sbcl Log


Commit Date  
[478afd] (sbcl_0_6_13) by William Harold Newman William Harold Newman

0.6.13:
"We release no software before its time.":-| (Hopefully the
interval between 0.6.13 and 0.7.0 will be shorter
than the interval between 0.6.12 and 0.6.13..)

2001-08-06 18:02:54 Tree
[5b0638] by William Harold Newman William Harold Newman

0.6.12.65:
removed redundant-in-SBCL FBOUNDP 'COMPILE test in
pprint.lisp's COMPUTE-TEST-FN
merged MNA INSPECT patches (sbcl-devel 2001-08-02)
renamed INSPECT-related service functions from DESCRIBE-FOO
to INSPECTED-FOO
renamed SHOW-HELP, INPUT-LOOP, and stuff to be more
INSPECT-related
rewrote the CONS method so that (INSPECT '(1 . 2)) works
renamed NUMBERED-PARTS to NAMED-P
redid the list return value from INSPECTED-PARTS as multiple
VALUES
made INSPECT work recursively instead of maintaining an
explicit *INSPECT-OBJECT-STACK*
defined SB-EXT:*INSPECTED*
removed 'redisplay'/'recompute' distinction in INSPECT,
since I can't see what it's good for (but I'd
cheerfully restore it if someone can explain)

2001-08-05 16:27:39 Tree
[525126] by William Harold Newman William Harold Newman

0.6.12.64:
removed duplicate DEFKNOWNs for other compiler/x86/system.lisp
stuff following the same approach as in 0.6.12.63
removed other duplicate DEFKNOWNs by simple deletion
tweaked DEF!MACRO DEFKNOWN so that it now handles duplicate
DEFKNOWNs with CERROR instead of WARN

2001-08-04 23:52:15 Tree
[0c54ea] by William Harold Newman William Harold Newman

0.6.12.63:
added code to report what's being overwritten by a duplicate
DEFKNOWN
removed duplicate DEFKNOWN for
(SETF %FUNCALLABLE-INSTANCE-FUNCTION), at the cost of
much entrail-reading and a lengthy KLUDGE note

2001-08-02 18:29:46 Tree
[b80843] by William Harold Newman William Harold Newman

0.6.12.62:
made DEFMACRO DEFKNOWN check for duplicate DEFKNOWNs
(I want to remove the duplicate DEFKNOWNs, but currently doing
that makes make.sh die in src/pcl/walk.lisp. So I've
put the duplicate DEFKNOWNs back, and things work
again. Since by now I have a strong conditioned reflex
"working version"=>"CVS checkin", I'll check this in
before going on to try to figure out which duplicate
DEFKNOWNs are important and why.)

2001-08-02 14:18:09 Tree
[b1a508] by William Harold Newman William Harold Newman

0.6.12.61:
copied transforms for TRUNCATE, FLOOR, and CEILING from
CMU CL 18c. (Like various other efficiency fixes,
they're in contrib/compiler-extras.lisp instead
of the main system. My plan is to merge all the
contrib/*-extra.lisp stuff into the main system
in version 0.7.x.)

2001-08-01 18:02:29 Tree
[853b48] by William Harold Newman William Harold Newman

0.6.12.60:
merged MNA clean-up-PCL-code-walker patch

2001-08-01 17:39:39 Tree
[bde2d0] by William Harold Newman William Harold Newman

0.6.12.59:
simplified *N-BYTES-FREED-OR-PURIFIED-PCOUNTER* to
*N-BYTES-FREED-OR-PURIFIED*, an ordinary UNSIGNED-BYTE
cleared *N-BYTES-FREED-OR-PURIFIED* on program startup

2001-08-01 02:06:18 Tree
[3bd7a9] by William Harold Newman William Harold Newman

0.6.12.49:
made verbose GC output report GET-INTERNAL-RUN-TIME when each
GC happens, so that when you reading a log containing
many verbose GC messages, you have a better chance of
grokking what happened
fixed POSIX-GETENV doc string after Alexey Dejneka pointed out
that it was wrong

0.6.12.50:
removed some #+OpenBSD stubifications, since FILE-LENGTH is no
longer completely broken on OpenBSD now
fixed the skip-trailing-whitespace logic in READ so it no
longer requires a second Ctrl-D char to return EOF

0.6.12.51:
merged MNA HANDLER-CASE patch: Since the compiler seems to be
smart enough to handle it now, use lexical scoping
again instead of the previous dynamic scoping
workaround; and also do #+X86 (FLOAT-WAIT).

0.6.12.52:
quasimerged MNA fix-GCC-warnings patch (sbcl-devel
2001-07-17), editing the source by hand
also made other tweaks to fix GCC warnings
fixed memory leak in wrapped_readlink()

0.6.12.53:
merged MNA port of rtoy COERCE and ARRAY-ELEMENT-TYPE
DEFOPTIMIZERs from CMU CL (sbcl-devel 2001-07-16)
merged MNA port of rtoy irrational math fixes (sbcl-devel
2001-07-16)
added MNA regression tests of irrational functions

0.6.12.54:
merged MNA port of cachopo COPY-READTABLE fix (sbcl-devel
2001-07-17)
merged MNA port of mai DESCRIBE-a-class fix (sbcl-devel
2001-07-17)

0.6.12.55:
added distclean.sh to remove stuff like
customize-target-features.lisp
DEFCONSTANT *FOO* now issues a STYLE-WARNING.
factored out LOOKS-LIKE-NAME-OF-SPECIAL-VAR-P to support this
removed call to no-longer-defined OUTPUT-INTERPRETED-FUNCTION
in PRINT-OBJECT method for INTERPRETED-FUNCTION (which
will hopefully go away real soon anyway)

0.6.12.56:
fixed bug 26: ARRAY-DISPLACEMENT now returns (VALUES NIL 0)
for undisplaced arrays.
fixed bug 55: DEFMACRO-MUNDANELY no longer ignores DOC.
DEFMACRO-MUNDANELY should be in SB-INT.

0.6.12.57:
set default for *DERIVE-FUNCTION-TYPES* to NIL (i.e. ANSI
behavior)
overrode the default within the cross-compiler, so that SBCL
itself is still built the old static efficient way

0.6.12.58:
fixed the DCONSING carry case in PROFILE
ported CMU CL fix to FILL-POINTER-OUCH (Tim Moore's fix
to Janne Rinta-Manty's bug report)
added MNA fix for FIX-CORE-SOURCE-INFO

2001-07-30 19:30:33 Tree
[a1a2c0] by William Harold Newman William Harold Newman

0.6.12.49:
made verbose GC output report GET-INTERNAL-RUN-TIME when each
GC happens, so that when you reading a log containing
many verbose GC messages, you have a better chance of
grokking what happened
fixed POSIX-GETENV doc string after Alexey Dejneka pointed out
that it was wrong

2001-07-30 16:15:21 Tree
[81cfdf] by William Harold Newman William Harold Newman

0.6.12.48:
made ldb conditional on :SB-LDB in *FEATURES*. (Dan Barlow
likes having the system stop on an ldb> prompt even
when running as a batch program, but for me it's a
nuisance. I want sh make.sh to terminate when it hits
an error, period, since typically I make it beep when
it's done. And since ldb isn't really appropriate for
the problems that ordinary users have to deal with,
it should be reasonable to suppress it by default,
which is an overkill way to keep my system from
stopping on an ldb> prompt.)
more gencgc tidying..
..hardwired enable_pointer_filter conditionalization to
always be enabled
..renamed the PTR() and Pointerp() macros to
native_pointer() and is_lisp_pointer() inline
functions

2001-07-14 04:25:14 Tree
[a946a4] by William Harold Newman William Harold Newman

0.6.12.47:
more poking around in gencgc, tweaking comments and formatting
and names
deleted "#ifdef CONTROL_STACKS" stuff

2001-07-13 03:39:38 Tree
[fbe6e2] by William Harold Newman William Harold Newman

0.6.12.46:
(got confused, thought GENCGC was excessively conservative,
tried to fix it, then backed out)
tweaked a lot of comments and some names in gencgc.c
made GENESIS systematically propagate *SHEBANG-FEATURES*
entries into #define's in sbcl.h. (This isn't currently
used for anything -- I wrote it to support my confused
make-GENCGC-less-conservative effort, which no longer
exists. However, it seems harmless and could be useful
e.g. for making C-level code depend on SB-SHOW, so I
left it in.)

2001-07-12 23:15:12 Tree
[6879a3] by William Harold Newman William Harold Newman

0.6.12.45:
(This version/commit actually includes some stuff from the Dave
MacDonald patch logged in the 0.6.12.44 message, since
CVS got stuck somehow and didn't finish that commit.)
(and also left a stale lock in place for days..)
some *EVAL-STACK* fixes..
..removed CURRENT-STACK-POINTER macro, since (1) its name is
confusing (suggesting the system stack, not the eval
stack; especially since the CURRENT-STACK-POINTER vop
*does* refer to the system stack!) and (2) abstracting
away the fact that this is a special variable is,
given the INTERPRETER-GC-HOOK hack to scrub the eval
stack, an unfortunate invitation to subtle GC bugs
..renamed STACK-COPY to EVAL-STACK-COPY
..made *EVAL-STACK-TRACE* stuff dependent on #!+SB-SHOW
..renamed EVAL-STACK-SET-TOP to EVAL-STACK-RESET-TOP
I also made *INTERNAL-APPLY-NODE-TRACE* stuff dependent on
#!+SB-SHOW, even though in retrospect that was
probably silly, since it's only in IR1 interpreter
code, which I hope to blow away next month anyway.

2001-07-11 00:03:02 Tree
[d40305] by William Harold Newman William Harold Newman

0.6.12.44:
merged Dave MacDonald patch (sbcl-devel 2001-07-04): cleanup
of COMPILE-FILE cruft left over from transition to
compiling only one file at a time; plus a few
WHN modifications..
..INDEX isn't actually ignored in FIND-SOURCE-ROOT
..(IF (PATHNAMEP X) X (PATHNAME X)) => (PATHNAME X)
..more-explicit names in general
..expanded remaining FIND-FILE-INFO call into something which
still works now that DEFUN FIND-FILE-INFO is gone
Also, while I'm at it, SOURCE-INFO-FILE-INFO is a better
name than SOURCE-INFO-FILE.

2001-07-06 13:30:59 Tree
[75d948] by William Harold Newman William Harold Newman

0.6.12.43:
made PEEK-CHAR check its argument type explicitly, so
it will issue a TYPE-ERROR on e.g.
(WITH-INPUT-FROM-STRING (S "SZ[19]")
(PEEK-CHAR S))
instead doing PEEK-CHAR on *STANDARD-INPUT*
fixed completely-broken bignum-sized-consing code path in
profile.lisp (Code coverage testing? What's that?:-)
merged MNA/CMUCL compiler bug collection (sbcl-devel
2001-06-25) to BUGS

2001-06-30 21:47:55 Tree
[b4dcc8] by William Harold Newman William Harold Newman

0.6.12.42:
more profiling stuff..
..rearranged DCONSING arithmetic in profile wrappers so that
accounting for small amounts of wrapped consing almost
always works without extra consing overhead in the
wrapper
..made calibrate-PROFILE-overhead code run longer
MNA pointed out that bug 59 was fixed a while ago.
MNA comment clarifications sbcl-devel 2001-06-25
removed bogus QUIT from stream.pure.lisp, so that the
subsequent *.pure.lisp tests get executed

2001-06-25 16:05:39 Tree
[4719b7] by William Harold Newman William Harold Newman

0.6.12.41:
removed various REMOVEME stuff
more profiling/PCOUNTER stuff..
..restored "grep ';.*decla.*type'" stuff in profile.lisp,
and did other un-fluidifications of pcounter.lisp
and profile.lisp
..converted various PCOUNTER-related DECLAIMs to DEFKNOWNs

2001-06-24 15:54:39 Tree
[ed1c91] by William Harold Newman William Harold Newman

0.6.12.40:
fixed bug 107: (WRITE #*101 :RADIX T :BASE 36) now does
the right thing.

2001-06-24 13:13:42 Tree
[4a466c] by William Harold Newman William Harold Newman

0.6.12.39:
removed some stuff checking for profile monotonicity (since
now I know that OpenBSD get-internal-run-time *isn't*
monotonic, since OpenBSD getrusage() isn't monotonic:
bug kernel/1065. I've now patched my OpenBSD 2.9 to
fix this enough to make GET-INTERNAL-RUN-TIME
monotonic, and I'll submit the patch to the OpenBSD
maintainers once I straighten out how to diff against
the patch branch.)
tweaked GET-BYTES-CONSED (aiming to minimize consing in
profile.lisp)..
..replaced *TOTAL-BYTES-CONSED* with
*N-BYTES-FREED-OR-PURIFIED-PCOUNTER*
..deleted *LAST-BYTES-IN-USE*
..added DEFKNOWN GET-BYTES-CONSED
..incompatible change: made GET-BYTES-CONSED return the
number of bytes consed since the system was started,
not since the first time the function was called
..renamed GC-COLD-INIT-OR-REINIT, since it's only actually
called at reinit time
fixed *BYTES-CONSED-BETWEEN-GCS* FIXME; and now
DEFAULT-BYTES-CONSED-BETWEEN-GCS no longer needs
to be a named definition
deleted unused *GC-INHIBIT-HOOK*

2001-06-24 01:32:55 Tree
[f3af39] by William Harold Newman William Harold Newman

0.6.12.38:
tweaked PCOUNTER stuff in anticipation of future changes
in GET-BYTES-CONSED..
..moved PCOUNTER code from profile.lisp to new pcounter.lisp
..exported PCOUNTER stuff from SB-INT

2001-06-23 16:26:55 Tree
[b33fd6] by William Harold Newman William Harold Newman

0.6.12.37;
fixed bug 24: COMPILE-FILE handles READER-ERROR now, instead
of dropping you into the debugger.
READ-FOR-COMPILE-FILE now fills the role of the old
CAREFUL-READ code.
deleted now-unused NORMAL-READ-ERROR, IGNORE-ERROR-FORM,
and UNEXPECTED-EOF-ERROR
made slam.sh use a crude make-ish timestamp hack to choose
which files to re-cross-compile

2001-06-22 23:34:21 Tree
[83fd55] by William Harold Newman William Harold Newman

0.6.12.36:
fixed DIRECTORY/TRUENAME/symlink failure as proposed by DB
(sbcl-devel 2001-06-13, plus following discussion),
mostly by tweaking UNIX-READ-LINKS
In tweaking UNIX-READ-LINKS, I also redid it so that it would
handle pathnames of any length.
rewrote UNIX-READLINK to use wrapped_readlink(), handling
paths of any length
incompatible change: The old CMU-CL-style DIRECTORY options
(:ALL, :FOLLOW-LINKS, and :CHECK-FOR-SUBDIRS) are no
longer supported, so DIRECTORY always does the abstract
Common-Lisp-y thing, i.e. :ALL T :FOLLOW-LINKS T
:CHECK-FOR-SUBDIRS T.
added some DIRECTORY and TRUENAME test cases
fixed stupid gross errors in 0.6.12.35 test cases (Evidently I
neglected to run them before..)

2001-06-22 16:21:22 Tree
[e37366] by William Harold Newman William Harold Newman

0.6.12.35:
null-*PRINT-LENGTH* bugfix in structure printing (from Alexey
Dejneka sbcl-devel 2001-06-14)
merged Eric Marsden ANSI compliance fixes (from cmucl-imp
2001-06-15); except that I skipped the ones in
eval-comp.lisp because I hope to get rid of the IR1
interpreter completely sometime in the next few months,
and I skipped COMPILER-ERROR-is-PROGRAM-ERROR because
(1) I think not all compiler errors are program errors
(e.g. type errors or i/o errors) and (2) SBCL mostly
handles COMPILER-ERRORs before the user sees them
anyway, so I couldn't think of a test case where this
matters

2001-06-16 13:18:13 Tree
[7a5fb7] by William Harold Newman William Harold Newman

0.6.12.34:
belatedly bumped fasl file version number because the inline
type test patches changed the layout of STREAM
Bug 23 is fixed, so delete it from BUGS.
merged MNA fixes for tests/interface.lisp (sbcl-devel
2001-06-12)

2001-06-13 17:19:45 Tree
Older >