on Windows.

17 years ago0.9.13.45:
William Harold Newman [Wed, 14 Jun 2006 19:18:45 +0000 (19:18 +0000)]
0.9.13.45:
comment tweak inspired by Gisle.Salensminde message to sbcl-help

17 years ago0.9.13.44:
William Harold Newman [Wed, 14 Jun 2006 13:52:05 +0000 (13:52 +0000)]
0.9.13.44:
tweaked text around READ-ERROR for undefined #n# label

17 years ago0.9.13.43: install sbcl.info-1 (and any other *.info-* files)
Nikodemus Siivola [Mon, 12 Jun 2006 16:32:46 +0000 (16:32 +0000)]
0.9.13.43: install sbcl.info-1 (and any other *.info-* files)
   * makeinfo splits large info-files into separate parts, so
     we'd better deal with them.

17 years ago0.9.13.42:
Juho Snellman [Sat, 10 Jun 2006 05:07:24 +0000 (05:07 +0000)]
0.9.13.42:
Don't generate (or include) genesis/lutex.h when building without
        SB-THREAD. Should fix the FreeBSD 4 build problems reported
        on sbcl-devel.

        Whitespace fixes.

17 years ago0.9.13.41:
Juho Snellman [Sat, 10 Jun 2006 01:17:35 +0000 (01:17 +0000)]
0.9.13.41:
Improvements to the disassembly of INC and DEC on x86-64 (patch
        by Lutz Euler on sbcl-devel)

17 years ago0.9.13.40: RUN-PROGRAM tweak
Nikodemus Siivola [Sat, 10 Jun 2006 01:00:58 +0000 (01:00 +0000)]
0.9.13.40: RUN-PROGRAM tweak
  * If the streams passed to RUN-PRGRAM are composite streams,
    try to extract the underlying FD-STREAM before falling
    back to the tempfile/pipe strategy.

17 years ago0.9.13.39:
Juho Snellman [Sat, 10 Jun 2006 00:26:14 +0000 (00:26 +0000)]
0.9.13.39:
Fix FP exception handling on FreeBSD (patch by NIIMI Satoshi
on sbcl-devel)

17 years ago0.9.13.38:
Juho Snellman [Sat, 10 Jun 2006 00:04:05 +0000 (00:04 +0000)]
0.9.13.38:
Reintroduce the unused "linux_no_threads_p" variable (deleted
        with some lutex-related cleanups), since the latest released
        version of Slime wants to access it.

17 years ago0.9.13.37:
Juho Snellman [Fri, 9 Jun 2006 23:58:27 +0000 (23:58 +0000)]
0.9.13.37:
Add an explicit -fPIC to the gcc options for building a shared
        library for x86-64 in threads.impure.lisp, as apparently required
        for some versions of gcc. (thanks to Marco Monteiro)

17 years ago0.9.13.36: global policy / null-lexenv confusion fix
Nikodemus Siivola [Fri, 9 Jun 2006 20:59:44 +0000 (20:59 +0000)]
0.9.13.36: global policy / null-lexenv confusion fix
  * Do not store the global policy in null-lexenv, but include
    it in subsequent lexenvs. Fixes visibility of global policy
    in LOCALLY and MACROLET.
  * Also actually enable the SB-LDB in default build instead of
    just providing the framework to make this a good idea.
    (Accidentally left out from from 0.9.13.33)

17 years ago0.9.13.35:
William Harold Newman [Fri, 9 Jun 2006 17:07:37 +0000 (17:07 +0000)]
0.9.13.35:
logged bug 403
added workaround for bug 403 when printing *DEBUG-CONDITION*

17 years ago0.9.13.34: Class objects as specializers
Nikodemus Siivola [Wed, 7 Jun 2006 19:08:30 +0000 (19:08 +0000)]
0.9.13.34: Class objects as specializers
 * As reported by Pascal Costanze on sbcl-devel.
 * Also record PCL code walker bug wrt. user-defined declarations.

17 years ago0.9.13.33: :SB-LDB in default build
Nikodemus Siivola [Wed, 7 Jun 2006 16:25:09 +0000 (16:25 +0000)]
0.9.13.33: :SB-LDB in default build
 * DISABLE-DEBUGGER now also turns LDB off, so scriptability remains the same.
 * Use lose() instead of monitor_or_something() in the runtime, and replace
   set_lossage_handler() with enable/disable_lossage_handler().

17 years ago0.9.13.32: restarts for slot-unbound
Nikodemus Siivola [Wed, 7 Jun 2006 09:58:37 +0000 (09:58 +0000)]
0.9.13.32: restarts for slot-unbound
 * USE-VALUE and STORE-VALUE have obvious meanings here, so use them.
 * Missing NEWS for .31.

17 years ago0.9.13.31: native-namestring bugfix
Nikodemus Siivola [Wed, 7 Jun 2006 09:49:54 +0000 (09:49 +0000)]
0.9.13.31: native-namestring bugfix
 * The native-namestring should be same for :DIRECTORY (:RELATIVE) and
   NIL.

18 years ago0.9.13.30
Cyrus Harmon [Tue, 6 Jun 2006 21:27:46 +0000 (21:27 +0000)]
0.9.13.30
   * restore macos/x86 buildability without sb-thread by making the
     lutex primitive object defintion and %make-lutex and lutexp
     definitions be #!+sb-lutex not #!+(and sb-lutex sb-thread).

18 years ago0.9.13.29:
Juho Snellman [Tue, 6 Jun 2006 17:11:16 +0000 (17:11 +0000)]
0.9.13.29:
Fix some regressions:

        * x86 build was broken by the ALWAYS-TRANSLATABLE defknown changes
        * Some symbol package issues in the test framework caused
          --break-on-failure to stop working (patch by Yaroslav Kavenchuk)

18 years ago0.9.13.28:
Nathan Froyd [Tue, 6 Jun 2006 02:42:51 +0000 (02:42 +0000)]
0.9.13.28:
Move full-call checking responsibilities to DEFKNOWN.
... new ALWAYS-TRANSLATABLE flag in DEFKNOWN;
... change PONDER-FULL-CALL to check said flag;
... declare flag in appropriate cases;
... by analogy with %INSTANCE-{REF,SET}, flag their %RAW-*
  counterparts as well (they were not flagged previously)..

There are probably other cases where it's applicable, too.

18 years ago0.9.13.27:
Nathan Froyd [Tue, 6 Jun 2006 01:54:42 +0000 (01:54 +0000)]
0.9.13.27:
Restore buildability on OS X/PPC.

18 years ago0.9.13.26: Life after heap exhaustion on GENCGC
Nikodemus Siivola [Mon, 5 Jun 2006 12:25:54 +0000 (12:25 +0000)]
0.9.13.26: Life after heap exhaustion on GENCGC
 * If gc_find_freeish_pages fails, but managed to find _some_
   memory, and we are not in GC, signal a HEAP-EXHAUSTED-ERROR.
 * A few stray s/0/FREE_PAGE_FLAG/'s.
 * Fix fprintf control string in print_generation_stats.
 * *CONTROL-STACK-EXHAUSTION-SAP* is unused, get rid of it.

18 years ago0.9.13.25:
Nathan Froyd [Sun, 4 Jun 2006 14:25:06 +0000 (14:25 +0000)]
0.9.13.25:
Remove vestiges of raw-slot indirection in structures for non-HPPA
  ports.
... #!{-,+}hppa where appropriate;
... purge %RAW-REF-FOO and associated SET VOPs on #-HPPA ports;
... tested on PPC.

18 years ago0.9.13.24:
Juho Snellman [Sat, 3 Jun 2006 21:31:33 +0000 (21:31 +0000)]
0.9.13.24:
Fix a bug with the EAGAIN/EINTR handling of SB-BSD-SOCKETS:SOCKET-SEND.
        Add EINTR handling to SOCKET-ACCEPT and SOCKET-RECEIVE.

18 years ago0.9.13.23:
Thiemo Seufer [Sat, 3 Jun 2006 20:59:00 +0000 (20:59 +0000)]
0.9.13.23:
Unwhitespaceify.

18 years ago0.9.13.22:
Juho Snellman [Sat, 3 Jun 2006 20:26:52 +0000 (20:26 +0000)]
0.9.13.22:
Implement SB-THREAD mutexes and waitqueues using only pthread
        functionality on platforms that don't support Linux futexes. New
        platforms that can be compiled with SB-THREAD:

          * Solaris/x86 (seems to be as stable as SBCL threads on Linux)
          * OS X/x86 (some known stability problems, but doesn't fail on the
            thread regression tests every time)
          * FreeBSD/x86 (reportedly "flat-out broken", tends to cause
            kernel panics)

        While I (Juho) am doing the final merge from lutex-branch to
        HEAD, much of the work was done by Cyrus Harmon, based on an
        initial implementation by Nathan Froyd. The Solaris work was
        funded by Tellme Networks, Inc.

18 years ago0.9.13.21:
Juho Snellman [Thu, 1 Jun 2006 12:49:00 +0000 (12:49 +0000)]
0.9.13.21:
Make TIME show the amount of run-time spent in GC. SBCLs of old used
        to have this feature, but (as far as I can tell) it was unintentionally
        disabled in 0.pre8.119.thread-gc.1.

18 years ago0.9.13.20:
Juho Snellman [Thu, 1 Jun 2006 12:01:18 +0000 (12:01 +0000)]
0.9.13.20:
Merge sbcl-devel patch "NAME-CHAR and unrecognised symbols" by
        Robert J. Macomber, fixing an issue with NAME-CHAR signaling an error
        when given an invalid symbol.

18 years ago0.9.13.19:
Juho Snellman [Thu, 1 Jun 2006 11:41:33 +0000 (11:41 +0000)]
0.9.13.19:
Fix "[Sbcl-devel] Bug: :element-type '(signed-byte 32) on an AMD64"
        reported by Marco Monteiro.

18 years ago0.9.13.18:
Juho Snellman [Thu, 1 Jun 2006 11:12:38 +0000 (11:12 +0000)]
0.9.13.18:
Give a better diagnostic for the commo x86-64 "mprotect: Unable
        to allocate memory" error, as suggested by Max Mikhanosha on sbcl-devel.

18 years ago0.9.13.17:
Juho Snellman [Thu, 1 Jun 2006 10:38:57 +0000 (10:38 +0000)]
0.9.13.17:
Support running the impure tests on win32, by replacing usages of
        fork in the test framework with run-program. (patch by Yaroslav
        Kavenchuk on sbcl-devel)

18 years ago0.9.13.16: preliminary Windows installer builder
Nikodemus Siivola [Thu, 1 Jun 2006 09:14:25 +0000 (09:14 +0000)]
0.9.13.16: preliminary Windows installer builder
 * Added "free software & no warranty" summary to top of COPYING.

 * tools-for-build/rtf.lisp turns COPYING into License.rtf needed
   for the installer.

 * tools-for-build/wxs.lisp generates the XML from which
   the installer is built.

 * refactor good-for-lisp pathname logic from make-target-contrib.sh
   to sbcl-pwd.sh.

 * make-windows-installer.sh builds sbcl.msi into output/,
   assuming WiX (2.0) is installed in $PROGRAMFILES/WiX or
   $WIX_PATH.

   The installer installs sbcl.exe, sbcl.core, and contribs
   into $PROGRAMFILES/Steel Bank Common Lisp/<sbcl-version>/
   by default, though the location is configurable.

   .lisp and .fasl files are associated with the installed
   SBCL, action being to start SBCL and load the file.

   The installation directory is added to PATH.

   SBCL_HOME is set to the installation directory.

   A shortcut is added to the start-menu. The shortcut refers
   to the core explicitly, and will continue to work even if
   a newer SBCL is installed.

18 years ago0.9.13.15:
Juho Snellman [Thu, 1 Jun 2006 09:08:01 +0000 (09:08 +0000)]
0.9.13.15:
Add x86-64 disassembler support for the #x66 prefix (patch by
        Lutz Euler on sbcl-devel)

18 years ago0.9.13.14:
Juho Snellman [Thu, 1 Jun 2006 08:52:44 +0000 (08:52 +0000)]
0.9.13.14:
Fix saving >2GB cores on x86-64 (reported by Benjamin Lambert
on sbcl-help)

18 years ago0.9.13.13: docstrings.lisp robustification
Nikodemus Siivola [Wed, 31 May 2006 10:51:51 +0000 (10:51 +0000)]
0.9.13.13: docstrings.lisp robustification
 * Add replacement sequences for > and < in pathnames, and
   deal with orphaned docstrings. (Needed for Windows.)

18 years ago0.9.13.12: GNU make finding tweak
Nikodemus Siivola [Wed, 31 May 2006 10:46:46 +0000 (10:46 +0000)]
0.9.13.12: GNU make finding tweak
 * Try "make" before "gmake": should not (hopefully!) break
   anything, and makes build work out of the box on MSYS when
   Cygwin is installed and in PATH.

18 years ago0.9.13.11:
Nathan Froyd [Wed, 31 May 2006 02:26:34 +0000 (02:26 +0000)]
0.9.13.11:
Merge "Patch: A code size microoptimization for x86-64 and x86
  (EMIT-EA)" from Lutz Euler, sbcl-devel, 2006-28-05.

18 years ago0.9.13.10: better SB-BSD-SOCKETS on Windows
Nikodemus Siivola [Tue, 30 May 2006 12:03:00 +0000 (12:03 +0000)]
0.9.13.10: better SB-BSD-SOCKETS on Windows
 * ...now using SB-GROVEL, thanks to Timothy Ritchey.

18 years ago0.9.13.9: expand portability fixes
Nikodemus Siivola [Tue, 30 May 2006 11:58:43 +0000 (11:58 +0000)]
0.9.13.9: expand portability fixes
 * $DEVNULL, as BSD expand hangs without < /dev/null, and Windows needs
   NUL.

18 years ago0.9.13.8:
Christophe Rhodes [Sun, 28 May 2006 14:35:43 +0000 (14:35 +0000)]
0.9.13.8:
Fix the compiler transform for %set-sap-ref-foo and sap+

18 years ago0.9.13.7:
Christophe Rhodes [Sun, 28 May 2006 14:27:23 +0000 (14:27 +0000)]
0.9.13.7:
Give TRACE a chance (... to print return values nicely indented)
... by binding *print-pretty* to T around the bits where we
use the pretty printer.

18 years ago0.9.13.6:
Christophe Rhodes [Sun, 28 May 2006 14:16:59 +0000 (14:16 +0000)]
0.9.13.6:
Adjust asdf "make up" target for new Sourceforge layout.
... and update asdf, while we're at it.

18 years ago0.9.13.5:
Christophe Rhodes [Sun, 28 May 2006 12:24:04 +0000 (12:24 +0000)]
0.9.13.5:
Fix for bug reported by James Y Knight sbcl-devel 2006-05-17
"merge-pathnames bug"
... don't let :back scribble over :relative

18 years ago0.9.13.4:
Christophe Rhodes [Sun, 28 May 2006 11:02:02 +0000 (11:02 +0000)]
0.9.13.4:
Explicitly disallow specializations on new-value for (setf
slot-value-using-class), as suggested by Levente Mezaros
sbcl-devel 2006-05-09.
... documentation
... internals documentation
... also one or two other small cleanups: don't keep dfun
information in special-case generic functions; make
metaobject-protocol-violation an error, not just a
condition.

18 years ago0.9.13.3:
Christophe Rhodes [Sun, 28 May 2006 10:22:22 +0000 (10:22 +0000)]
0.9.13.3:
Fix for bug reported by Pascal Bourguignon sbcl-help 2006-05-22
"cannot take the namestring of a logical pathname"
... be more careful before applying sequence functions to
pathname components.
... test case.

18 years ago0.9.13.2:
Alexey Dejneka [Sun, 28 May 2006 04:16:09 +0000 (04:16 +0000)]
0.9.13.2:
        * Fix bug MISC.641: unused optional entry deletion after
          physical environment analysis can cause LET-conversion, so
          the latter must deal with NLX-INFOs.

18 years ago0.9.13.1:
Nathan Froyd [Sun, 28 May 2006 03:25:51 +0000 (03:25 +0000)]
0.9.13.1:
Improve SAP ref/set VOPs, inspired by Alistair Bridgewater's
  comments in #lisp:
... MOV{S,Z}X can take EAs; don't bother with a temporary
    (the equivalent VOPs for DATA-VECTOR-REF already did this;
     so much for OAOO...);
... take this opportunity to consolidate VOPs and shrink the core.

18 years ago0.9.13:
William Harold Newman [Sat, 27 May 2006 20:46:40 +0000 (20:46 +0000)]
0.9.13:
release, tagged as sbcl_0_9_13

18 years ago0.9.12.31:
Thiemo Seufer [Fri, 26 May 2006 17:54:10 +0000 (17:54 +0000)]
0.9.12.31:
- Clean up endianness dependency.
- Use an explicit alien type for context register.

18 years ago0.9.12.30:
Thiemo Seufer [Fri, 26 May 2006 17:50:21 +0000 (17:50 +0000)]
0.9.12.30:
Add for mips/linux 64k guard gaps between the memory spaces.

18 years ago0.9.12.29:
Nathan Froyd [Tue, 23 May 2006 21:33:39 +0000 (21:33 +0000)]
0.9.12.29:
Brain-dump some OPTIMIZATIONS.

18 years ago0.9.12.28:
Nathan Froyd [Sun, 21 May 2006 02:50:48 +0000 (02:50 +0000)]
0.9.12.28:
Micro-optimize type testing on x86 by doing a direct comparison
  with memory in the common case, which is slightly smaller.

18 years ago0.9.12.27
Rudi Schlatte [Wed, 17 May 2006 12:06:06 +0000 (12:06 +0000)]
0.9.12.27
  Really make install.sh work on win32.  "oops"

18 years ago0.9.12.26:
Brian Mastenbrook [Mon, 15 May 2006 16:11:38 +0000 (16:11 +0000)]
0.9.12.26:
The argument to IN-PACKAGE is a string designator; stringify
it early in case it's a symbol interned in a package which no
longer exists at FASL load time. (Thanks to Zach Beane)

18 years ago0.9.12.25:
Nathan Froyd [Mon, 15 May 2006 11:56:21 +0000 (11:56 +0000)]
0.9.12.25:
Micro-optimization of BOUNDP for x86 and x86-64 (#-sb-thread only).

18 years ago0.9.12.24:
Nathan Froyd [Mon, 15 May 2006 11:47:44 +0000 (11:47 +0000)]
0.9.12.24:
Be sure to place a constant arg second when dealing with EQL.

18 years ago0.9.12.23
Rudi Schlatte [Sun, 14 May 2006 18:44:01 +0000 (18:44 +0000)]
0.9.12.23
  install.sh on win32
  - installs under c:/Program Files/sbcl, or wherever your Windows
    installation believes programs should go
  - need to set SBCL_HOME accordingly (e.g. to "c:/Program
    Files/sbcl/lib/sbcl") to run it

18 years ago0.9.12.22:
Thiemo Seufer [Sat, 13 May 2006 20:02:55 +0000 (20:02 +0000)]
0.9.12.22:
MIPS alien callbacks.

18 years ago0.9.12.21:
Thiemo Seufer [Sat, 13 May 2006 19:50:32 +0000 (19:50 +0000)]
0.9.12.21:
Fix syntax for fails-on tag, add :mips, which fails as well.

18 years ago0.9.12.20:
Juho Snellman [Sat, 13 May 2006 19:48:17 +0000 (19:48 +0000)]
0.9.12.20:
Get rid of the PCL method-lambda macrolets completely (see also
        0.9.12.6).

        * Replace PV-ENV with a global macro that switches its behaviour
          based on the environment where it's expanded.
        * The macrolets created by BIND-*-LEXICAL-METHOD-MACROS
          were only used by BIND-LEXICAL-METHOD-FUNCTIONS. Merge them into a
          BIND-*-LEXICAL-METHOD-FUNCTIONS without the macrolets. Also remove
          some dead code in the macros (I'm sure it made sense 15 years ago).

18 years ago0.9.12.19:
Thiemo Seufer [Sat, 13 May 2006 19:46:45 +0000 (19:46 +0000)]
0.9.12.19:
Add a newline, just because.

18 years ago0.9.12.18:
Thiemo Seufer [Sat, 13 May 2006 19:44:04 +0000 (19:44 +0000)]
0.9.12.18:
Fix typos in comment.

18 years ago0.9.12.17:
Thiemo Seufer [Sat, 13 May 2006 19:34:27 +0000 (19:34 +0000)]
0.9.12.17:
Add copyright information header.

18 years ago0.9.12.16:
Thiemo Seufer [Sat, 13 May 2006 19:29:59 +0000 (19:29 +0000)]
0.9.12.16:
Minor code improvement.

18 years ago0.9.12.15:
Juho Snellman [Sat, 13 May 2006 18:43:42 +0000 (18:43 +0000)]
0.9.12.15:
Use component address in addition to the debug-name to distinguish
        functions in SB-SPROF. Call-graphs tended to be pretty useless due
        to all (LAMBDA ())'s getting clustered into the same profiler node.

18 years ago0.9.12.14:
Juho Snellman [Sat, 13 May 2006 18:33:40 +0000 (18:33 +0000)]
0.9.12.14:
        Get rid of some low-level hotspots in the compiler:

        * Use simple-arrays for assembler segment buffers.
        * Move the checking of policy variable name validity from run-time
          to compile-time where possible.
        * In POLICY, don't fetch the values of optimization qualities that
          aren't used in the body.
        * When creating debug-names, don't call FORMAT when encountering
          values of unhandled types. Instead mark them with a suitable
          s-exp.

18 years ago0.9.12.13:
Juho Snellman [Sat, 13 May 2006 17:20:03 +0000 (17:20 +0000)]
0.9.12.13:
        Add a simple compiler from some common toplevel forms directly
to fasl bytecode operations, instead of going through the real
compiler. Shrinks fasls and speeds up COMPILE-FILE and fasl
        loading.

18 years ago0.9.12.12:
Cyrus Harmon [Sat, 13 May 2006 16:27:50 +0000 (16:27 +0000)]
0.9.12.12:
        * Fix bug 401: IR1-transform for TYPEP aborts transformation
          if invalid type is encountered.
        NOTE: reapplying changes lost in the sf.net CVS outage

18 years ago0.9.12.11:
Cyrus Harmon [Sat, 13 May 2006 16:25:45 +0000 (16:25 +0000)]
0.9.12.11:
Simplify lui calculation.
        NOTE: reapplying changes lost in the sf.net CVS outage

18 years ago0.9.12.10:
Cyrus Harmon [Sat, 13 May 2006 16:23:53 +0000 (16:23 +0000)]
0.9.12.10:
Add package specifiers.
        NOTE: reapplying changes lost in the sf.net CVS outage

18 years ago0.9.12.9:
Cyrus Harmon [Sat, 13 May 2006 16:21:50 +0000 (16:21 +0000)]
0.9.12.9:
Improve code legibility.
        NOTE: reapplying changes lost in the sf.net CVS outage

18 years ago0.9.12.8:
Cyrus Harmon [Sat, 13 May 2006 16:18:40 +0000 (16:18 +0000)]
0.9.12.8:
Improve code legibility. Minor instruction shceduling improvements.
        NOTE: reapplying changes lost in the sf.net CVS outage

18 years ago0.9.12.7:
Cyrus Harmon [Sat, 13 May 2006 16:16:49 +0000 (16:16 +0000)]
0.9.12.7:
Mips assembly micro optimisations.
        NOTE: reapplying changes lost in the sf.net CVS outage

18 years ago0.9.12.6:
Juho Snellman [Sat, 6 May 2006 22:40:31 +0000 (22:40 +0000)]
0.9.12.6:
Speed up CLOS compilation. (For example, 45% speedup for compiling
        asdf.lisp, 30% for compiling McCLIM on x86-64).

        The code generated by PCL in MAKE-METHOD-LAMBDA has lots of
        macrolets, which for the most part are never expanded. Modify
        it to only create the macrolets that are really used in the
        body, so that the useless local macro-functions don't need to
        be compiled.

        You might wonder why this is done in PCL, rather than as a
        general purpose compiler change by lazily compiling the
        definitions when they're first expanded. I tried that first,
        and while it worked, the end result was rather messy. Since
        users can access the macro-functions through the environment,
        we need to minimally compile them to be ansixly correct, and
        we don't really have much useful minimal compilation
        infrastructure for at the moment. Ensuring that the source of
        the macro-functions is stored properly, e.g. for (MACROLET
        ((FOO ...)) (DECLAIM (INLINE BAR)) (DEFUN BAR () (FOO))) is
        also somewhat tricky.

18 years ago0.9.12.5:
Thiemo Seufer [Sun, 30 Apr 2006 19:21:24 +0000 (19:21 +0000)]
0.9.12.5:
More .cvsignore-age.

18 years ago0.9.12.4:
William Harold Newman [Fri, 28 Apr 2006 13:08:50 +0000 (13:08 +0000)]
0.9.12.4:
tweaked whitespace in compiler while hunting unsuccessfully
for a quick fix for bug 401
logged bug 401

18 years ago0.9.12.3:
Juho Snellman [Thu, 27 Apr 2006 15:56:50 +0000 (15:56 +0000)]
0.9.12.3:
Make all internals calls to RANDOM use their own random states instead
        of *RANDOM-STATE*.

18 years ago0.9.12.2:
Juho Snellman [Thu, 27 Apr 2006 13:15:21 +0000 (13:15 +0000)]
0.9.12.2:
Generate source path information for macro-expansion errors (patch
        by Helmut Eller on slime-devel)

18 years ago0.9.12.1:
Juho Snellman [Thu, 27 Apr 2006 11:00:17 +0000 (11:00 +0000)]
0.9.12.1:
sendto()-support broke SB-BSD-SOCKETS build on Solaris. Add
        some compiler flags to get the right header versions pulled in.

18 years ago0.9.12:
William Harold Newman [Wed, 26 Apr 2006 16:29:23 +0000 (16:29 +0000)]
0.9.12:
release, will be tagged as sbcl_0_9_12

18 years ago0.9.11.47:
Christophe Rhodes [Mon, 24 Apr 2006 13:51:16 +0000 (13:51 +0000)]
0.9.11.47:
Bandage over the signal trampoline for sparc/linux.  (Builds and
builds itself once more -- first time since 0.9.5.50 or so)

18 years ago0.9.11.46
Rudi Schlatte [Sun, 23 Apr 2006 12:21:12 +0000 (12:21 +0000)]
0.9.11.46
  Disable (non-working) stack guard page protection on win32
  ... No more VirtualProtect error 0x1e7
  ... thanks to Alastair Bridgewater and Yaroslav Kavenchuk

18 years ago0.9.11.45:
Christophe Rhodes [Thu, 20 Apr 2006 17:29:13 +0000 (17:29 +0000)]
0.9.11.45:
More intuitive FINALIZE-INHERITANCE behaviour
... the change for 0.9.11 fixing forward references and TYPEP
made finalize-inheritance call itself recursively on
subclasses.  This broke clg, which is now fixed, but
still...
... might as well fix the odd behaviour, and add a test for it.

18 years ago0.9.11.44:
Christophe Rhodes [Thu, 20 Apr 2006 14:51:01 +0000 (14:51 +0000)]
0.9.11.44:
More upstream asdf fixes
... win32 support;
... more paranoid file-write-date usage

18 years ago0.9.11.43:
Christophe Rhodes [Thu, 20 Apr 2006 14:25:44 +0000 (14:25 +0000)]
0.9.11.43:
Fix bug reported by Levente Meszaros sbcl-devel 2004-04-19:
cache filling with negative offset.
... when precomputing cache contents, don't include classes with
invalid wrappers.
... irony of ironies: I'm pretty sure that there's a bug lurking
somewhere else in any case, because the problem was
being exhibited in the computation of a cache for
SLOT-BOUNDP-USING-CLASS, which doesn't (or shouldn't)
use a cacheing dfun, but instead has its own specialized
dfun which basically calls the boundp function from the
slot definition.  Hmm...
... comments and minor tidying in cache.lisp

18 years ago0.9.11.42:
Christophe Rhodes [Thu, 20 Apr 2006 08:48:24 +0000 (08:48 +0000)]
0.9.11.42:
Merge fix from upstream ASDF, keeping (for now) the local win32
modification which should have been sent upstream

18 years ago0.9.11.41:
Alexey Dejneka [Thu, 20 Apr 2006 07:53:22 +0000 (07:53 +0000)]
0.9.11.41:
        * Repair RUN-PROGRAM on Linux.

18 years ago0.9.11.40: constant folding
Nikodemus Siivola [Wed, 19 Apr 2006 14:16:05 +0000 (14:16 +0000)]
0.9.11.40: constant folding
 * Out-of-line versions for %SIN, %COS, and %TAN.