Taylor R Campbell [Mon, 13 Sep 2010 15:34:17 +0000 (15:34 +0000)]
Simplify idiom for implementing dynamic binding.
Taylor R Campbell [Mon, 13 Sep 2010 15:28:31 +0000 (15:28 +0000)]
New macro BEGIN0.
(BEGIN0 <expression> <command> ...) evaluates <expression>, executes
<command> ..., and then returns the value of <expression>.
(Cf. Common Lisp's PROG1.)
Arthur A. Gleckler [Thu, 9 Sep 2010 06:15:43 +0000 (23:15 -0700)]
Moved declaration inside a block in order to conform to C89.
(Thanks to Taylor for pointing out the need for this.)
Arthur A. Gleckler [Tue, 7 Sep 2010 15:03:28 +0000 (08:03 -0700)]
Use <error_unimplemented_primitive> in case uint64 is not available.
Arthur A. Gleckler [Tue, 7 Sep 2010 05:57:30 +0000 (22:57 -0700)]
Conditionalized compilation of the flonum casting primitives.
The primitives `cast-ieee754-double-to-integer' and
`cast-integer-to-ieee754-double' are omitted when the type
<uint64_t> is not available.
Arthur A. Gleckler [Mon, 6 Sep 2010 05:32:55 +0000 (22:32 -0700)]
Added <AC_INT64_T> and <AC_UINT64_T> to "configure.ac". The latter is
now required by "flonum.c" because of the new primitives I added, and
CPH asked me to include the former, too.
Taylor R Campbell [Sun, 5 Sep 2010 18:57:11 +0000 (18:57 +0000)]
Shrink hashtb.{bin,com,bci} further with judicious use of syntax.
Taylor R Campbell [Sun, 5 Sep 2010 18:23:04 +0000 (18:23 +0000)]
Work around bug that makes DECLARE and SYNTAX-RULES not mix.
Taylor R Campbell [Sun, 5 Sep 2010 05:04:47 +0000 (05:04 +0000)]
Fix total botch of last commit.
Make the GC cleverer about reallocating the ephemeron array, in order
to avoid potentially quadratic-time behaviour in MAKE-EPHEMERON and
fasloading files with ephemerons in them.
Now fasls and bands with ephemerons in them should really work, this
time with feeling...
Taylor R Campbell [Sat, 4 Sep 2010 05:10:17 +0000 (05:10 +0000)]
New fasl version to support ephemerons.
In the process, complete the transition to the STACK_END fasl format.
In the EPHEMERON fasl format, the fasl header has an extra field for
the number of ephemerons stored in the fasl, for which the fasloader
reserves space in ephemeron_array.
The fasdumper chooses between the C_CODE, STACK_END, or EPHEMERON
fasl format for maximum compatibility:
- If there are any ephemerons in the fasl, the fasdumper chooses the
EPHEMERON format. Older microcodes don't know about ephemerons and
thus can't handle such fasls anyway.
- If dumping a band, the fasdumper chooses the STACK_END format,
since the only differences between the C_CODE format and the
STACK_END format matter only for bands. Support for reading the
STACK_END format was added in version 15 of the microcode; any
newly created bands are not likely to be used in older microcodes
than that anyway.
- Otherwise, the fasdumper chooses the C_CODE format, like before.
Arthur A. Gleckler [Fri, 3 Sep 2010 03:57:08 +0000 (20:57 -0700)]
Removed NaN tests because they cause traps on Linux. (It turns out
that they are supposed to on OS X, too, but a bug prevents that).
Taylor is considering adding support for controlling the
floating-point exception mask and flags, in which case we'll be able
to manipulate NaN values and I can add these tests back.
Arthur A. Gleckler [Thu, 2 Sep 2010 06:56:41 +0000 (23:56 -0700)]
Added tests for infinities.
Arthur A. Gleckler [Thu, 2 Sep 2010 05:57:36 +0000 (22:57 -0700)]
Fixed typo in comments.
Arthur A. Gleckler [Thu, 2 Sep 2010 05:43:46 +0000 (22:43 -0700)]
Added trivial tests for conversion to floating-point infinities by
`cast-integer-to-ieee754-single' and `cast-integer-to-ieee754-double'.
Arthur A. Gleckler [Thu, 2 Sep 2010 05:29:34 +0000 (22:29 -0700)]
Fixed code formatting.
Arthur A. Gleckler [Thu, 2 Sep 2010 05:23:39 +0000 (22:23 -0700)]
* Renamed `cast-flonum-to-integer' to `cast-ieee754-double-to-integer'.
* Renamed `cast-integer-to-flonum' to `cast-integer-to-ieee754-double'.
* Changed both to use <uint64_t> instead of <unsigned long> for
conversion. That way, they should work on 32-bit systems. (Thanks to
Taylor for pointing out that bug.)
* Added `cast-ieee754-single-to-integer' and `cast-integer-to-ieee754-single'.
* Updated existing tests.
* Added tests for casts to and from single-precision floating-point numbers.
Arthur A. Gleckler [Thu, 2 Sep 2010 03:48:14 +0000 (20:48 -0700)]
Removed the divisions by zero that I had used to create floating-point
infinities since those are supposed to trap, not actually return
infinities. On OS X, the trap doesn't happen, but I shouldn't depend
on a bug.
Arthur A. Gleckler [Wed, 1 Sep 2010 04:17:30 +0000 (21:17 -0700)]
Merge branch 'master' of git://git.savannah.gnu.org/mit-scheme
Arthur A. Gleckler [Wed, 1 Sep 2010 04:16:03 +0000 (21:16 -0700)]
Added primitives `cast-flonum-to-integer' and `cast-integer-to-flonum'
and tests.
Joe Marshall [Tue, 31 Aug 2010 20:25:58 +0000 (13:25 -0700)]
Softer landing for system-library-directory-pathname if sudirectory doesn't exist.
Taylor R Campbell [Tue, 31 Aug 2010 15:26:30 +0000 (15:26 +0000)]
Merge branch 'master' of ssh://git.sv.gnu.org/srv/git/mit-scheme
Taylor R Campbell [Tue, 31 Aug 2010 15:25:00 +0000 (15:25 +0000)]
Exploit OBJECT->FIXNUM's independence of tag on i386 and x86-64.
Taylor R Campbell [Tue, 31 Aug 2010 15:24:00 +0000 (15:24 +0000)]
Kill infinite parsing loop by using + inside *, not * inside *.
Arthur A. Gleckler [Tue, 31 Aug 2010 05:15:15 +0000 (22:15 -0700)]
Fixed comment in shadowing test.
Taylor R Campbell [Mon, 30 Aug 2010 21:02:38 +0000 (21:02 +0000)]
Merge branch 'master' of ssh://git.sv.gnu.org/srv/git/mit-scheme
Chris Hanson [Mon, 30 Aug 2010 09:10:55 +0000 (02:10 -0700)]
Fix missing -f in test.
Chris Hanson [Mon, 30 Aug 2010 06:25:17 +0000 (23:25 -0700)]
Add unit test to confirm that shadowing fix worked.
Chris Hanson [Mon, 30 Aug 2010 05:52:55 +0000 (22:52 -0700)]
Merge branch 'master' of git.sv.gnu.org:/srv/git/mit-scheme
Taylor R Campbell [Sun, 29 Aug 2010 17:28:53 +0000 (17:28 +0000)]
Extend hash table entry types to support ephemeral hash tables.
Add some tests for correctness against red/black trees.
Still missing are tests for weak and ephemeral entries types.
Taylor R Campbell [Sat, 28 Aug 2010 20:58:19 +0000 (20:58 +0000)]
Fix SVM's write_cc_entry_offset.
Since this is used only to initialize trampolines, the starting
instruction is irrelevant, and will be garbage anyway because the
trampoline blocks are always freshly allocated.
Taylor R Campbell [Sat, 28 Aug 2010 20:56:36 +0000 (20:56 +0000)]
Fix SVM interpreter's INDEX-FIXNUM? predicate.
Taylor R Campbell [Sat, 28 Aug 2010 20:56:03 +0000 (20:56 +0000)]
Fix SVM machine's conditional floating-point jumps.
These want to use the float registers, not the word registers.
Taylor R Campbell [Sat, 28 Aug 2010 20:55:04 +0000 (20:55 +0000)]
Fix SVM interpreter's decoding of doubles.
They are encoded as 64-bit significand and 16-bit exponent, not as
machine (IEEE 754) doubles.
Taylor R Campbell [Sat, 28 Aug 2010 20:52:26 +0000 (20:52 +0000)]
In FLO:VECTOR-LENGTH, coerce length to fixnum before OBJECT->FIXNUM.
The previous open-coding worked whenever OBJECT->FIXNUM merely
shifted left by the tag width, ignoring whatever tag was there
before. The SVM does not do this, however.
Taylor R Campbell [Sat, 28 Aug 2010 20:49:33 +0000 (20:49 +0000)]
Omit continuation code words after interpreter cache references.
The RTL already arranges their continuations to follow them; putting
an extra continuation code word causes the machine to try to execute
the real one as if it were an instruction.
Taylor R Campbell [Sat, 28 Aug 2010 20:48:47 +0000 (20:48 +0000)]
PRE-INCREMENT is only ever a memory reference, not address, in RTL.
Taylor R Campbell [Sat, 28 Aug 2010 20:47:08 +0000 (20:47 +0000)]
Use NON-POINTER-OBJECT? and CAREFUL-OBJECT-DATUM in the compiler.
Using OBJECT-NON-POINTER? causes primitives to be put into blocks as
non-marked constants, so that the fasloader fails to renumber them.
Taylor R Campbell [Mon, 23 Aug 2010 19:13:02 +0000 (19:13 +0000)]
Update `last update' dates on the reference and user manuals.
Taylor R Campbell [Mon, 23 Aug 2010 18:58:51 +0000 (18:58 +0000)]
Document the stack sampler.
Chris Hanson [Mon, 23 Aug 2010 09:20:00 +0000 (02:20 -0700)]
Draft fix for over-shadowing bug.
Taylor R Campbell [Fri, 20 Aug 2010 02:57:36 +0000 (02:57 +0000)]
Add tests for ephemerons with compiled entries for keys and data.
Taylor R Campbell [Fri, 20 Aug 2010 02:57:01 +0000 (02:57 +0000)]
Fix bug in the GC's ephemeron scanning.
When queueing ephemerons for the key stored at some address, compare
the address to the weak referent address of each ephemeron's key --
not to the object address of each ephemeron's key, which causes us to
miss ephemerons whose keys are compiled entries, because we pass the
addresses of compiled blocks, never those of the compiled entries
stored in them, to queue_ephemerons_for_key.
Taylor R Campbell [Fri, 20 Aug 2010 02:45:31 +0000 (02:45 +0000)]
Minor fixes to unit-testing.scm.
- Update copyright years.
- Ignore unused variable CONDITION in ASSERT-ERROR.
- Use RUN-SUB-TEST to run test procedures in RUN-SUB-TESTS
The last one is probably wrong, but without it, running unit tests
crashes in a bizarre way for me. Example:
(run-unit-tests "runtime/test-ephemeron")
;Loading "runtime/test-ephemeron.com"... done
no-gc: failed 2 sub-tests out of 2 in .32 seconds:
.0.0:
;Ill-formed failure: key
Taylor R Campbell [Fri, 20 Aug 2010 02:43:13 +0000 (02:43 +0000)]
Omit needless definition of REFERENCE-BARRIER in test-ephemeron.scm.
Due to a bug in the microcode's environment operations, this
definition, which should be harmless, actually renders the tests
unrunnable because it makes REFERENCE-BARRIER spin in an infinite
loop.
Taylor R Campbell [Thu, 19 Aug 2010 02:55:13 +0000 (02:55 +0000)]
Eliminate losing bisection crock in find_block_address.
This wasted several hours of my day in exchange for several
milliseconds of time (`on average') searching through the heap
after a trap that will trigger an interactive debugger anyway
and sit waiting for I/O.
Taylor R Campbell [Thu, 19 Aug 2010 02:46:02 +0000 (02:46 +0000)]
Fix plausible_cc_block_p some more.
Now it detects COLUMN->Y in edwin/bufwmc.scm.
Taylor R Campbell [Wed, 18 Aug 2010 16:30:25 +0000 (16:30 +0000)]
Add a couple more consts to debug.c.
Taylor R Campbell [Wed, 18 Aug 2010 16:19:14 +0000 (16:19 +0000)]
Merge branch 'master' of ssh://git.sv.gnu.org/srv/git/mit-scheme
Joe Marshall [Wed, 18 Aug 2010 00:02:48 +0000 (17:02 -0700)]
Merge branch 'master' of ssh://git.sv.gnu.org/srv/git/mit-scheme
Joe Marshall [Wed, 18 Aug 2010 00:02:05 +0000 (17:02 -0700)]
Add appropriate const qualifiers for anal compiler.
Taylor R Campbell [Tue, 17 Aug 2010 19:05:16 +0000 (19:05 +0000)]
Remove spurious ignore declaration in i386 GENERATE/CLOSURE-HEADER.
Taylor R Campbell [Tue, 17 Aug 2010 18:46:26 +0000 (18:46 +0000)]
Implement and document SET-EPHEMERON-KEY!.
Taylor R Campbell [Mon, 16 Aug 2010 22:51:45 +0000 (22:51 +0000)]
New procedure REFERENCE-BARRIER.
(REFERENCE-BARRIER <x>) guarantees that the garbage collector will
assume <x> to be live at least until the call to REFERENCE-BARRIER,
if control can ever reach it.
This is implemented as an alias for the identity procedure currently,
but a more efficient implementation not involving a general procedure
call is conceivable for the future.
Taylor R Campbell [Fri, 13 Aug 2010 20:58:42 +0000 (20:58 +0000)]
Ignore out-of-date uncompressed debugging info files.
Fixes
(load "x")
(pp foo)
;Output: (named-lambda (foo x y z) ...)
;; Edit x.scm and recompile.
(load "x")
(pp foo)
;Output: #[compiled-procedure foo ...]
Taylor R Campbell [Fri, 13 Aug 2010 20:25:58 +0000 (20:25 +0000)]
New stack-sampling profiler.
For usage, see runtime/stack-sample.scm. May be worthwhile to write
documentation in the user manual.
This is *not* a replacement for a PC-sampler, which can give
information that is more precise at one level and less precise at
another level. See the comments for details.
Taylor R Campbell [Fri, 13 Aug 2010 20:18:56 +0000 (20:18 +0000)]
Rework the hash table entry abstraction.
This simplifies some of the code (at the expense of a longer
definition for the abstraction leading to a longer hashtb.scm
altogether), and makes it less prone to mistakes with using the
keys and data of entries without checking their validity -- which
will matter especially for hash tables with ephemeron entries when
those are implemented.
Check the results of the key hash table function. Sprinkle some
guarantees throughout the code. Disable type and range checks in
the hash table methods, where it is safe to do so.
With type and range checks still enabled, performance on strong eq
and eqv hash tables is no different from what it was last week.
Performance on weak hash tables is ever so slightly worse, but that
is because weak hash tables were incorrect before.
New tests check for some regressions.
Taylor R Campbell [Fri, 13 Aug 2010 04:22:46 +0000 (04:22 +0000)]
Protect each use of ENTRY-{KEY,DATUM} in hashtb.scm by ENTRY-VALID?.
Taylor R Campbell [Thu, 12 Aug 2010 22:29:28 +0000 (22:29 +0000)]
Fix SRFI 69 HASH-TABLE-FOLD.
The specification makes no mention of any constraint that the combine
procedure not update the hash table, so we must be prepared for that
case. The old definition for HASH-TABLE-FOLD now has the new name
%HASH-TABLE-FOLD, used internally by the procedures that need to fold
but need not update the hash table while doing so.
Joe Marshall [Thu, 12 Aug 2010 22:00:06 +0000 (15:00 -0700)]
Restore warning as per Taylor.
Joe Marshall [Wed, 11 Aug 2010 16:45:10 +0000 (09:45 -0700)]
Add declarations to silence warnings.
Joe Marshall [Wed, 11 Aug 2010 16:43:17 +0000 (09:43 -0700)]
Add declaration to silence warning.
Taylor R Campbell [Tue, 10 Aug 2010 19:58:30 +0000 (19:58 +0000)]
Document ephemerons.
Taylor R Campbell [Tue, 10 Aug 2010 18:43:44 +0000 (18:43 +0000)]
Implement ephemerons.
Ephemerons are like weak pairs, except the datum is dropped if the
key is dropped, and references to the key through the datum don't
count if the only references to the datum are through the ephemeron.
In other words, the weak references to the key and datum are dropped
iff the key can be proven dead; conversely, the references to the key
and datum are preserved iff somebody else cares about the key.
Joe Marshall [Mon, 9 Aug 2010 19:55:47 +0000 (12:55 -0700)]
Declare IGNORABLE NB to silence warning.
Taylor R Campbell [Mon, 9 Aug 2010 17:45:10 +0000 (17:45 +0000)]
Missed some vestiges of archaic microcode state space support.
Taylor R Campbell [Mon, 9 Aug 2010 17:38:57 +0000 (17:38 +0000)]
Eliminate some unused return codes.
Taylor R Campbell [Mon, 9 Aug 2010 16:47:03 +0000 (16:47 +0000)]
Eliminate all traces of the archaic microcode state space support.
Taylor R Campbell [Mon, 9 Aug 2010 16:44:03 +0000 (16:44 +0000)]
Set -u in the setup scripts and make them more robust to interrupts.
Taylor R Campbell [Mon, 9 Aug 2010 16:36:32 +0000 (16:36 +0000)]
Check for termcap.h too.
Assume that term.h or termcap.h defines BC, UP, PC, & ospeed.
Work around !*@#@^!*#%!^&$@!%# stupidity in NetBSD with ospeed.
Taylor R Campbell [Sun, 8 Aug 2010 02:56:56 +0000 (02:56 +0000)]
Move the check for termcap-related headers.
This puts the warning messages about emulating termcap or tparm
closer to the checks, which makes configure's output make a trifle
more sense.
Taylor R Campbell [Sun, 8 Aug 2010 01:53:07 +0000 (01:53 +0000)]
Add configure option `--with-termcap=LIB' to select termcap library.
LIB=no disables any termcap library (causing the use of the built-in
termcap emulation); LIB=yes, the default, makes configure choose
whatever termcap library looks best.
I wanted to go further and make the termcap primitives dynamically
loadable, so that the compile-time choice of termcap library can be
made just for some prtermcap.so module separately from the libraries
against which the microcode is linked, but uxtty.c uses termcap for
screen-clearing and window-sizing.
If ^L worked everywhere to clear the screen, and TIOCGWINSZ worked
everywhere to find the screen dimensions, then we could get rid of
those uses of termcap and move it off to a module. But alas, that is
not the case. E.g., ^L doesn't clear the screen in screen.
Taylor R Campbell [Fri, 6 Aug 2010 01:36:09 +0000 (01:36 +0000)]
Fix i386 code generated for flonum rounding primitives.
The default has not been round-to-even since primitives were added to
change it, but the generated code assumed it was.
Remember to run at least a two-stage build so that the runtime is
compiled with the new compiler.
Before:
(flo:with-rounding-mode 'TOWARD-ZERO (lambda () (flo:ceiling 1.2)))
;Value: 1.
After:
(flo:with-rounding-mode 'TOWARD-ZERO (lambda () (flo:ceiling 1.2)))
;Value: 2.
Taylor R Campbell [Fri, 6 Aug 2010 01:18:01 +0000 (01:18 +0000)]
Emulate C99 fenv.h on amd64 through the MXCSR.
Taylor R Campbell [Fri, 6 Aug 2010 01:17:07 +0000 (01:17 +0000)]
Emulate C99 fenv.h on i386 through the i387 control word.
Taylor R Campbell [Fri, 6 Aug 2010 01:15:30 +0000 (01:15 +0000)]
Emulate C99 fenv.h with BSD ieeefp.h if necessary.
Taylor R Campbell [Fri, 6 Aug 2010 00:38:40 +0000 (00:38 +0000)]
New procedures FLO:EXPM1 and FLO:LOG1P just like libm's.
Defined only on an interval about 0 of radius 1/log 2 and 1/sqrt 2,
respectively; intended for computing (exp x) - 1 and log (1 + x) for
very small x with high precision, unlike FLO:EXP and FLO:LOG.
Taylor R Campbell [Thu, 5 Aug 2010 18:57:53 +0000 (18:57 +0000)]
Be slightly more consistent about references to MIT_SCHEME_EXE.
It is to be interpreted as a pathname, provided that no character in
it be a make meta-character, or a shell meta-character that has
meaning within double-quotes. It shouldn't have such constraints,
but Unix sucks. To set the library path, use MITSCHEME_LIBRARY_PATH
rather than sticking a `--library' argument in MIT_SCHEME_EXE.
Taylor R Campbell [Thu, 5 Aug 2010 17:15:17 +0000 (17:15 +0000)]
Ensure that signal handlers see the C stack, not the Scheme stack.
Do this by wrapping all the signal handlers in stubs that call an
assembly hook to make the stack pointer point into the C stack rather
than the Scheme stack if necessary. To indicate that this is not
necessary, define SIGNAL_HANDLERS_CAN_USE_SCHEME_STACK. For now, I'm
leaving that undefined by default, because it is the safer option.
This solves a problem on operating systems such as NetBSD that store
the current pthread identifier in the stack pointer. When Scheme's
signal handler calls routines that are pthread cancellation points,
such as waitpid, they try to find the current pthread identifier in a
stack pointer that points off into oblivion (into Scheme's stack) and
promptly crash -- or, worse, trigger SIGSEGV, to be handled by a
signal handler while the stack pointer still points into Scheme's
stack, with the same problem.
I am told that this will be fixed in NetBSD 6 (since it interferes
not just with Scheme but also with sigaltstack, makecontext, and
anything else that wants to mess with the stack pointer), but only on
i386 and amd64 for certain, and in any case, this workaround will
work on any other systems that try to use the same trick to store the
current pthread identifier, of which I believe there may be several.
(E.g., older versions of GNU/Linux with LinuxThreads.)
Taylor R Campbell [Tue, 3 Aug 2010 22:35:33 +0000 (22:35 +0000)]
Use intmax and uintmax to support 64-bit off_t, time_t, &c.
Eliminate BIGNUM_NO_ULONG. Twenty years ago it may have made sense;
now too much of the system relies on having unsigned longs for it to
matter.
New routines convert between integers and intmax/uintmax. The old
ones for long/unsigned long are still there, because they are likely
to be much faster on 32-bit systems. Only a few uses of long have
been replaced by intmax -- the ones where it was obvious we were
converting between time_t, off_t, ino_t, &c., and long. Others may
still be there. Before:
(file-length "fnord")
;Value: 1235
After:
(file-length "fnord")
;Value:
4294968531
(`fnord' was created by seeking to byte 2^32 + 1234 or something and
writing a single byte there.)
Chris Hanson [Wed, 28 Jul 2010 06:53:46 +0000 (23:53 -0700)]
Don't delete runtime.com from the distribution.
Chris Hanson [Wed, 28 Jul 2010 05:27:12 +0000 (22:27 -0700)]
Merge branch 'master' of git.sv.gnu.org:/srv/git/mit-scheme
Joe Marshall [Tue, 27 Jul 2010 19:33:58 +0000 (12:33 -0700)]
Implement SYMBOL>?
Taylor R Campbell [Tue, 27 Jul 2010 16:52:43 +0000 (16:52 +0000)]
Don't make lib/lib a symlink to ../microcode; this breaks liarc.
Chris Hanson [Tue, 27 Jul 2010 05:15:40 +0000 (22:15 -0700)]
Properly fix handling of pstring.
Chris Hanson [Tue, 27 Jul 2010 05:01:52 +0000 (22:01 -0700)]
Fix typo.
Taylor R Campbell [Tue, 27 Jul 2010 01:56:36 +0000 (01:56 +0000)]
Fix microcode's `--version' output.
Write a newline, and force console output.
Chris Hanson [Mon, 26 Jul 2010 07:18:36 +0000 (00:18 -0700)]
Fix bug: change to MODULE_LDFLAGS broke liarc-ld on MacOS.
Chris Hanson [Mon, 26 Jul 2010 07:17:34 +0000 (00:17 -0700)]
Nuke trailing whitespace.
Chris Hanson [Sun, 25 Jul 2010 07:54:24 +0000 (00:54 -0700)]
Broaden handling of pstring to accept NIL.
Taylor R Campbell [Fri, 23 Jul 2010 18:33:32 +0000 (18:33 +0000)]
Use _init and etext on NetBSD to find the text segment's bounds.
Taylor R Campbell [Fri, 23 Jul 2010 02:30:25 +0000 (02:30 +0000)]
Eliminate vestiges of i386 code from cmpintmd/x86-64.h.
Taylor R Campbell [Fri, 23 Jul 2010 01:28:59 +0000 (01:28 +0000)]
Unmask IEEE 754 floating-point exceptions on x86-64.
Should this instead be part of Scheme's dynamic environment, like the
floating-point rounding mode?
Taylor R Campbell [Thu, 22 Jul 2010 23:12:37 +0000 (23:12 +0000)]
Relax plausible_cc_block_p.
Some leaf procedures and continuations do not have GC trap prefixes,
so don't require the block in question to have one.
On i386, this makes the trap handler successfully figure out where it
is in (FIX:QUOTIENT 1 0). FIX:QUOTIENT is a leaf procedure, unlike
FLO:/, although perhaps FIX:QUOTIENT ought to do checks and defer to
the microcode primitive like FLO:/.
Taylor R Campbell [Thu, 22 Jul 2010 21:16:08 +0000 (21:16 +0000)]
Merge branch 'master' of ssh://git.sv.gnu.org/srv/git/mit-scheme
Taylor R Campbell [Thu, 22 Jul 2010 20:25:19 +0000 (20:25 +0000)]
Reorganize Linux section of uxtrap.h and add x86-64 support.
Also add signal codes for SIGSEGV, SIGBUS, and SIGTRAP.
Taylor R Campbell [Thu, 22 Jul 2010 20:23:04 +0000 (20:23 +0000)]
Exploit common mcontext structure shared between NetBSD ports.
Taylor R Campbell [Thu, 22 Jul 2010 20:20:53 +0000 (20:20 +0000)]
Omit explicit disabling of trap recovery on x86-64.
There are still no platforms that support it, but it doesn't need to
be explicitly disabled.
Joe Marshall [Thu, 22 Jul 2010 20:16:12 +0000 (13:16 -0700)]
Restore the classical exit message.
Taylor R Campbell [Thu, 22 Jul 2010 19:51:14 +0000 (19:51 +0000)]
Enable trap recovery on NetBSD.
Not really tested, but this triggers only when you're already
screwed, so enabling trap recovery probably won't make anything
worse.
Taylor R Campbell [Thu, 22 Jul 2010 17:28:07 +0000 (17:28 +0000)]
Improve trap handling under NetBSD.