bpt/guile.git
22 years ago* lib.scm: Move module the system directives `export',
Mikael Djurfeldt [Sun, 21 Oct 2001 09:49:19 +0000 (09:49 +0000)]
* lib.scm: Move module the system directives `export',
  `export-syntax', `re-export' and `re-export-syntax' into the
  `define-module' form.  This is the recommended way of exporting
  bindings.

* srfi-2.scm, srfi-4.scm, srfi-8.scm, srfi-9.scm, srfi-10.scm,
srfi-11.scm, srfi-14.scm, srfi-16.scm: Move module the system
directives `export', `export-syntax', `re-export' and
`re-export-syntax' into the `define-module' form.  This is the
recommended way of exporting bindings.

* goops.scm, goops/active-slot.scm, goops/compile.scm,
goops/composite-slot.scm, goops/describe.scm, goops/dispatch.scm,
goops/old-define-method.scm, goops/save.scm, goops/util.scm: Move
module the system directives `export', `export-syntax',
`re-export' and `re-export-syntax' into the `define-module' form.
This is the recommended way of exporting bindings.

* slib.scm (array-indexes): New procedure.
(*features*): Extend.  (Probably some of these options should be
set elsewhere.)  (Thanks to Aubrey Jaffer.)

* and-let-star-compat.scm, and-let-star.scm, calling.scm,
  channel.scm, common-list.scm, debug.scm, debugger.scm,
  expect.scm, hcons.scm, lineio.scm, ls.scm, mapping.scm,
  null.scm, optargs.scm, poe.scm, popen.scm, pretty-print.scm,
  q.scm, r5rs.scm, rdelim.scm, regex.scm, runq.scm, safe-r5rs.scm,
  safe.scm, session.scm, slib.scm, streams.scm, string-fun.scm,
  syncase.scm, threads.scm: Move module the system directives
  `export', `export-syntax', `re-export' and `re-export-syntax'
  into the `define-module' form.  This is the recommended way of
  exporting bindings.

22 years ago* tests/syntax.test: Added test cases for 'cond =>' syntax with
Dirk Herrmann [Thu, 18 Oct 2001 21:59:29 +0000 (21:59 +0000)]
* tests/syntax.test:  Added test cases for 'cond =>' syntax with
else clause.  Changed some tests and comments related to the
'case' form to reflect recent changes in the implementation.

22 years ago* eval.c (scm_lookupcar, scm_m_letstar, scm_m_do, iqq,
Dirk Herrmann [Thu, 18 Oct 2001 21:38:04 +0000 (21:38 +0000)]
* eval.c (scm_lookupcar, scm_m_letstar, scm_m_do, iqq,
scm_m_define, scm_m_letrec1, scm_m_let, scm_m_expand_body,
scm_macroexp, unmemocopy, scm_eval_args, scm_deval_args,
SCM_CEVAL, scm_map, scm_init_eval):  When building lists, prefer
scm_list_<n> over scm_cons[2]?.

(scm_unmemocar, scm_m_cond, scm_m_letstar, scm_m_letrec1,
scm_m_let, scm_m_atbind, unmemocopy, SCM_CEVAL, SCM_APPLY):  Use
SCM_C[AD][AD]R instead of explicit form.

(scm_m_set_x, scm_m_cond, scm_m_letstar, scm_m_do):  Reordered
comparison parameters.

(scm_m_case, scm_m_cond, scm_m_letstar, scm_m_do, SCM_CEVAL):  Use
!SCM_NULLP instead of SCM_NIMP.

(scm_m_case):  Don't copy the form.  Renamed proc to clause and
minimized its scope.  Renamed x to clauses.  Removed side
effecting operation from macro call.

(scm_m_cond):  Don't copy the form.  Renamed arg1 to clause and
minimized its scope.  Renamed x to clauses.  Minimized the scope
of variable 'len'.  Make sure the else clause is treated specially
even in case of '=>' occurences.  Don't change the else to #t in
order to be able to distinguish this case in the evaluator.  Leave
type checking of the recipient to the evaluator.

(scm_c_improper_memq):  Made the comment somewhat clearer.

(scm_m_lambda):  Renamed proc to formals.  Removed unnecessary
test for SCM_IM_LET at the place of the formal parameters.
Simplified the formal parameter checking.

(scm_m_letstar):  Added Comment.  Renamed proc to bindings.
Renamed arg1 to binding and minimized its scope.  Eliminated
unnecessary consing.

(scm_m_do):  Renamed proc to bindings.  Minimized the scope of
variable 'len'.

(build_binding_list):  New static function.

(unmemocopy):  Don't use SCM_TYP7 on pairs (it's unclean).
Further, split up the 'letrec' unmemoizing code to the
corresponding parts for 'do', 'let' and 'letrec', adding comments
to each form.  Cleanup the handling of the do form (This removes
some *real* code :-).

(SCM_CEVAL):  Removed side effecting operation from macro call.
Handle the 'else clause of the 'cond form specially - the symbol
'else is not replaced with #t any more.

22 years ago* boot-9.scm (process-define-module): New options: :export-syntax,
Mikael Djurfeldt [Wed, 17 Oct 2001 12:02:13 +0000 (12:02 +0000)]
* boot-9.scm (process-define-module): New options: :export-syntax,
:re-export-syntax

22 years ago * version.c (scm_version): use sprintf instead of snprintf,
Gary Houston [Sun, 14 Oct 2001 21:28:23 +0000 (21:28 +0000)]
* version.c (scm_version): use sprintf instead of snprintf,
for portability.  thanks to Bill Schottstaedt.

22 years ago * configure.in: include sys/types.h when testing uint32_t.
Gary Houston [Sun, 14 Oct 2001 21:21:20 +0000 (21:21 +0000)]
* configure.in: include sys/types.h when testing uint32_t.
thanks to Bill Schottstaedt.

22 years agoAdded Michael Carmack. Removed "Chris Cramer" since he is already
Marius Vollmer [Sun, 14 Oct 2001 20:15:07 +0000 (20:15 +0000)]
Added Michael Carmack.  Removed "Chris Cramer" since he is already
listed as "Christopher".

22 years ago*** empty log message ***
Marius Vollmer [Sun, 14 Oct 2001 20:08:20 +0000 (20:08 +0000)]
*** empty log message ***

22 years agoDo not use an absolute path for <unistd.h> when checking for return
Marius Vollmer [Sun, 14 Oct 2001 20:08:08 +0000 (20:08 +0000)]
Do not use an absolute path for <unistd.h> when checking for return
type of usleep.  Thanks to Michael Carmack.

22 years ago* read.c (scm_lreadr): When user-defined hash procedure returns
Mikael Djurfeldt [Sun, 14 Oct 2001 16:05:04 +0000 (16:05 +0000)]
* read.c (scm_lreadr): When user-defined hash procedure returns
SCM_UNSPECIFIED: Fall back to standard handling instead of raising
an exception.  (This prevents parsing of uniform vectors from
interfering with parsing of numbers.)

* arrays.scm (read:uniform-vector): Return *unspecified* instead
of raising an exception if hash extend character isn't followed by
the array list. (This prevents parsing of uniform vectors from
interfering with parsing of numbers.)

22 years ago* tests/syntax.test: Added test cases for 'lambda' syntax.
Dirk Herrmann [Sat, 13 Oct 2001 23:59:27 +0000 (23:59 +0000)]
* tests/syntax.test:  Added test cases for 'lambda' syntax.

22 years ago* tests/syntax.test: Added test cases for 'case' syntax.
Dirk Herrmann [Sat, 13 Oct 2001 17:02:01 +0000 (17:02 +0000)]
* tests/syntax.test:  Added test cases for 'case' syntax.

22 years agoThank Ken Raeburn and Chris Cramer. Removed Jim and Maciej from
Marius Vollmer [Sat, 13 Oct 2001 15:43:38 +0000 (15:43 +0000)]
Thank Ken Raeburn and Chris Cramer.  Removed Jim and Maciej from
maintainer list.

22 years ago*** empty log message ***
Marius Vollmer [Sat, 13 Oct 2001 15:42:52 +0000 (15:42 +0000)]
*** empty log message ***

22 years ago(EXTRA_libguile_la_SOURCES): Added "mkstemp.c".
Marius Vollmer [Sat, 13 Oct 2001 15:42:07 +0000 (15:42 +0000)]
(EXTRA_libguile_la_SOURCES): Added "mkstemp.c".

22 years ago(display_backtrace_file_and_line): Only use scm_basename when POSIX
Marius Vollmer [Sat, 13 Oct 2001 15:41:40 +0000 (15:41 +0000)]
(display_backtrace_file_and_line): Only use scm_basename when POSIX
support is compiled in.  Thanks to Chris Cramer.

22 years ago(scm_m_atbind): First try to find the variable without defining it
Marius Vollmer [Sat, 13 Oct 2001 15:41:12 +0000 (15:41 +0000)]
(scm_m_atbind): First try to find the variable without defining it
locally; when it has not been found, define it locally.

22 years ago(module_variable): Pass over variables that exist but are unbound.
Marius Vollmer [Sat, 13 Oct 2001 15:40:29 +0000 (15:40 +0000)]
(module_variable): Pass over variables that exist but are unbound.

22 years agoSet NO_PREPRO_MAGIC when defining our version of PTRDIFF_MIN. Thanks
Marius Vollmer [Sat, 13 Oct 2001 15:39:53 +0000 (15:39 +0000)]
Set NO_PREPRO_MAGIC when defining our version of PTRDIFF_MIN.  Thanks
to Ken Raeburn.

22 years ago* numbers.c (mem2uinteger): Return number read so far when coming
Dirk Herrmann [Sat, 13 Oct 2001 12:39:26 +0000 (12:39 +0000)]
* numbers.c (mem2uinteger):  Return number read so far when coming
across a hexdigit after having read a # or if not reading a hex
value.  This will enable the calling code to correctly handle
forms like 1e2.  (The background is, that the exponent markers d,
e and f are also hexdigits.)  Thanks to Mikael Djurfeldt for
providing this patch.

(mem2complex):  Fix erroneous double-negation.  Now, numbers like
1-i will be read correctly.

22 years ago* tests/numbers.test: Added a test case that checks if valid
Dirk Herrmann [Sat, 13 Oct 2001 12:29:44 +0000 (12:29 +0000)]
* tests/numbers.test:  Added a test case that checks if valid
number strings are transformed correctly by string->number.

22 years ago* debug.c (scm_mem_to_proc): Fixed typo in previous change.
Mikael Djurfeldt [Fri, 12 Oct 2001 09:02:03 +0000 (09:02 +0000)]
* debug.c (scm_mem_to_proc): Fixed typo in previous change.

22 years ago* validate.h (SCM_VALIDATE_DOUBLE_DEF_COPY): New macro.
Mikael Djurfeldt [Fri, 12 Oct 2001 08:55:05 +0000 (08:55 +0000)]
* validate.h (SCM_VALIDATE_DOUBLE_DEF_COPY): New macro.

22 years ago* print.c (scm_print_state_vtable, print_state_pool):
Dirk Herrmann [Thu, 11 Oct 2001 23:20:52 +0000 (23:20 +0000)]
* print.c (scm_print_state_vtable, print_state_pool):
Initialize.  These variables are now registered as gc roots.

(scm_current_pstate):  Update documentation.

(scm_current_pstate, scm_make_print_state, scm_free_print_state,
scm_prin1, scm_init_print):  print_state_pool is registered as a
gc root and thus does not need to be protected by a surrounding
pair any more.

(make_print_state):  The car of print_state_pool no longer holds
the scm_print_state_vtable.

(scm_current_pstate, scm_make_print_state, print_circref,
scm_iprin1, scm_prin1, scm_iprlist):  Prefer !SCM_<foo> over
SCM_N<foo>.

(scm_prin1):  When building lists, prefer scm_list_<n> over
scm_cons[2]?.

(scm_iprlist):  Removed a redundant SCM_IMP test.

(scm_simple_format):  Use SCM_EQ_P to compare SCM values.

22 years ago* debug.c (scm_make_iloc): Prefer !SCM_<foo> over SCM_N<foo>.
Dirk Herrmann [Thu, 11 Oct 2001 22:26:25 +0000 (22:26 +0000)]
* debug.c (scm_make_iloc):  Prefer !SCM_<foo> over SCM_N<foo>.

(scm_memcons, scm_mem_to_proc):  When building lists, prefer
scm_list_<n> over scm_cons[2]?.

(scm_mem_to_proc):  Prefer SCM_CONSP over SCM_NIMP.

(scm_procedure_name):  Use SCM_CADR instead of explicit form.

(debugobj_print):  Coerce scm_intprint arg 1 to long, not int.
Thanks to Rob Browning for the patch (see log entry 2001-09-21) -
for some reason his patch didn't make it into the cvs.

22 years ago* numbers.c (mem2decimal_from_point): Cleaned up the parsing a
Dirk Herrmann [Tue, 9 Oct 2001 20:56:36 +0000 (20:56 +0000)]
* numbers.c (mem2decimal_from_point):  Cleaned up the parsing a
little bit - should even be somewhat more accurate now.

22 years ago* gc.c: support ia64 register backing store.
Rob Browning [Tue, 9 Oct 2001 03:33:16 +0000 (03:33 +0000)]
* gc.c: support ia64 register backing store.
(SCM_MARK_BACKING_STORE): new macro.

22 years ago* continuations.h: support ia64 register backing store.
Rob Browning [Tue, 9 Oct 2001 03:33:06 +0000 (03:33 +0000)]
* continuations.h: support ia64 register backing store.
(struct scm_t_contregs): add ia64 register backing store.

22 years ago* continuations.c: support ia64 register backing store.
Rob Browning [Tue, 9 Oct 2001 03:32:20 +0000 (03:32 +0000)]
* continuations.c: support ia64 register backing store.
(continuation_mark): mark ia64 register backing store.
(continuation_free): free ia64 register backing store.
(scm_make_continuation): capture ia64 register backing store.
(copy_stack_and_call): copy ia64 register backing store.

22 years ago*** empty log message ***
Rob Browning [Tue, 9 Oct 2001 03:32:12 +0000 (03:32 +0000)]
*** empty log message ***

22 years ago* hashtab.c (scm_hash_fn_create_handle_x): The result of assoc_fn
Dirk Herrmann [Mon, 8 Oct 2001 21:23:00 +0000 (21:23 +0000)]
* hashtab.c (scm_hash_fn_create_handle_x):  The result of assoc_fn
is known to be #f if no entry is found.  Thus, use !SCM_FALSEP
instead of SCM_NIMP to test for that case.

* strings.h (SCM_SET_STRING_LENGTH):  Cast the length to
scm_t_bits instead of long.

22 years ago* emacs.scm (%%load-port, %%emacs-load, %%emacs-eval-request,
Mikael Djurfeldt [Mon, 8 Oct 2001 12:40:14 +0000 (12:40 +0000)]
* emacs.scm (%%load-port, %%emacs-load, %%emacs-eval-request,
%%emacs-select-frame, %%emacs-frame-eval, %%emacs-symdoc,
%%apropos-internal): Use module-define! instead of
builtin-variable.

22 years agoCorrect cut and paste error: use SCM_POSFIXABLE for unsigned types.
Marius Vollmer [Sat, 6 Oct 2001 20:49:10 +0000 (20:49 +0000)]
Correct cut and paste error: use SCM_POSFIXABLE for unsigned types.

22 years agoCorrected definition of SCM_MOST_POSITIVE_FIXNUM. Boy was I confused.
Marius Vollmer [Sat, 6 Oct 2001 20:48:27 +0000 (20:48 +0000)]
Corrected definition of SCM_MOST_POSITIVE_FIXNUM.  Boy was I confused.

22 years ago*** empty log message ***
Marius Vollmer [Sat, 6 Oct 2001 20:01:36 +0000 (20:01 +0000)]
*** empty log message ***

22 years ago(INTEGRAL2NUM): Let the preprocessor test
Marius Vollmer [Sat, 6 Oct 2001 20:00:43 +0000 (20:00 +0000)]
(INTEGRAL2NUM): Let the preprocessor test
whether the integral type fits in a fixnum, not the compiler.
This removes a spurious compiler warning.  Also, honor the
NO_PREPRO_MAGIC flag to suppress any preprocessor tests.  This is
needed for `long long's.

22 years agoDefine NO_PREPRO_MAGOC when including
Marius Vollmer [Sat, 6 Oct 2001 19:58:24 +0000 (19:58 +0000)]
Define NO_PREPRO_MAGOC when including
num2integral.c.i for `long long' and `signed long long'.

22 years ago* tags.h (SCM_T_BITS_MAX, SCM_T_SIGNED_BITS_MAX,
Marius Vollmer [Sat, 6 Oct 2001 19:58:02 +0000 (19:58 +0000)]
* tags.h (SCM_T_BITS_MAX, SCM_T_SIGNED_BITS_MAX,
SCM_T_SIGNED_BITS_MIN): New.
* numbers.h (SCM_MOST_POSITIVE_FIXNUM, SCM_MOST_NEGATIVE_FIXNUM):
Use them to make these macros computable by the preprocessor.

22 years ago* coop.c (mother_awake_p): New variable.
Mikael Djurfeldt [Sat, 6 Oct 2001 16:30:20 +0000 (16:30 +0000)]
* coop.c (mother_awake_p): New variable.
(coop_create): Set mother_awake_p before creating or signalling
mother; wait until mother is going to sleep before returning.
(mother): Reset mother_awake_p before going to sleep.

22 years ago*** empty log message ***
Thien-Thi Nguyen [Sat, 6 Oct 2001 06:45:54 +0000 (06:45 +0000)]
*** empty log message ***

22 years ago(quoted?, clump): New procs, exported.
Thien-Thi Nguyen [Sat, 6 Oct 2001 06:44:48 +0000 (06:44 +0000)]
(quoted?, clump): New procs, exported.

22 years ago* options.c (protected_objects, scm_init_options): The content of
Dirk Herrmann [Fri, 5 Oct 2001 20:18:30 +0000 (20:18 +0000)]
* options.c (protected_objects, scm_init_options):  The content of
protected_objects is now protected from garbage collection using
scm_gc_register_root instead of scm_permanent_object.

(get_option_setting):  New static function that computes an option
setting as it was formerly done in the function scm_options.

(get_documented_option_setting):  New static function that
returns option documentation as it was formerly done in the
function scm_options.  Note that documentation C strings are no
longer precomputed into SCM objects.  Instead, they are converted
into SCM strings every time get_documented_option_setting is
called.

(change_option_setting):  New static functions that modifies the
option setting as it was formerly done in the function
scm_options.  The function is now exception safe, i. e. won't
cause a memory leak when interrupted.  Further, only non-immediate
option values are added to the protection list.

(scm_options):  This function now has only the purpose to dispatch
to to get_option_setting, get_documented_option_setting or
change_option_setting, depending on the arguments given to
scm_options.

(scm_init_opts):  Don't convert documentation C strings into SCM
strings.  Further, don't protect any object values:  They _must_
be immediate values, otherwise there is no guarantee that they
have not been collected before anyway.

* options.[ch] (scm_t_option):  Made type unsigned, name into a
constant char* and val into a scm_t_bits type.

(scm_options, scm_init_opts):  The number of options is guaranteed
to be larger or equal to zero.  Thus, the type is changed to
unsigned.

22 years ago* num2integral.i.c (NUM2INTEGRAL): Eliminated some warnings about
Dirk Herrmann [Fri, 5 Oct 2001 18:26:46 +0000 (18:26 +0000)]
* num2integral.i.c (NUM2INTEGRAL):  Eliminated some warnings about
testing an unsigned value for being >= 0.

22 years ago* Remove doc for read-and-eval! and eval2.
Neil Jerram [Fri, 5 Oct 2001 18:05:10 +0000 (18:05 +0000)]
* Remove doc for read-and-eval! and eval2.

22 years ago* numbers.h: Removed old comment about using SCM_CAR to access
Dirk Herrmann [Fri, 5 Oct 2001 17:43:49 +0000 (17:43 +0000)]
* numbers.h:  Removed old comment about using SCM_CAR to access
non-pair cells.

(SCM_MOST_POSITIVE_FIXNUM, SCM_MOST_NEGATIVE_FIXNUM):  Make sure
the return value is signed.  Thanks to Brian Crowder for the bug
report.

(SCM_SRS):  Avoid unnecessary casting and don't unpack input
values.  With this patch, SCM_SRS can be safely used for other
types than scm_t_signed_bits.  However, it should still better be
an internal macro and thus be renamed to SCM_I_SRS.

(SCM_MAKINUM, SCM_INUM):  Use proper casting.

22 years ago*** empty log message ***
Thien-Thi Nguyen [Fri, 5 Oct 2001 16:00:24 +0000 (16:00 +0000)]
*** empty log message ***

22 years ago(Writing): Add entry for `display'. Include in R5RS Index.
Thien-Thi Nguyen [Fri, 5 Oct 2001 15:58:51 +0000 (15:58 +0000)]
(Writing): Add entry for `display'.  Include in R5RS Index.

22 years ago * continuations.h, unif.h: in the descriptions of the bit patterns
Gary Houston [Wed, 3 Oct 2001 20:45:44 +0000 (20:45 +0000)]
* continuations.h, unif.h: in the descriptions of the bit patterns
of the heap cells, make bit 0 the least significant.

22 years ago*** empty log message ***
Thien-Thi Nguyen [Sun, 30 Sep 2001 22:33:38 +0000 (22:33 +0000)]
*** empty log message ***

22 years ago(EXTRA_DIST): Add BUGS file.
Thien-Thi Nguyen [Sun, 30 Sep 2001 22:32:13 +0000 (22:32 +0000)]
(EXTRA_DIST): Add BUGS file.

22 years agoInitial revision
Thien-Thi Nguyen [Sun, 30 Sep 2001 22:29:56 +0000 (22:29 +0000)]
Initial revision

22 years ago*** empty log message ***
Thien-Thi Nguyen [Sun, 30 Sep 2001 18:25:17 +0000 (18:25 +0000)]
*** empty log message ***

22 years ago(module-name->filename-frag, display-module-commentary): New procs.
Thien-Thi Nguyen [Sun, 30 Sep 2001 18:24:10 +0000 (18:24 +0000)]
(module-name->filename-frag, display-module-commentary): New procs.
(display-commentary): Also handle refs that look like module names.

22 years agoSome cosmetic doc fixes.
Martin Grabmüller [Wed, 26 Sep 2001 17:24:54 +0000 (17:24 +0000)]
Some cosmetic doc fixes.

22 years ago*** empty log message ***
Thien-Thi Nguyen [Wed, 26 Sep 2001 03:15:44 +0000 (03:15 +0000)]
*** empty log message ***

22 years ago(SCM_MAKE_CHAR): Use `scm_t_bits' instead of `intptr_t'.
Thien-Thi Nguyen [Wed, 26 Sep 2001 03:14:15 +0000 (03:14 +0000)]
(SCM_MAKE_CHAR): Use `scm_t_bits' instead of `intptr_t'.
Thanks to Golubev I. N.

22 years ago*** empty log message ***
Thien-Thi Nguyen [Wed, 26 Sep 2001 02:46:30 +0000 (02:46 +0000)]
*** empty log message ***

22 years ago(Syntax Rules): Add `cindex' directive.
Thien-Thi Nguyen [Wed, 26 Sep 2001 02:45:17 +0000 (02:45 +0000)]
(Syntax Rules): Add `cindex' directive.
Thanks to suggestion by Alex Schroeder.

22 years ago*** empty log message ***
Thien-Thi Nguyen [Wed, 26 Sep 2001 02:19:28 +0000 (02:19 +0000)]
*** empty log message ***

22 years agoAdd bugfix item to "Eventually".
Thien-Thi Nguyen [Wed, 26 Sep 2001 02:17:12 +0000 (02:17 +0000)]
Add bugfix item to "Eventually".

22 years ago * ports.c (scm_drain_input): extended the docstring. thanks to
Gary Houston [Tue, 25 Sep 2001 22:46:13 +0000 (22:46 +0000)]
* ports.c (scm_drain_input): extended the docstring.  thanks to
Alex Schroeder and Thien-Thi Nguyen.

22 years ago* boot-9.scm (process-define-module): Added :re-export.
Mikael Djurfeldt [Mon, 24 Sep 2001 11:12:23 +0000 (11:12 +0000)]
* boot-9.scm (process-define-module): Added :re-export.

22 years ago* validate.h (SCM_NUM2FLOAT, SCM_NUM2DOUBLE,
Mikael Djurfeldt [Sun, 23 Sep 2001 21:25:26 +0000 (21:25 +0000)]
* validate.h (SCM_NUM2FLOAT, SCM_NUM2DOUBLE,
SCM_VALIDATE_FLOAT_COPY, SCM_VALIDATE_DOUBLE_COPY): New
macros. (The NUM names might soon change.)

* numbers.h: Added missing declarations.

22 years ago* Makefile.am: Distribute num2float.i.c.
Mikael Djurfeldt [Sat, 22 Sep 2001 21:39:42 +0000 (21:39 +0000)]
* Makefile.am: Distribute num2float.i.c.

* num2float.i.c: New file, multiply included by numbers.c, used
to "templatize" the float <-> num conversion routines.

* numbers.c: New functions: scm_num2float, scm_float2num,
scm_num2double, scm_double2num.

22 years ago* srfi-19.scm (priv:split-real): Inserted missing call to
Mikael Djurfeldt [Sat, 22 Sep 2001 21:17:31 +0000 (21:17 +0000)]
* srfi-19.scm (priv:split-real): Inserted missing call to
inexact->exact.

22 years ago*** empty log message ***
Rob Browning [Fri, 21 Sep 2001 17:58:04 +0000 (17:58 +0000)]
*** empty log message ***

22 years ago* .cvsignore: really add version.h
Rob Browning [Fri, 21 Sep 2001 17:57:07 +0000 (17:57 +0000)]
* .cvsignore: really add version.h

22 years ago* chars.h (SCM_MAKE_CHAR): coerce value to intptr_t.
Rob Browning [Fri, 21 Sep 2001 17:56:39 +0000 (17:56 +0000)]
* chars.h (SCM_MAKE_CHAR): coerce value to intptr_t.

22 years ago* numbers.c (scm_gcd): change "k" to a long from an int.
Rob Browning [Fri, 21 Sep 2001 17:56:31 +0000 (17:56 +0000)]
* numbers.c (scm_gcd): change "k" to a long from an int.
Otherwise it fails on the alpha.  However, we might rather choose
this size conditionally.

22 years ago* strings.h (SCM_SET_STRING_LENGTH): coerce "l" to a long.
Rob Browning [Fri, 21 Sep 2001 17:56:17 +0000 (17:56 +0000)]
* strings.h (SCM_SET_STRING_LENGTH): coerce "l" to a long.
Otherwise it fails on the alpha.  However, we might rather choose
this size conditionally.

22 years ago* srfi-14.c (SCM_CHARSET_SET): need 1L, not just 1 in "<<".
Rob Browning [Fri, 21 Sep 2001 17:55:07 +0000 (17:55 +0000)]
* srfi-14.c (SCM_CHARSET_SET): need 1L, not just 1 in "<<".
(scm_char_set_hash): val needs to be long, not just unsigned.
(scm_char_set): need 1L, not just 1 in "<<".
(scm_list_to_char_set): need 1L, not just 1 in "<<".
(scm_list_to_char_set_x): need 1L, not just 1 in "<<".
(scm_list_to_char_set_x): FUNC_NAME was wrong - added a _x.
(scm_string_to_char_set): string length var needed to be
scm_sizet, not int.
(scm_string_to_char_set): need 1L, not just 1 in "<<".
(scm_string_to_char_set_x): string length var needed to be
scm_sizet, not int.
(scm_string_to_char_set_x): need 1L, not just 1 in "<<".
(scm_char_set_filter): need 1L, not just 1 in "<<".
(scm_char_set_filter_x): need 1L, not just 1 in "<<".
(scm_ucs_range_to_char_set): need 1L, not just 1 in "<<".
(scm_ucs_range_to_char_set_x): need 1L, not just 1 in "<<".
(scm_char_set_adjoin): need 1L, not just 1 in "<<".
(scm_char_set_delete): need 1L, not just 1 in "<<".
(scm_char_set_adjoin_x): need 1L, not just 1 in "<<".
(scm_char_set_delete_x): need 1L, not just 1 in "<<".

22 years ago* srfi-14.h (SCM_CHARSET_GET): need 1L, not just 1 in "<<".
Rob Browning [Fri, 21 Sep 2001 17:54:56 +0000 (17:54 +0000)]
* srfi-14.h (SCM_CHARSET_GET): need 1L, not just 1 in "<<".

22 years ago* tests/bit-operations.test (fixnum-bit): compute dynamically.
Rob Browning [Fri, 21 Sep 2001 17:54:35 +0000 (17:54 +0000)]
* tests/bit-operations.test (fixnum-bit): compute dynamically.

22 years ago* tests/numbers.test (fixnum-bit): compute dynamically.
Rob Browning [Fri, 21 Sep 2001 17:54:27 +0000 (17:54 +0000)]
* tests/numbers.test (fixnum-bit): compute dynamically.

22 years ago* numbers.c (scm_integer_expt): Accept inexact integer in second
Mikael Djurfeldt [Thu, 20 Sep 2001 18:53:32 +0000 (18:53 +0000)]
* numbers.c (scm_integer_expt): Accept inexact integer in second
argument.  (Thanks to Bill Schottstaedt.)

22 years ago*** empty log message ***
Rob Browning [Thu, 20 Sep 2001 16:48:02 +0000 (16:48 +0000)]
*** empty log message ***

22 years ago* configure.in (AC_CONFIG_FILES): add libguile/version.h.
Rob Browning [Thu, 20 Sep 2001 16:47:41 +0000 (16:47 +0000)]
* configure.in (AC_CONFIG_FILES): add libguile/version.h.

22 years ago* version.c
Rob Browning [Thu, 20 Sep 2001 16:47:01 +0000 (16:47 +0000)]
* version.c
(scm_major_version): support integer *_VERSION macros.
(scm_minor_version): support integer *_VERSION macros.
(scm_micro_version): support integer *_VERSION macros.
(scm_version): support integer *_VERSION macros.

22 years ago* version.h: renamed to version.h.in.
Rob Browning [Thu, 20 Sep 2001 16:46:52 +0000 (16:46 +0000)]
* version.h: renamed to version.h.in.

22 years ago* version.h.in: renamed from version.h.
Rob Browning [Thu, 20 Sep 2001 16:46:37 +0000 (16:46 +0000)]
* version.h.in: renamed from version.h.
(SCM_GUILE_MAJOR_VERSION): new public macro.
(SCM_GUILE_MINOR_VERSION): new public macro.
(SCM_GUILE_MICRO_VERSION): new public macro.

22 years ago* versiondat.h.in: removed (obsolete).
Rob Browning [Thu, 20 Sep 2001 16:46:25 +0000 (16:46 +0000)]
* versiondat.h.in: removed (obsolete).

22 years ago* error.c, error.h: Made error keys globally accessible.
Mikael Djurfeldt [Thu, 20 Sep 2001 08:18:26 +0000 (08:18 +0000)]
* error.c, error.h: Made error keys globally accessible.
Applications might want to test for these or use them in a direct
call to scm_error.

22 years ago* num2integral.i.c (NUM2INTEGRAL): Report an error when these
Mikael Djurfeldt [Thu, 20 Sep 2001 08:17:25 +0000 (08:17 +0000)]
* num2integral.i.c (NUM2INTEGRAL): Report an error when these
routines are passed an inexact.  This change in behavior is
motivated by concordance with R5RS: It is more common that a
primitive doesn't want to accept an inexact for an exact.

22 years ago*** empty log message ***
Thien-Thi Nguyen [Wed, 19 Sep 2001 13:36:48 +0000 (13:36 +0000)]
*** empty log message ***

22 years agoFix number typo in "Jump Start" section.
Thien-Thi Nguyen [Wed, 19 Sep 2001 13:36:05 +0000 (13:36 +0000)]
Fix number typo in "Jump Start" section.

22 years ago*** empty log message ***
Thien-Thi Nguyen [Wed, 19 Sep 2001 13:15:38 +0000 (13:15 +0000)]
*** empty log message ***

22 years agoFix improper `@result' usage.
Thien-Thi Nguyen [Wed, 19 Sep 2001 13:14:43 +0000 (13:14 +0000)]
Fix improper `@result' usage.

22 years ago*** empty log message ***
Thien-Thi Nguyen [Wed, 19 Sep 2001 13:04:22 +0000 (13:04 +0000)]
*** empty log message ***

22 years agoCommentary fix; nfc.
Thien-Thi Nguyen [Wed, 19 Sep 2001 13:03:25 +0000 (13:03 +0000)]
Commentary fix; nfc.

22 years ago*** empty log message ***
Thien-Thi Nguyen [Wed, 19 Sep 2001 12:57:38 +0000 (12:57 +0000)]
*** empty log message ***

22 years ago(process-use-modules): Fix typo.
Thien-Thi Nguyen [Wed, 19 Sep 2001 12:56:28 +0000 (12:56 +0000)]
(process-use-modules): Fix typo.

22 years ago* Makefile.am (psyntax.pp): Reference compile-psyntax.scm in
Mikael Djurfeldt [Tue, 18 Sep 2001 18:17:17 +0000 (18:17 +0000)]
* Makefile.am (psyntax.pp): Reference compile-psyntax.scm in
$(srcdir) in order to support separate build trees.

22 years ago* Removed function scm_gc_mark_cell_conservatively.
Dirk Herrmann [Mon, 17 Sep 2001 20:32:53 +0000 (20:32 +0000)]
* Removed function scm_gc_mark_cell_conservatively.

22 years ago * root.h (scm_root_state): removed the continuation_stack and
Gary Houston [Sat, 15 Sep 2001 21:56:52 +0000 (21:56 +0000)]
* root.h (scm_root_state): removed the continuation_stack and
continuation_stack_ptr members, which have no apparent purpose.
(scm_continuation_stack, scm_continuation_stack_ptr): #defines
removed.
* root.c (root_mark), init.c (restart_stack, start_stack), gc
(scm_igc): remove all references to contination_stack and
continuation_stack_ptr, avoiding allocation of a vector and
useless processing during gc.

22 years ago* Some renamings and minor fixes.
Dirk Herrmann [Thu, 13 Sep 2001 22:29:44 +0000 (22:29 +0000)]
* Some renamings and minor fixes.

22 years ago* Simplified handling of static glocal SCM variable.
Dirk Herrmann [Thu, 13 Sep 2001 21:58:15 +0000 (21:58 +0000)]
* Simplified handling of static glocal SCM variable.

22 years ago* Fixed some typing bugs. Thanks to Rob Browning.
Dirk Herrmann [Thu, 13 Sep 2001 21:22:37 +0000 (21:22 +0000)]
* Fixed some typing bugs.  Thanks to Rob Browning.

22 years ago* Added function scm_str2string. Thanks to Martin Baulig.
Dirk Herrmann [Wed, 12 Sep 2001 19:53:57 +0000 (19:53 +0000)]
* Added function scm_str2string.  Thanks to Martin Baulig.

22 years ago2001-09-12 Gary Houston <ghouston@arglist.com>
Gary Houston [Wed, 12 Sep 2001 19:15:33 +0000 (19:15 +0000)]
2001-09-12  Gary Houston  <ghouston@arglist.com>

* srfi-1.scm (filter): change "caller" to "filter" in check-arg-type.

22 years agoAdd `arity' enhancement news.
Thien-Thi Nguyen [Sun, 9 Sep 2001 01:09:24 +0000 (01:09 +0000)]
Add `arity' enhancement news.