mit-scheme.git
11 years agomd5: A separately buildable FFI wrapper.
Matt Birkholz [Sun, 8 Sep 2013 01:46:56 +0000 (18:46 -0700)]
md5: A separately buildable FFI wrapper.

11 years agoblowfish: Random cleanup. Use guarantee-substring.
Matt Birkholz [Sat, 7 Sep 2013 23:56:48 +0000 (16:56 -0700)]
blowfish: Random cleanup.  Use guarantee-substring.

11 years agosrc/etc/std-makefile-prefix: Punt SHIM variables.
Matt Birkholz [Sat, 7 Sep 2013 23:50:44 +0000 (16:50 -0700)]
src/etc/std-makefile-prefix: Punt SHIM variables.

FFIs are no longer built inside the core build tree.

11 years agoinstall-shim: Create directory in first directory of library path.
Matt Birkholz [Sat, 7 Sep 2013 23:30:18 +0000 (16:30 -0700)]
install-shim: Create directory in first directory of library path.

Also, locate the directory for the -I CFLAG by searching the library
path for mit-scheme.h.

11 years agowith-system-library-directories: prepend to library directory path.
Matt Birkholz [Sat, 7 Sep 2013 22:59:44 +0000 (15:59 -0700)]
with-system-library-directories: prepend to library directory path.

This is for test scripts that want a new FFI shim to shadow an
installed shim, and requires canonicalize-debug-info-pathname to be
fixed to deal -- to SEARCH the library-directory-path.  At least, for
relative pathnames with at least one directory name, it now searches
the directory path for that name.

11 years agoblowfish: Typo in blowfish.cdecls.
Matt Birkholz [Sat, 7 Sep 2013 04:36:17 +0000 (21:36 -0700)]
blowfish: Typo in blowfish.cdecls.

11 years agosrc/Makefile.in: Punt $(FFIS) and build-ffis target.
Matt Birkholz [Sat, 7 Sep 2013 04:33:06 +0000 (21:33 -0700)]
src/Makefile.in: Punt $(FFIS) and build-ffis target.

Moved the test wrapper to tests/ffi/ though it is still commented out
of tests/check.scm.

11 years agogdbm: Update README. Fix test. Nomenclature.
Matt Birkholz [Fri, 6 Sep 2013 23:55:50 +0000 (16:55 -0700)]
gdbm: Update README.  Fix test.  Nomenclature.

Add installation instructions from the previous commit message.  Test
with the newly built shim (not an installed shim).  Call the Scheme
code that interfaces to C "the wrapper", as in blowfish/README.

11 years agoblowfish: A separately buildable FFI wrapper.
Matt Birkholz [Fri, 6 Sep 2013 23:52:14 +0000 (16:52 -0700)]
blowfish: A separately buildable FFI wrapper.

11 years agogdbm: Add generate-shim. Drop compile-bundle and install-bundle.
Matt Birkholz [Thu, 5 Sep 2013 17:19:56 +0000 (10:19 -0700)]
gdbm: Add generate-shim.  Drop compile-bundle and install-bundle.

LIAR/C's bundles cannot be built outside the core build tree.  So
compile-bundle is just e.g. compile.scm, and install-bundle is merged
with install-shim.  Generate-shim actually handles the whole process
from *.cdecl's to -shim.c and -const.bin, eliminating 4 tedious little
rules from every wrapper's Makefile.

11 years agogdbm: A separately buildable FFI wrapper.
Matt Birkholz [Sun, 1 Sep 2013 01:12:05 +0000 (18:12 -0700)]
gdbm: A separately buildable FFI wrapper.

This is a drop-in replacement for the gdbm microcode module and
runtime/gdbm.scm.  Run `make install' and install the following in
your optiondb.scm.

(define-load-option 'GDBM2 (guarded-system-loader '(gdbm) "gdbm"))

You will need to call it GDBM2 while GDBM refers to the original
microcode module wrapper.

11 years agoffi/build.scm.in: Support building shims separately.
Matt Birkholz [Sun, 1 Sep 2013 01:10:40 +0000 (18:10 -0700)]
ffi/build.scm.in: Support building shims separately.

A handful of new top-level bindings make a portickle Makefile.

gdbm-shim.so: gdbm-shim.o gdbm-adapter.o
echo "(link-shim)" | mit-scheme --batch-mode -- -o $@ $^ -lgdbm

The new procedures (generate-shim, compile-shim, link-shim and
install-shim) autoload ffi/build.scm, which captured the build
configuration of the machine.  Compile-bundle and install-bundle are
also provided.

11 years agoffi: Search the library path for shims and their C types/const.
Matt Birkholz [Sat, 31 Aug 2013 22:18:27 +0000 (15:18 -0700)]
ffi: Search the library path for shims and their C types/const.

11 years agoInstall all .pkd, not just <various>-unx.pkd.
Matt Birkholz [Sat, 31 Aug 2013 22:13:48 +0000 (15:13 -0700)]
Install all .pkd, not just <various>-unx.pkd.

11 years agocref: (package-set-pathname "dir/") => "dir/dir-OS.pkd"
Matt Birkholz [Sat, 31 Aug 2013 22:10:22 +0000 (15:10 -0700)]
cref: (package-set-pathname "dir/") => "dir/dir-OS.pkd"

11 years agocref: (global-definitions symbol) looks in library path.
Matt Birkholz [Sat, 31 Aug 2013 22:05:40 +0000 (15:05 -0700)]
cref: (global-definitions symbol) looks in library path.

(global-definitions "string") is still relative to the .pkg file's
directory.

11 years agoruntime/unxpth: Remove ./ when simplifying.
Matt Birkholz [Sat, 31 Aug 2013 21:59:26 +0000 (14:59 -0700)]
runtime/unxpth: Remove ./ when simplifying.

11 years agoCapitalize word after `Caution:'.
Taylor R Campbell [Fri, 30 Aug 2013 15:52:18 +0000 (15:52 +0000)]
Capitalize word after `Caution:'.

11 years agotexinfo strong note
Barak A. Pearlmutter [Mon, 12 Aug 2013 13:22:26 +0000 (15:22 +0200)]
texinfo strong note

Mollify makeinfo with regard to this issue:

$ makeinfo --output=mit-scheme-user.info user.texinfo
user.texinfo:1552: warning: @strong{Note...} produces a spurious cross-reference in Info; reword to avoid that.

The relevant portion of the manual states:

    Caution: Do not use `@strong' with the word `Note'; Info will
    mistake the combination for a cross reference.  (It's usually
    redundant, anyway.)  Use a phrase such as *Please notice* or
    *Caution* instead, or ...

11 years agoUse `texi2dvi --pdf', not `texi2pdf'.
Taylor R Campbell [Fri, 30 Aug 2013 15:45:21 +0000 (15:45 +0000)]
Use `texi2dvi --pdf', not `texi2pdf'.

No functional difference, but texi2pdf is a newer utility not found
in older versions of GNU Texinfo that are still prevalent.

11 years agoFix missing * in one definition of OS_nanotime_since_utc_epoch.
Taylor R Campbell [Fri, 30 Aug 2013 15:32:10 +0000 (15:32 +0000)]
Fix missing * in one definition of OS_nanotime_since_utc_epoch.

From Igor Pashev.

11 years agoMake `news-select-buffer' select in other window by default.
Taylor R Campbell [Fri, 30 Aug 2013 15:30:20 +0000 (15:30 +0000)]
Make `news-select-buffer' select in other window by default.

From Rich Loveland.

11 years agoShow procedure name too in Edwin `show-parameter-list' command.
Taylor R Campbell [Fri, 30 Aug 2013 15:27:12 +0000 (15:27 +0000)]
Show procedure name too in Edwin `show-parameter-list' command.

From Rich Loveland.

11 years agosvm: Add lost flonum-pred-2-args rule.
Matt Birkholz [Sat, 24 Aug 2013 22:03:06 +0000 (15:03 -0700)]
svm: Add lost flonum-pred-2-args rule.

The "new" rule escaped the copying from i386/rulflo.scm(?).  Also
added symlink src/lib/compiler, useful when debugging a broken build.

11 years agoRemove logic to lock IMAIL's IMAP folder cache.
Taylor R Campbell [Sat, 10 Aug 2013 21:11:17 +0000 (21:11 +0000)]
Remove logic to lock IMAIL's IMAP folder cache.

Keep WITH-FOLDER-LOCKED as an empty stub so we might use it later to
let the user control whether to use a cache at all.

There are no invariants that need to be preserved, so locking only
slows us down and reduces concurrency.  Two clients fetching the same
item may collide, but the collision is harmless -- unless the server
is bogus, they will be writing the same data.  Adding an entry to the
cache is already atomic (unless the system fails before the blocks
have hit the disk) through the use of a temporary directory.

11 years agoDon't assume there is a current thread when interrupted.
Taylor R Campbell [Mon, 1 Jul 2013 15:15:25 +0000 (15:15 +0000)]
Don't assume there is a current thread when interrupted.

Fixes error `No current thread!' when the thread timer interrupt
handler tries to find the interrupted thread's floating-point
environment and there is no current thread.

11 years agoFix makefile's IMAIL rule: Edwin must be loaded first.
Taylor R Campbell [Mon, 1 Jul 2013 14:23:47 +0000 (14:23 +0000)]
Fix makefile's IMAIL rule: Edwin must be loaded first.

11 years agoApply patch updating list of windows OS versions.
Chris Hanson [Sun, 23 Jun 2013 07:14:30 +0000 (00:14 -0700)]
Apply patch updating list of windows OS versions.

11 years agoFix MIT Scheme's web site URI in the user manual.
Taylor R Campbell [Thu, 20 Jun 2013 23:42:40 +0000 (23:42 +0000)]
Fix MIT Scheme's web site URI in the user manual.

From Richard M. Loveland.

11 years agoMerge branch 'lazy-floenv'
Taylor R Campbell [Thu, 20 Jun 2013 15:46:19 +0000 (15:46 +0000)]
Merge branch 'lazy-floenv'

11 years agoTweak floenv tests so they preserve the floating-point environment.
Taylor R Campbell [Thu, 20 Jun 2013 15:45:51 +0000 (15:45 +0000)]
Tweak floenv tests so they preserve the floating-point environment.

11 years agoImplement lazy switching of thread floating-point environments.
Taylor R Campbell [Wed, 29 May 2013 03:10:52 +0000 (03:10 +0000)]
Implement lazy switching of thread floating-point environments.

Should reduce the overhead of thread switching and avoid
platform-dependent objects in threads that don't mess with the
floating-point environment so that they can be dumped in bands.

11 years agoUse WITH-LIMITED-INTERRUPTS, not WITH-INTERRUPT-MASK, for IMAP update.
Taylor R Campbell [Thu, 20 Jun 2013 15:32:16 +0000 (15:32 +0000)]
Use WITH-LIMITED-INTERRUPTS, not WITH-INTERRUPT-MASK, for IMAP update.

11 years agoPass RTLD_NOW, not RTLD_LAZY, to dlopen.
Taylor R Campbell [Thu, 20 Jun 2013 15:18:19 +0000 (15:18 +0000)]
Pass RTLD_NOW, not RTLD_LAZY, to dlopen.

Let dlopen fail recoverably if there are missing symbols, rather than
aborting the hapless process when it tries to use them later.

11 years agoPut a space after the colon when reporting an undefined command.
Taylor R Campbell [Wed, 12 Jun 2013 20:09:02 +0000 (20:09 +0000)]
Put a space after the colon when reporting an undefined command.

11 years agoFix typo in i386/rulflo.scm signed zero test.
Taylor R Campbell [Tue, 11 Jun 2013 16:17:32 +0000 (16:17 +0000)]
Fix typo in i386/rulflo.scm signed zero test.

Missed these typos because although I checked for (constant 0.) and
(constant -0.) in the RTL, I didn't check whether they were being
immediately passed to OBJECT->FLOAT...

11 years agoThe gdbm error handler takes a const char *, not a char *.
Taylor R Campbell [Tue, 4 Jun 2013 20:52:23 +0000 (20:52 +0000)]
The gdbm error handler takes a const char *, not a char *.

11 years agoUpdate copyrights to 2013.
Chris Hanson [Sun, 9 Jun 2013 08:05:26 +0000 (01:05 -0700)]
Update copyrights to 2013.

11 years agoDon't allow AC_CHECK_DECL([__x86_64__]) to override user-specified cpu spec.
Chris Hanson [Sun, 9 Jun 2013 07:25:33 +0000 (00:25 -0700)]
Don't allow AC_CHECK_DECL([__x86_64__]) to override user-specified cpu spec.

11 years agoSome tweaks while debugging a problem.
Chris Hanson [Sun, 9 Jun 2013 07:24:05 +0000 (00:24 -0700)]
Some tweaks while debugging a problem.

11 years agoFix typo.
Chris Hanson [Sun, 9 Jun 2013 00:37:36 +0000 (17:37 -0700)]
Fix typo.

11 years agoSyntax and compile win32 separately in the top-level makefile.
Taylor R Campbell [Tue, 4 Jun 2013 17:19:21 +0000 (17:19 +0000)]
Syntax and compile win32 separately in the top-level makefile.

11 years agoMake NUMBER->STRING give negative flonum zero a minus sign.
Taylor R Campbell [Tue, 4 Jun 2013 17:18:36 +0000 (17:18 +0000)]
Make NUMBER->STRING give negative flonum zero a minus sign.

11 years agoMake EQV? distinguish signed flonum zeros.
Taylor R Campbell [Tue, 4 Jun 2013 17:18:11 +0000 (17:18 +0000)]
Make EQV? distinguish signed flonum zeros.

11 years agoFix i386 and amd64 lapgen rules to load negative zero correctly.
Taylor R Campbell [Tue, 4 Jun 2013 17:17:28 +0000 (17:17 +0000)]
Fix i386 and amd64 lapgen rules to load negative zero correctly.

11 years agoFix floating-point underflow tests with a suggestion from Bill Kahan.
Taylor R Campbell [Tue, 4 Jun 2013 08:50:24 +0000 (08:50 +0000)]
Fix floating-point underflow tests with a suggestion from Bill Kahan.

11 years agoDisregard ERANGE from libm log when x = 0.
Taylor R Campbell [Tue, 4 Jun 2013 08:47:43 +0000 (08:47 +0000)]
Disregard ERANGE from libm log when x = 0.

This enables log to usefully raise the IEEE 754 divide-by-zero
exception when the divide-by-zero exception is untrapped.

11 years agoAdd some tests for floating-point comparison screw cases.
Taylor R Campbell [Tue, 4 Jun 2013 05:29:03 +0000 (05:29 +0000)]
Add some tests for floating-point comparison screw cases.

11 years agoAdd unordered branches for NaN cases to amd64 comparison trampolines.
Taylor R Campbell [Tue, 4 Jun 2013 05:27:11 +0000 (05:27 +0000)]
Add unordered branches for NaN cases to amd64 comparison trampolines.

11 years agoFix various infinity and NaN screw cases for <, MIN, MAX, *, and /.
Taylor R Campbell [Tue, 4 Jun 2013 05:22:42 +0000 (05:22 +0000)]
Fix various infinity and NaN screw cases for <, MIN, MAX, *, and /.

Behaviour is as prescribed by IEEE 754.  Most of the changes that
involve FLO:NAN? use it only in mixed-exactness branches, which are
not particularly performance-critical anyway.  Eventually we ought to
make FLO:NAN? and FLO:FINITE? open-codable primitives, and add
FLO:INFINITE?, FLO:NORMAL?, and FLO:SUBNORMAL? likewise.

11 years agoOn Mac OS X, use mmap(MAP_FIXED) only for i386.
Taylor R Campbell [Mon, 3 Jun 2013 19:17:41 +0000 (19:17 +0000)]
On Mac OS X, use mmap(MAP_FIXED) only for i386.

11 years agoRework heap mmapping once more.
Taylor R Campbell [Mon, 3 Jun 2013 17:53:59 +0000 (17:53 +0000)]
Rework heap mmapping once more.

- Ensure that if we pass mmap a nonzero address, we are requesting
  MAP_FIXED or MAP_TRYFIXED.

- Try grovelling through /proc/self/maps on NetBSD too -- it might be
  there.

11 years agoFix thinko in BITS.
Taylor R Campbell [Sun, 2 Jun 2013 22:31:19 +0000 (22:31 +0000)]
Fix thinko in BITS.

11 years agoFix for fix for bug #36887: return '() from IGNORABLE parser.
Matt Birkholz [Mon, 3 Jun 2013 16:40:57 +0000 (09:40 -0700)]
Fix for fix for bug #36887: return '() from IGNORABLE parser.

The tail expr was accidentally deleted in the previous commit.

11 years agoFix for bug #36887: Ignore a free variable and sf fails.
Matt Birkholz [Mon, 3 Jun 2013 15:33:57 +0000 (08:33 -0700)]
Fix for bug #36887: Ignore a free variable and sf fails.

Ignoring a free variable now just causes a warning (and names the variable).

11 years agoDon't setsid Scheme.
Taylor R Campbell [Sun, 2 Jun 2013 15:33:41 +0000 (15:33 +0000)]
Don't setsid Scheme.

If you want to detach from the controlling terminal, use a trivial
wrapper program that calls setsid and then exec (e.g., pgrphack from
daemontools), or add a primitive to Scheme to daemonize without race
conditions.

Hitting ^C at the terminal now works to interrupt Scheme in batch
mode, e.g. when run under make during the build.

11 years agocompiler: Don't let continuation-analysis add a cycle...
Matt Birkholz [Sun, 2 Jun 2013 01:53:24 +0000 (18:53 -0700)]
compiler: Don't let continuation-analysis add a cycle...

...to the block tree, which it would do (causing a "maximum recursion
depth exceeded" abort during block->dbg-block) when compiling crazy
code like:

      (define (fubar param)
(define (closure) param)
(define (loop)
  (closure)
  (loop)
  ;; It doesn't break without this
  (tail-call)))
      (fubar sumpn)

This patch checks that the "Acceptable substitute" for a block's
stack-link is not inside the block.

11 years agoMissed a dependency: edwin needs win32 syntaxed first.
Taylor R Campbell [Fri, 31 May 2013 03:53:55 +0000 (03:53 +0000)]
Missed a dependency: edwin needs win32 syntaxed first.

11 years agoRevert tterm.c changes.
Matt Birkholz [Wed, 29 May 2013 22:56:48 +0000 (15:56 -0700)]
Revert tterm.c changes.

tterm.c is not used when not USE_TERMCAP, so these old changes are
unnecessary... and potentially confusing.

11 years agoAdd configure option `--without-termcap'.
Matt Birkholz [Wed, 29 May 2013 22:50:34 +0000 (15:50 -0700)]
Add configure option `--without-termcap'.

Punt the termcap primitives (all of microcode/tterm.c!) if configured
--with-termpcap=no.  The console port ops like clear and x-size were
already equipped with useful(?) defaults.  In fact Edwin (on X11)
still works!

11 years agoMerge branch 'make-jobs'
Taylor R Campbell [Wed, 29 May 2013 14:25:00 +0000 (14:25 +0000)]
Merge branch 'make-jobs'

11 years agoReduce loading verbiage in compile-imail and compile-xml targets.
Taylor R Campbell [Wed, 29 May 2013 14:10:05 +0000 (14:10 +0000)]
Reduce loading verbiage in compile-imail and compile-xml targets.

Their respective compile.scm scripts load everything anyway.

11 years agoSplit compile-edwin target into syntax-edwin/compile-edwin.
Taylor R Campbell [Wed, 29 May 2013 14:07:14 +0000 (14:07 +0000)]
Split compile-edwin target into syntax-edwin/compile-edwin.

11 years agoSplit compile-runtime target into syntax-runtime and compile-runtime.
Taylor R Campbell [Wed, 29 May 2013 14:06:05 +0000 (14:06 +0000)]
Split compile-runtime target into syntax-runtime and compile-runtime.

Dependents of runtime-*.pkd for cref need only syntax-runtime.
Increases parallelism.

11 years agoGenerate an intermediate lib/compiler.com.
Taylor R Campbell [Wed, 29 May 2013 14:04:37 +0000 (14:04 +0000)]
Generate an intermediate lib/compiler.com.

We stopped installing this ages ago and hence stopped building it,
but now that each subdirectory is built in a separate process, it's
useful to avoid reloading the compiler repeatedly during the build.

11 years agoFirst stab at making top-level makefile jobs-safe for native builds.
Taylor R Campbell [Wed, 29 May 2013 13:45:38 +0000 (13:45 +0000)]
First stab at making top-level makefile jobs-safe for native builds.

This enables coarse-grained parallelization of the build.  There's a
nontrivial amount of crap encoded in the makefile about how to do the
build, but that's better than the nontrivial amount of dependency
crap implied by the total ordering in etc/compile.scm, most of which
can go away once the liarc and svm builds are converted similarly.

This does not restructure the build; we still have various bootstrap
and staging botches going on which we need to get rid of.  Fixing
those nicely will require some changes to liar and sf so that we can
compile a fresh compiler into a subdirectory, say `boot', using the
host compiler, and then use that to compile the system.  Liarc
approximates that with the `boot-root' installation prefix, but that
kludge breaks incremental builds which means we never, ever try to
touch the liarc build because the turnaround time is too painful.

11 years agodoc: open-tcp-stream-socket
Matt Birkholz [Tue, 28 May 2013 18:17:41 +0000 (11:17 -0700)]
doc: open-tcp-stream-socket

This procedure no longer has buffer-size and line-translation parameters.
Thanks to nick.f.russell@gmail.com for noticing this.

11 years agodoc: accommodate texinfo 5.1
Matt Birkholz [Tue, 28 May 2013 18:03:43 +0000 (11:03 -0700)]
doc: accommodate texinfo 5.1

Thanks to Charles Moss <czm>.

11 years agoTemporarily: inform older hosts about new type code SEQUENCE.
Matt Birkholz [Tue, 28 May 2013 02:07:19 +0000 (19:07 -0700)]
Temporarily: inform older hosts about new type code SEQUENCE.

Like d4d8ea1 this patch allows the released 9.1.1 (LIAR/C
distribution) to compile the compiler.  Without it, make-liarc-dist.sh
fails.  It seems the hack in d4d8ea1 is saved in a native
x-runtime.com but not in a LIAR/C-based x-runtime.com. ?

11 years agoDon't add empty arguments to -I and -L if pg_config fails.
Taylor R Campbell [Mon, 27 May 2013 21:44:26 +0000 (21:44 +0000)]
Don't add empty arguments to -I and -L if pg_config fails.

11 years agoruntime: Make it nicer to (ge '(R3RS)).
Matt Birkholz [Sat, 27 Apr 2013 21:16:13 +0000 (14:16 -0700)]
runtime: Make it nicer to (ge '(R3RS)).

The default parser and unparser expect to find bindings for
e.g. *parser-table* in any given environment, but there are no such
bindings in a package with parent #f (unless you import them).  If you
don't, executing (ge '(R3RS)) puts the REPL into a tight little error
loop.  This patch makes the parser and unparser consult the (USER)
package if they don't find these bindings in the current environment.

Once "in" a package like (r3rs) it is tricky to get out(!).  Entering
(ge '(USER)) will just produce an error: unbound variable GE.  Thus
the default/repl-eval now looks for an unquoted expression and
evaluates it in the (USER) package, so you can enter ,(ge '(USER)) to
get back.

11 years agocref: one more place must allow #f for a package's parent.
Matt Birkholz [Wed, 24 Apr 2013 16:52:13 +0000 (09:52 -0700)]
cref: one more place must allow #f for a package's parent.

11 years agoffi: Temporay hack is no longer needed.
Matt Birkholz [Wed, 24 Apr 2013 00:49:31 +0000 (17:49 -0700)]
ffi: Temporay hack is no longer needed.

11 years agocref: Support (parent #f) packages. Punt system-global-package.
Matt Birkholz [Wed, 24 Apr 2013 00:48:31 +0000 (17:48 -0700)]
cref: Support (parent #f) packages.  Punt system-global-package.

The package-structure<? sort did not ensure that BOTH a package's
parent-in-name and parent-environment were created first.  An easier
sort ensures only that the parent-environment is created first.  The
parent-in-name was only needed to hang the "child" on a tree.  This
patch replaces the tree with a list: *packages*.  Thus a
parent-in-name is not needed and the only parent/child tree is the
environment tree.

To catch out old code searching for packages via package/children (as
in swank.scm) the system-global-package binding was removed, as was
package/child.  Package/add-child! seems popular and so is supported
(with only minor trouble, i.e. no APPENDing during the cold load).
Package creation by other means is... not supported.

11 years agoAdd (unused) attribute to Rdl.
Matt Birkholz [Fri, 19 Apr 2013 19:26:54 +0000 (12:26 -0700)]
Add (unused) attribute to Rdl.

This eliminates LAIR/C build log noise to the tune of 70,000 lines!

11 years agoTypo: thread-queue/dequeue-until!.
Matt Birkholz [Wed, 17 Apr 2013 16:07:28 +0000 (09:07 -0700)]
Typo: thread-queue/dequeue-until!.

11 years agoRevert this commit when the release can build without it.
Matt Birkholz [Wed, 17 Apr 2013 16:04:28 +0000 (09:04 -0700)]
Revert this commit when the release can build without it.

These last vestiges of TC_COMBINATION_1, TC_COMBINATION_2, TC_PCOMB0,
TC_PCOMB1, and TC_PCOMB2 remain only to allow building directly from
release 9.1.

11 years agoRemove all trace of type codes COMBINATION-2, PCOMB0, etc.
Matt Birkholz [Wed, 17 Apr 2013 15:42:50 +0000 (08:42 -0700)]
Remove all trace of type codes COMBINATION-2, PCOMB0, etc.

This finishes what 3fc580e started.  The resulting system runs but
does not build with the released version 9.1.  That version puts old
combinations in .exts that cannot be fasloaded by this new version.
As the old combination types are no longer compilable nor even
interpretable, it seems the build process loads them but does not use
them.

The following commit will put just the gc handlers back, which is
sufficient to allow building directly from 9.1.

11 years agoFix argument order in HASH-TABLE/MODIFY! documentation.
Taylor R Campbell [Tue, 9 Apr 2013 02:46:16 +0000 (02:46 +0000)]
Fix argument order in HASH-TABLE/MODIFY! documentation.

11 years agoFix tabification in (runtime hash-table) package description.
Taylor R Campbell [Tue, 9 Apr 2013 02:45:36 +0000 (02:45 +0000)]
Fix tabification in (runtime hash-table) package description.

11 years agoMerge Alexey's hash table documentation and tweaks from ages ago.
Taylor R Campbell [Tue, 9 Apr 2013 02:42:12 +0000 (02:42 +0000)]
Merge Alexey's hash table documentation and tweaks from ages ago.

There was some little issue but I forget what it was.

11 years agoWarn that weak references are scary and hard to get right.
Alexey Radul [Fri, 27 May 2011 21:04:45 +0000 (22:04 +0100)]
Warn that weak references are scary and hard to get right.

Redirect to provided higher-level abstractions.

11 years agoUpdate hash-table/count for multiple entry types.
Alexey Radul [Sun, 29 May 2011 22:22:41 +0000 (23:22 +0100)]
Update hash-table/count for multiple entry types.

11 years agoRecord the reason for the restrictions in hash-table/modify!
Alexey Radul [Sun, 29 May 2011 22:06:42 +0000 (23:06 +0100)]
Record the reason for the restrictions in hash-table/modify!

I found this non-obvious when I was writing the documentation, so it
seems appropriate to record it somewhere.  On the other hand, it does
not seem appropriate to burden the user with having to read it ---
just obey the restriction and you'll be fine.  Will anyone ever read
comments in the source of the manual?  Is there a better place to
record this knowledge?

11 years agoAdd some concept index entries that seem reasonable.
Alexey Radul [Sun, 29 May 2011 21:56:34 +0000 (22:56 +0100)]
Add some concept index entries that seem reasonable.

11 years agoUpdate the documentation of hash-table/clean!
Alexey Radul [Sun, 29 May 2011 21:48:26 +0000 (22:48 +0100)]
Update the documentation of hash-table/clean!

11 years agoDeprecate {strong/weak}-hash-table/constructor.
Alexey Radul [Sun, 29 May 2011 21:47:49 +0000 (22:47 +0100)]
Deprecate {strong/weak}-hash-table/constructor.

Flush the detailed description of what they do as having been copied
into the description of HASH-TABLE/CONSTRUCTOR.

11 years agoRewrite the hash table constructors section with more weakness options.
Alexey Radul [Sun, 29 May 2011 21:37:59 +0000 (22:37 +0100)]
Rewrite the hash table constructors section with more weakness options.

Define and export key-ephemeral-eq and key-ephemeral-eqv tables as
replacements for the key-weak versions.

Punt descriptions of old hash table constructor procedures to the
bottom of the section and say they are for backward compatibility.

One may object to MAKE-KEY-EPHEMERAL-EQ-HASH-TABLE on the grounds that
it invites a combinatorial explosion of names:
make-key/datum-weak-eqv-hash-table
make-datum-ephemeral-string-hash-table
make-key&datum-ephemeral-equal-hash-table (!?)
and so on ad nauseam.

Where will it end?  The criterion I used to decide which names to
export and document and which names to leave alone and defer to the
general HASH-TABLE/CONSTRUCTOR mechanism was simply to update the
existing documentation.  The manual already listed
MAKE-WEAK-EQ-HASH-TABLE.  It is now named MAKE-KEY-WEAK-EQ-HASH-TABLE,
so that name is included.  But really, a key-weak table is just a
performance optimization over a key-ephemeral table, to save work when
you know the data will not hold the keys strongly.  So
MAKE-KEY-EPHEMERAL-EQ-HASH-TABLE is in; it was, in fact, the reason I
wanted Taylor to implement ephemerons in the first place.
MAKE-KEY-EPHEMERAL-EQV-HASH-TABLE is in to preserve the symmetry
between eq? and eqv? that was already present in the manual.  But the
rest of them are out, because they weren't there before.  If
datum-weak tables were so important that their constructor really must
be given a name here, then (arguably) why were they not already
implemented and documented?  The fact that MIT Scheme only supported
strong and key-weak tables for a long time suggests that those kinds
are the most common, and therefore the most deserving of slots in the
name space.

11 years agoRelocate and update the examples of defining hash table contructors.
Alexey Radul [Sun, 29 May 2011 20:48:27 +0000 (21:48 +0100)]
Relocate and update the examples of defining hash table contructors.

Now in terms of HASH-TABLE/CONSTRUCTOR and entry types.

11 years agoDraft documentation of HASH-TABLE/CONSTRUCTOR and all the entry types.
Alexey Radul [Fri, 27 May 2011 22:56:42 +0000 (23:56 +0100)]
Draft documentation of HASH-TABLE/CONSTRUCTOR and all the entry types.

Will adjust the surrounding text to fit it in in a separate commit.

11 years agoRedirect cross reference to Weak References
Alexey Radul [Fri, 27 May 2011 20:42:38 +0000 (21:42 +0100)]
Redirect cross reference to Weak References

on the grounds that that is now the landing page for the discussion of
strength and weakness.

11 years agoImprove documentation of hash-table/{modify,intern}!
Alexey Radul [Fri, 27 May 2011 08:31:28 +0000 (09:31 +0100)]
Improve documentation of hash-table/{modify,intern}!

By the mechanism of implementing Taylor's suggestions.  Also add
sentences about average running times.

11 years agoDeparenthesize.
Alexey Radul [Thu, 26 May 2011 11:05:25 +0000 (12:05 +0100)]
Deparenthesize.

11 years agoCollect "Weak Pairs" and "Ephemerons" into a section called "Weak References".
Alexey Radul [Thu, 26 May 2011 11:03:25 +0000 (12:03 +0100)]
Collect "Weak Pairs" and "Ephemerons" into a section called "Weak References".

State the common motivation, define the common terminology, and
describe the contrasts between weak pairs and ephemerons in the common
section, and tighten the "Weak Pairs" and "Ephemerons" sections
themselves back to just describing their data structures.

11 years agoRewrite the description of intricacies with ephemeron-broken? per
Alexey Radul [Thu, 26 May 2011 08:48:10 +0000 (09:48 +0100)]
Rewrite the description of intricacies with ephemeron-broken? per
Taylor's recommendation to name the ephemeron explicitly and use more
active language.

11 years agoDocumenting hash-table/modify! and hash-table/intern!. Also slightly
Alexey Radul [Tue, 24 May 2011 22:39:06 +0000 (23:39 +0100)]
Documenting hash-table/modify! and hash-table/intern!.  Also slightly
reworded the introduction to hash-table/lookup, to make it more
positive.

11 years agoDocumenting the fact that MIT Scheme implements SRFI 69, the
Alexey Radul [Tue, 24 May 2011 12:13:07 +0000 (13:13 +0100)]
Documenting the fact that MIT Scheme implements SRFI 69, the
relationship between the MIT Scheme native hash table API and the one
specified by SRFI 69, and the reason why it is appropriate to provide
both (it's not pure backwards compatibility).  This includes
documenting the bugs in SRFI 69.

11 years agoGrammar.
Alexey Radul [Tue, 24 May 2011 09:39:49 +0000 (10:39 +0100)]
Grammar.

11 years agoPointing out what ephemerons are for; emphasizing the special feature
Alexey Radul [Mon, 23 May 2011 11:27:14 +0000 (12:27 +0100)]
Pointing out what ephemerons are for; emphasizing the special feature
of ephemerons, and the distinction between ephemerons and weak pairs.

11 years agoClarifying that ephemeron-broken? indicates that an ephemeron is whole by returning #f.
Alexey Radul [Mon, 23 May 2011 11:24:59 +0000 (12:24 +0100)]
Clarifying that ephemeron-broken? indicates that an ephemeron is whole by returning #f.