projects/cmucl/cmucl.git
27 hours agoUpdate to ASDF 3.0.1. master
Raymond Toy [Fri, 24 May 2013 02:36:45 +0000] 
Update to ASDF 3.0.1.

4 days agoRegenerated from current sources.
Raymond Toy [Tue, 21 May 2013 05:15:29 +0000] 
Regenerated from current sources.

4 days agoFix ticket:82: If the form numbers are too large to be encoded, don't
Raymond Toy [Tue, 21 May 2013 02:24:53 +0000] 
Fix ticket:82:  If the form numbers are too large to be encoded, don't
try to set the source location information.  We try to print a warning
in such cases.

4 days agoDon't reverse surrogate pairs in strings; it's not allowed by the
Raymond Toy [Tue, 21 May 2013 02:08:31 +0000] 
Don't reverse surrogate pairs in strings; it's not allowed by the
CLHS.

6 days agoUpdate.
Raymond Toy [Sun, 19 May 2013 00:47:39 +0000] 
Update.

6 days agoFix ticket:81 and fix ticket:83.
Raymond Toy [Sun, 19 May 2013 00:44:02 +0000] 
Fix ticket:81 and fix ticket:83.

From ticket 81, the tests are now:

{{{
(time (prog1 t (time-rev *s*)))
; Evaluation took:
;   0.49 seconds of real time
;   0.481813 seconds of user run time
;   0.003624 seconds of system run time
;   1,490,776,936 CPU cycles
;   [Run times include 0.13 seconds GC run time]
;   0 page faults and
;   200,073,704 bytes consed.

(time (prog1 t (time-rev *s2*)))
; Evaluation took:
;   0.97 seconds of real time
;   0.965893 seconds of user run time
;   0.005139 seconds of system run time
;   2,980,415,911 CPU cycles
;   [Run times include 0.23 seconds GC run time]
;   0 page faults and
;   400,005,560 bytes consed.
}}}

So the new string-reverse* is 20 times faster for strings without
surrogates and 10 times faster for strings containing only surrogates.

8 days agoForgot to commit changes to code/x86-vm.lisp to:
Raymond Toy [Fri, 17 May 2013 02:17:56 +0000] 
Forgot to commit changes to code/x86-vm.lisp to:

Wrap exports in eval-when for x86 as was done for sparc and add
CHAR-BYTES to x86-x86 cross-compile script.

10 days agoWrap exports in eval-when for x86 as was done for sparc and add
Raymond Toy [Wed, 15 May 2013 03:30:16 +0000] 
Wrap exports in eval-when for x86 as was done for sparc and add
CHAR-BYTES to x86-x86 cross-compile script.

Only the change for CHAR-BYTES is needed for cross-compiling from x86
to x86, but I think the other changes will be needed if you do a
cross-compile from some other arch to x86.

10 days agoWrap exports in eval-when for ppc as was done for sparc and add
Raymond Toy [Wed, 15 May 2013 03:20:44 +0000] 
Wrap exports in eval-when for ppc as was done for sparc and add
CHAR-BYTES to ppc cross-compile script.

This is untested.

10 days agoWrap exports in eval-when.
Raymond Toy [Wed, 15 May 2013 03:02:28 +0000] 
Wrap exports in eval-when.

 src/compiler/generic/objdef.lisp::
 src/compiler/generic/utils.lisp:;
 src/compiler/generic/vm-macs.lisp::
 src/compiler/sparc/c-callback.lisp::
 src/compiler/sparc/parms.lisp::

  When export lost its compile-time effects and became a normal
  function, the exports in many files no longer took affect while
  compiling the file.  This change makes the compile-time effects
  happen as before in selected files. With this change, the sparc port
  can be cross-compiled from x86 again.

 src/tools/cross-scripts/cross-x86-sparc.lisp::
  Need to frob CHAR-BYTES, which is needed by BYTE-BASH-COPY. .

2 weeks agoClean up CC snapshot-2013-05
Raymond Toy [Sun, 5 May 2013 04:37:33 +0000] 
Clean up CC

 * GNUmakefile
 * Config.solaris_sunc
   * Remove assignment of CC, letting it default.
 * Config.solaris
   * Force CC = gcc in case cc is Sun C

2 weeks agoMake the C stack not executable
Raymond Toy [Sun, 5 May 2013 04:32:35 +0000] 
Make the C stack not executable

 * Config.sparc_common
   * OS_LINK_FLAGS is not common, so remove it.
 * Config.sparc_gcc
 * Config.sparc_sunc
   * Set OS_LINK_FLAGS appropriately.

2 weeks agoRemove DEPENDS and set DEPENDS_FLAG for ppc.
Raymond Toy [Sun, 5 May 2013 04:16:44 +0000] 
Remove DEPENDS and set DEPENDS_FLAG for ppc.

2 weeks agoClean up DEPENDS
Raymond Toy [Sun, 5 May 2013 04:15:20 +0000] 
Clean up DEPENDS

 * GNUmakefile
   * Replace $(DEPENDS) with $(CC)
 * Config.x86_common
   * Set DEPENDS_FLAG to -MM

2 weeks agoSupport older versions of git that don't support describe --dirty.
Raymond Toy [Sun, 5 May 2013 04:03:34 +0000] 
Support older versions of git that don't support describe --dirty.
Just use plain describe, which is close enough.

2 weeks agoAdd -z noexecstack to disable exec flag for the C stack.
Raymond Toy [Sat, 4 May 2013 21:13:46 +0000] 
Add -z noexecstack to disable exec flag for the C stack.

2 weeks agoDon't need to set CC or CPP. But add a note that we implicitly assume
Raymond Toy [Sat, 4 May 2013 21:11:25 +0000] 
Don't need to set CC or CPP.  But add a note that we implicitly assume
cc is Sun's C compiler, not an alias to gcc.

2 weeks agoMake stacks not executable.
Raymond Toy [Sat, 4 May 2013 21:08:17 +0000] 
Make stacks not executable.

2 weeks agoFilter out the Config.*common files from the list of lisp variants.
Raymond Toy [Sat, 4 May 2013 19:55:14 +0000] 
Filter out the Config.*common files from the list of lisp variants.

2 weeks agoRemove old Config files that are no longer used or supported.
Raymond Toy [Sat, 4 May 2013 19:51:34 +0000] 
Remove old Config files that are no longer used or supported.

2 weeks agoRemove the -M option that was recently added.
Raymond Toy [Sat, 4 May 2013 19:45:30 +0000] 
Remove the -M option that was recently added.

This isn't needed since we no longer set CC in the Config files.

 bin/build.sh::
 * Revert change that added -M

 bin/rebuild-lisp.sh::
 * Remove the -M option, but leave in the getops and usage changes so
   we get a little command help if needed.

2 weeks agoDon't set CC. The default is good enough and we'll use either gcc on
Raymond Toy [Sat, 4 May 2013 19:38:30 +0000] 
Don't set CC.  The default is good enough and we'll use either gcc on
older systems or clang on newer ones.

2 weeks agoDon't set CC; let it default or let the individual Config files set up
Raymond Toy [Sat, 4 May 2013 19:37:36 +0000] 
Don't set CC; let it default or let the individual Config files set up
CC if necessary.

2 weeks agoFix compiler warning on sparc about 1st arg to madvise.
Raymond Toy [Sat, 4 May 2013 19:36:31 +0000] 
Fix compiler warning on sparc about 1st arg to madvise.

2 weeks agoClean up variables.
Raymond Toy [Sat, 4 May 2013 19:35:32 +0000] 
Clean up variables.

 Config.sparc_common::
 * Use += instead of = for CPPFLAGS and CFLAGS

 Config.sparc_gcc:
 * Move options from CC to CFLAGS.

 Config.sparc_gcc:
 * Move options from CC to CFLAGS.
 * Add DEPEND_FLAGS with appropriate options.

3 weeks agoUse clang instead of gcc on Mac OSX.
Raymond Toy [Fri, 3 May 2013 04:46:02 +0000] 
Use clang instead of gcc on Mac OSX.

3 weeks agoAdd -M option to pass args to make.
Raymond Toy [Fri, 3 May 2013 04:43:09 +0000] 
Add -M option to pass args to make.

3 weeks agoAdd -M option to pass extra flags to make.
Raymond Toy [Fri, 3 May 2013 04:42:51 +0000] 
Add -M option to pass extra flags to make.

Use getopts to parse the options and also print slightly better error
messages.

3 weeks agoInclude interr.h (and os.h) to get rid of a compiler warning about
Raymond Toy [Fri, 3 May 2013 03:15:57 +0000] 
Include interr.h (and os.h) to get rid of a compiler warning about
utf16_output being undeclared.

3 weeks agoJust remove a blank line and reformat a strangely formatted line.
Raymond Toy [Fri, 3 May 2013 03:15:16 +0000] 
Just remove a blank line and reformat a strangely formatted line.

3 weeks agoFix some issues with zero check in gc_free_heap.
Raymond Toy [Fri, 3 May 2013 03:09:04 +0000] 
Fix some issues with zero check in gc_free_heap.

In gc_free_heap, the check that unallocated pages are zero can only be
done if gencgc_unmap_zero is MODE_MAP or MODE_MEMSET.  The other modes
are known not to zero memory.

Also, don't limit the check to the first 16K pages, and also check the
entire page of GC_PAGE_SIZE bytes, not just the first 1024 words.

3 weeks agoClean up madvise implementation
Raymond Toy [Sat, 27 Apr 2013 16:08:11 +0000] 
Clean up madvise implementation

 * Remove PAGE_MADVISE_MASK
 * Remove code using PAGE_MADVISE_MASK
 * Move some #defines from gencgc.c to gencgc.h
 * Add new or better comments

4 weeks agoTurn on zero checking if gencgc_unmap_zero is lazy.
Raymond Toy [Fri, 26 Apr 2013 04:04:16 +0000] 
Turn on zero checking if gencgc_unmap_zero is lazy.

This should only be temporary so we can get some testing with lazy
mode.

4 weeks ago * Change default to MODE_LAZY
Raymond Toy [Wed, 24 Apr 2013 04:14:40 +0000] 
 * Change default to MODE_LAZY
 * Disable gencgc_debug_madvise.

4 weeks agoClean up implementation of new gencgc_unmap_zero.
Raymond Toy [Sun, 21 Apr 2013 21:34:40 +0000] 
Clean up implementation of new gencgc_unmap_zero.

 * Add MODE_LAZY
 * Rename PAGE_MADVISE flag to PAGE_MADVISE_MASK
 * Remove some #ifdef'ed out code.
 * Change default gencgc_unmap_zero to MODE_MEMSET for all OS/archs.

4 weeks ago * Add support for linux and solaris/sparc. Linux isn't working with
Raymond Toy [Sun, 21 Apr 2013 05:39:48 +0000] 
 * Add support for linux and solaris/sparc. Linux isn't working with
   madvise, though.  I get an error about no transport function for
   some object.

4 weeks ago * Add variable to control debug prints for madvise
Raymond Toy [Sat, 20 Apr 2013 19:07:33 +0000] 
 * Add variable to control debug prints for madvise
 * Actually enable call to madvise instead of doing memset.

4 weeks agoFirst cut at using madvise instead of memset or mmap/munmap to zero
Raymond Toy [Sat, 20 Apr 2013 18:03:34 +0000] 
First cut at using madvise instead of memset or mmap/munmap to zero
out freed pages.

5 weeks agoOops. Always build asdf and friends. snapshot-2013-04
Raymond Toy [Sun, 14 Apr 2013 05:12:36 +0000] 
Oops. Always build asdf and friends.

5 weeks agoUpdate from logs.
Raymond Toy [Sun, 14 Apr 2013 03:16:32 +0000] 
Update from logs.

5 weeks agoUpdate from logs.
Raymond Toy [Sat, 13 Apr 2013 21:57:27 +0000] 
Update from logs.

5 weeks agoUpdate -u option so that asdf is always built.
Raymond Toy [Sat, 13 Apr 2013 21:54:15 +0000] 
Update -u option so that asdf is always built.

 bin/build-utils.sh::
 * asdf is not part of build-utils.sh

 bin/build.sh::
 * Build asdf as part of the normal build because it's always part of
   the main tarball, not the extras tarball.

6 weeks agoUpdate from logs.
Raymond Toy [Mon, 8 Apr 2013 15:28:20 +0000] 
Update from logs.

6 weeks agoFix ticket:80
Raymond Toy [Sun, 7 Apr 2013 18:03:06 +0000] 
Fix ticket:80

 * Increase the size of *powers-of-ten* a bit.
 * In expt-ten, handle the case where the exponent exceeds the size of
   the *powers-of-ten* array.

2 months agoFix ticket:77 correctly, using the supplied patch link.
Raymond Toy [Mon, 25 Mar 2013 03:13:56 +0000] 
Fix ticket:77 correctly, using the supplied patch link.

2 months agoFix ticket:77 by adding the code given in the ticket.
Raymond Toy [Sun, 24 Mar 2013 16:43:33 +0000] 
Fix ticket:77 by adding the code given in the ticket.

2 months agoUpdate from logs.
Raymond Toy [Sat, 23 Mar 2013 17:38:29 +0000] 
Update from logs.

2 months agoFix ticket:79
Raymond Toy [Sat, 23 Mar 2013 17:27:43 +0000] 
Fix ticket:79

 * Initialize in-length to in-buffer-length, not 0.
 * Added a few more debugging prints.

2 months agoTry to be careful about extracting the linux version from the (uname)
Raymond Toy [Sat, 23 Mar 2013 03:10:16 +0000] 
Try to be careful about extracting the linux version from the (uname)
release.  Some Debian versions have a release name like "3.7-trunk",
which is missing the patch version.

2 months agoSupport ppc.
Raymond Toy [Sun, 10 Mar 2013 05:24:25 +0000] 
Support ppc.

2 months agoUpdate. rtoy-unicode-6.2 snapshot-2013-03-a
Raymond Toy [Thu, 7 Mar 2013 03:18:04 +0000] 
Update.

2 months agoReindent STRING-NEXT-WORD-BREAK neatly.
Raymond Toy [Wed, 6 Mar 2013 08:37:50 +0000] 
Reindent STRING-NEXT-WORD-BREAK neatly.

2 months agoMerge branch 'master' into rtoy-unicode-6.2
Raymond Toy [Wed, 6 Mar 2013 08:30:11 +0000] 
Merge branch 'master' into rtoy-unicode-6.2

2 months agoUpdate.
Raymond Toy [Wed, 6 Mar 2013 08:29:09 +0000] 
Update.

2 months agoFix ticket:76
Raymond Toy [Wed, 6 Mar 2013 08:27:44 +0000] 
Fix ticket:76

Missed one place in DO-UNARY-BYTE-BASH to adjust the call to END-MASK
to use a bit offset instead of a byte offset.  This affects anything
that was using DO-UNARY-BYTE-BASH, including REPLACE.

2 months agoImplement Rule WB13c for regional indicators.
Raymond Toy [Wed, 6 Mar 2013 08:13:22 +0000] 
Implement Rule WB13c for regional indicators.

2 months agoFix PARSE-WORD-BREAK-LINE to handle codepoints outside the BMP. The
Raymond Toy [Wed, 6 Mar 2013 06:15:43 +0000] 
Fix PARSE-WORD-BREAK-LINE to handle codepoints outside the BMP.  The
count needs to be incremented one to adjust for the UTF-16 encoding of
strings that we use.

2 months agoUpdate from logs.
Raymond Toy [Wed, 6 Mar 2013 03:52:16 +0000] 
Update from logs.

2 months agoUpdate to ASDF 2.32.
Raymond Toy [Wed, 6 Mar 2013 03:46:38 +0000] 
Update to ASDF 2.32.

2 months agoNote ticket #74 fixed, and move a Change item to a Bugfix item.
Raymond Toy [Tue, 5 Mar 2013 06:08:38 +0000] 
Note ticket #74 fixed, and move a Change item to a Bugfix item.

2 months agoOops. Remove debugging echo.
Raymond Toy [Tue, 5 Mar 2013 06:07:06 +0000] 
Oops.  Remove debugging echo.

2 months agoUpdate to Unicode 6.2.
Raymond Toy [Tue, 5 Mar 2013 05:54:28 +0000] 
Update to Unicode 6.2.

Still needs work because the word-break tests fail.

2 months agoUpdate from logs and cleanup some of the text. snapshot-2013-03
Raymond Toy [Mon, 4 Mar 2013 02:21:37 +0000] 
Update from logs and cleanup some of the text.

2 months agoIgnore some TeX output files.
Raymond Toy [Mon, 4 Mar 2013 02:20:52 +0000] 
Ignore some TeX output files.

2 months agoFix ticket:74
Raymond Toy [Wed, 27 Feb 2013 04:44:18 +0000] 
Fix ticket:74

When accounting for the octets left in the in-buffer that we haven't
read (or converted to characters), we were subtracting the index from
the total in-buffer length.  This is wrong if the file is less than
the total in-buffer length.  We should have subtracted from the actual
number of octets in the in-buffer.

2 months agoUpdate from commit logs.
Raymond Toy [Sun, 24 Feb 2013 04:44:45 +0000] 
Update from commit logs.

2 months agoSignal an error if the readtable case of the standard readtable is
Raymond Toy [Sun, 24 Feb 2013 04:44:23 +0000] 
Signal an error if the readtable case of the standard readtable is
changed.

* Rename the slot to %READTABLE-CASE (from READTABLE-CASE).
* Add READTABLE-CASE and (SETF READTABLE-CASE) functions, as required.
* Check for the standard readtable in (SETF READTABLE-CASE).

2 months agoFix ticket:73
Raymond Toy [Sun, 24 Feb 2013 02:14:18 +0000] 
Fix ticket:73

error.lisp::
* Create two new conditions, one for modifying the readtable and one
  for the pprint dispatch table.

exports.lisp::
* Export the two new conditions.

pprint.lisp::
* Add check to SET-PPRINT-DISPATCH to disallow modifying the standard
  pprint dispatch table.
* Allow PPRINT-INIT to modify the standard pprint dispatch table.

print.lisp::
* In WITH-STANDARD-IO-SYNTAX, don't copy a new dispatch table; bind
  *print-pprint-dispatch* to the standard table.

reader.lisp::
* Add check to disallow modifying the standard readtable.
* Allow INIT-STD-LISP-READTABLE to modify the standard readtable.

2 months agoUpdate asdf version.
Raymond Toy [Sat, 23 Feb 2013 16:42:13 +0000] 
Update asdf version.

2 months agoUpdate to asdf 2.30.
Raymond Toy [Sat, 23 Feb 2013 16:41:33 +0000] 
Update to asdf 2.30.

2 months agoUpdate from commit logs.
Raymond Toy [Sat, 23 Feb 2013 16:40:00 +0000] 
Update from commit logs.

2 months agoRemove old stuff.
Raymond Toy [Sat, 23 Feb 2013 16:31:57 +0000] 
Remove old stuff.

* Remove support for :athlon in fp-pop.
* Remove all of the enhanced pentium pro fp comparison code.  Use the
  non ppro everywhere.

2 months agoOops. Use the pentium version of the log1p vop and remove the other
Raymond Toy [Sat, 23 Feb 2013 16:17:19 +0000] 
Oops.  Use the pentium version of the log1p vop and remove the other
one.

2 months agoRemove :PENTIUM from *FEATURES*
Raymond Toy [Sat, 23 Feb 2013 15:38:18 +0000] 
Remove :PENTIUM from *FEATURES*

code/misc.lisp::
* Don't register :PENTIUM feature.

code/time.lisp::
* Replace :pentium with :x86; we assume all x86's have the rdtsc
  instruction.

compiler/x86/float.lisp::
compiler/x86/system.lisp::
* Remove the vop guard on :pentium; assume we're always running on a
  pentium or better.

3 months agoRemove the :i486 feature.
Raymond Toy [Fri, 22 Feb 2013 06:21:10 +0000] 
Remove the :i486 feature.

code/misc.lisp::
* Remove :i486 from *features*

code/multi-proc.lisp::
* Change :i486 to :x86 on the off chance that multi-proc will be
  ported to other archs.

compiler/x86/cell.lisp::
compiler/x86/macros.lisp::
compiler/x86/memory.lisp::
* Remove guard on backend-featurep :i486 since we got rid of :i486 and
  have it always enabled now.

3 months agoUpdate from commit logs.
Raymond Toy [Fri, 22 Feb 2013 05:46:40 +0000] 
Update from commit logs.

3 months agoClean up one- and two-arg-derive-type.
Raymond Toy [Fri, 22 Feb 2013 05:43:50 +0000] 
Clean up one- and two-arg-derive-type.

srctran.lisp::
 * The optional convert-type arg is always true, so remove it and
   update the local functions to convert always.
 * Clean up some comments as well.

float-tran.lisp::
 * Fix up one place where we were supplying a value for the optional
   convert-type arg.

3 months agoRemove :conservative-float-type from *features* and update code
Raymond Toy [Fri, 22 Feb 2013 04:28:37 +0000] 
Remove :conservative-float-type from *features* and update code
accordingly.

3 months agoRemove :COMMON from *FEATURES*
Raymond Toy [Fri, 22 Feb 2013 03:26:15 +0000] 
Remove :COMMON from *FEATURES*

 lispinit.lisp::
 * Remove :COMMON from *FEATURES*
 misc.lisp::
 * Remove unused defvar of *features*.

3 months agoRemove old fast-truncate-c vops; the new ones work.
Raymond Toy [Wed, 20 Feb 2013 04:25:04 +0000] 
Remove old fast-truncate-c vops; the new ones work.

3 months agoForgot to update this when x86/parms.lisp was updated with new
Raymond Toy [Wed, 20 Feb 2013 02:27:13 +0000] 
Forgot to update this when x86/parms.lisp was updated with new
constants.

3 months agoOptimize shifts that are known to be right shifts.
Raymond Toy [Tue, 19 Feb 2013 03:54:07 +0000] 
Optimize shifts that are known to be right shifts.

Without this, the general shift vop is used which has to test for the
sign of the shift before shifting.  This micro-optimization removes
the test when we know the sign of the shift.

3 months agoAdd some fixnum contants for x86, as was done for sparc.
Raymond Toy [Tue, 19 Feb 2013 03:27:47 +0000] 
Add some fixnum contants for x86, as was done for sparc.

 code/export.lisp::
 * Export new symbols

 compiler/x86/parms.lisp::
 * Define constants for useful the number of lowtag bits and masks and
   the number of fixnum tag bits and masks.

3 months agoOptimize TRUNCATE when the second arg is a compile-time constant.
Raymond Toy [Tue, 19 Feb 2013 00:17:24 +0000] 
Optimize TRUNCATE when the second arg is a compile-time constant.

3 months agoUpdate to asdf 2.29.
Raymond Toy [Mon, 18 Feb 2013 21:29:44 +0000] 
Update to asdf 2.29.

3 months agoUpdate.
Raymond Toy [Sat, 2 Feb 2013 18:50:45 +0000] 
Update.

3 months agoMake the version arg optional.
Raymond Toy [Sat, 2 Feb 2013 18:49:53 +0000] 
Make the version arg optional.

If the version is not specified, determine a default version based on
the git hash returned by git describe.  Thus, for tagged snapshot
builds, the version will automatically be chosen from the tag, with
"snapshot-" removed.

3 months agoUpdate to ASDF 2.28 snapshot-2013-02
Raymond Toy [Sat, 2 Feb 2013 04:33:30 +0000] 
Update to ASDF 2.28

 src/contrib/asdf/asdf.lisp::
 * Update

 src/general-info/release-20e.txt::
 * Update

3 months agoRegenerated.
Raymond Toy [Sat, 2 Feb 2013 03:21:24 +0000] 
Regenerated.

3 months agoUpdate from commit logs.
Raymond Toy [Sat, 2 Feb 2013 03:20:41 +0000] 
Update from commit logs.

3 months agoFix typo in string.
Raymond Toy [Sat, 2 Feb 2013 03:10:52 +0000] 
Fix typo in string.

3 months agoFix ticket:72 : SIGFPE with no exceptions enabled
Raymond Toy [Fri, 1 Feb 2013 05:27:36 +0000] 
Fix ticket:72 : SIGFPE with no exceptions enabled

 code/float-trap.lisp::
  * In the default case, check FOP to see if it's defined.  This means
    it's a divide exception.

 code/x86-vm.lisp::
  * Also check to see if the offending instruction is a DIV or IDIV,
    which means we got an integer overflow.  Return the appropriate
    values in this case.

3 months agoFix ticket:65 some more.
Raymond Toy [Fri, 25 Jan 2013 04:55:45 +0000] 
Fix ticket:65 some more.

Apply the expt transform in more places.  The test script in the
ticket now passes.

3 months agoFix ticket:71
Raymond Toy [Fri, 25 Jan 2013 04:28:09 +0000] 
Fix ticket:71

The reader conditionals in {{{NOT-MORE-CONTAGIOUS}}} were incorrrectly
handling the case when the second type was a double-float. It was
always returning T in that case, but this is wrong if the first type
is a double-double-float.

3 months agoUpdate from commit logs.
Raymond Toy [Thu, 24 Jan 2013 05:25:07 +0000] 
Update from commit logs.

3 months agoFix ticket:65. Implement the deftransform in the expt function.
Raymond Toy [Thu, 24 Jan 2013 05:22:24 +0000] 
Fix ticket:65.  Implement the deftransform in the expt function.

3 months agoFix ticket:70 by binding *d-p-d* to #p"".
Raymond Toy [Thu, 24 Jan 2013 04:20:18 +0000] 
Fix ticket:70 by binding *d-p-d* to #p"".

This allows logical pathnames (which always have a host to parse
against) and search-lists (which don't have an explicit pathname to
parse against) to be parsed correctly.  See the comment for some
possible issues.

4 months agoForgot to update this when alien-callback.lisp was added.
Raymond Toy [Wed, 23 Jan 2013 04:43:03 +0000] 
Forgot to update this when alien-callback.lisp was added.

4 months agoFix up the file names in the file-comment.
Raymond Toy [Wed, 23 Jan 2013 04:41:56 +0000] 
Fix up the file names in the file-comment.

4 months agoUpdate due to new docstring for rename-file.
Raymond Toy [Wed, 23 Jan 2013 03:55:56 +0000] 
Update due to new docstring for rename-file.