bpt/guile.git
26 years ago* backtrace.c (display_error_body, display_backtrace_body),
Mikael Djurfeldt [Mon, 2 Feb 1998 15:00:14 +0000 (15:00 +0000)]
* backtrace.c (display_error_body, display_backtrace_body),
coop-threads.c (scheme_body_bootstrip, c_body_bootstrip),
gh_eval.c (eval_str_wrapper, eval_file_wrapper), init.c
(invoke_main_func), root.c (cwdr_body), throw.c (cwss_body,
scm_body_thunk, hbpca_body): Removed the second jmpbuf arg on body
functions.

26 years ago*** empty log message ***
Mikael Djurfeldt [Mon, 2 Feb 1998 14:59:26 +0000 (14:59 +0000)]
*** empty log message ***

26 years ago*** empty log message ***
Mikael Djurfeldt [Sun, 1 Feb 1998 21:05:14 +0000 (21:05 +0000)]
*** empty log message ***

26 years ago * async.c (async_pending): Removed declaration.
Mikael Djurfeldt [Sat, 31 Jan 1998 13:30:10 +0000 (13:30 +0000)]
* async.c (async_pending): Removed declaration.

26 years ago* coop.c, iselect.c: Since thread switches are now performed with
Mikael Djurfeldt [Fri, 30 Jan 1998 21:08:26 +0000 (21:08 +0000)]
* coop.c, iselect.c: Since thread switches are now performed with
interrupts masked, we can't use the old mechanism of delivering
signals immediately when they arrive.  Signals must instead be
delivered when the asyncs run *after* the end of the critical
section in scm_internal_select.  But this also means after context
switch so that the signal will be delivered to a different thread.
To avoid this, I have changed the protocol of
coop_wait_for_runnable_thread and friends so that they are allowed
to return the original thread.  So, if a signal arrives during
scm_internal_select, we won't any longer be forced do a context
switch, but can remain in the same thread and deliver the signal
to it.

* iselect.c: Small fixes.

26 years ago* eval.c (scm_eval, scm_deval), eq.c (scm_equal_p): Use SCM_TICK
Mikael Djurfeldt [Fri, 30 Jan 1998 21:07:51 +0000 (21:07 +0000)]
* eval.c (scm_eval, scm_deval), eq.c (scm_equal_p): Use SCM_TICK
instead of SCM_ASYNC_TICK.

26 years ago* dynwind.c (scm_wind_chain): New debug function.
Mikael Djurfeldt [Fri, 30 Jan 1998 21:07:14 +0000 (21:07 +0000)]
*  dynwind.c (scm_wind_chain): New debug function.

26 years ago* coop.c, iselect.c: Since thread switches are now performed with
Mikael Djurfeldt [Fri, 30 Jan 1998 21:07:00 +0000 (21:07 +0000)]
* coop.c, iselect.c: Since thread switches are now performed with
interrupts masked, we can't use the old mechanism of delivering
signals immediately when they arrive.  Signals must instead be
delivered when the asyncs run *after* the end of the critical
section in scm_internal_select.  But this also means after context
switch so that the signal will be delivered to a different thread.
To avoid this, I have changed the protocol of
coop_wait_for_runnable_thread and friends so that they are allowed
to return the original thread.  So, if a signal arrives during
scm_internal_select, we won't any longer be forced do a context
switch, but can remain in the same thread and deliver the signal
to it.

* coop.c (coop_mutex_init, coop_mutex_lock, coop_mutex_unlock,
coop_condition_variable_init, coop_condition_variable_wait,
coop_condition_variable_signal): Changed return type from `void'
to `int'.  This is to adhere closer to the pthreads interface.
This, in turn, is part of an attempt to provide C versions of the
mutex and condition variable primitives which can be part of a
frontend to COOP or pthreads.

* coop.c (coop_mutex_destroy, coop_condition_variable_wait_mutex,
coop_condition_variable_destroy): New functions.

26 years ago* coop-threads.c (scheme_launch_data, scheme_body_bootstrip,
Mikael Djurfeldt [Fri, 30 Jan 1998 21:06:47 +0000 (21:06 +0000)]
* coop-threads.c (scheme_launch_data, scheme_body_bootstrip,
scheme_handler_bootstrip, scheme_launch_thread, c_launch_data,
c_body_bootstrip, c_handler_bootstrip, c_launch_thread): Add an
extra layer of functions around the body and handler of a thread.
This extra layer makes sure that the handler is called in the
dynamic context of the surround (= empty dynwind list), but under
the *dynamic root* of the body.  We can not use the dynamic root
of the surround since that root belongs to another thread => stack
is not handled correctly.  It may seem ugly to use this extra
layer, but the extra cost in terms of execution time is really
negligible compared to the total amount of time required to create
a thread, and it would reduce maintainability to duplicate the
crucial and complicated steps performed by cwdr.

* coop-threads.c (scm_wait_condition_variable): Use
coop_condition_variable_wait_mutex.

26 years ago* async.c, async.h (asyncs_pending): Renamed asyncs_pending -->
Mikael Djurfeldt [Fri, 30 Jan 1998 21:06:31 +0000 (21:06 +0000)]
* async.c, async.h (asyncs_pending): Renamed asyncs_pending -->
scm_asyncs_pending and made it global.

26 years ago* __scm.h (SCM_ASYNC_TICK): Removed thread switching code.
Mikael Djurfeldt [Fri, 30 Jan 1998 21:06:16 +0000 (21:06 +0000)]
* __scm.h (SCM_ASYNC_TICK): Removed thread switching code.
(SCM_ALLOW_INTS): Added thread switching code before interrupts
get re-enabled.  The important effect of this is that interrupts
are blocked during thread switching so that thread data structures
don't risk getting messed up by an unfortunate signal.
(SCM_REDEFER_INTS, SCM_REALLOW_INTS): It turned out that gcc-2.8.0
seems to do more aggressive optimization which actually move
instructions around in these macros in a fatal way.  Therefore:
Introduce Anthony's SCM_FENCE macro!  (And I who thought he was
just superstitious...)
(SCM_TICK): Maybe do a context switch and take care of asyncs.
This macro should be used instead of SCM_ASYNC_TICK since the
latter doesn't do context switches any more.

26 years ago* threads.scm: Added simple error and signal handler.
Mikael Djurfeldt [Fri, 30 Jan 1998 21:05:53 +0000 (21:05 +0000)]
* threads.scm: Added simple error and signal handler.
(make-thread, begin-handler): Use this handler.  The most
important effect of this is that signals get unmasked.
Previously, when a signal was thrown signals remained masked
(signals get masked when a signal is taken) which influenced other
threads.

26 years agoSmall fix to scm_getc: Move FD_ZERO into the select loop
Mikael Djurfeldt [Fri, 30 Jan 1998 21:02:36 +0000 (21:02 +0000)]
Small fix to scm_getc: Move FD_ZERO into the select loop

26 years ago*** empty log message ***
Mikael Djurfeldt [Mon, 26 Jan 1998 01:43:43 +0000 (01:43 +0000)]
*** empty log message ***

26 years ago* coop.c (coop_mutex_init, coop_mutex_lock, coop_mutex_unlock,
Mikael Djurfeldt [Mon, 26 Jan 1998 01:43:16 +0000 (01:43 +0000)]
* coop.c (coop_mutex_init, coop_mutex_lock, coop_mutex_unlock,
coop_condition_variable_init, coop_condition_variable_wait,
coop_condition_variable_signal): Changed return type from `void'
to `int'.  This is to adhere closer to the pthreads interface.
This, in turn, is part of an attempt to provide C versions of the
mutex and condition variable primitives which can be part of a
frontend to COOP or pthreads.

* coop.c (coop_mutex_destroy, coop_condition_variable_wait_mutex,
coop_condition_variable_destroy): New functions.

* coop-threads.c (scm_wait_condition_variable): Use
coop_condition_variable_wait_mutex.

* coop-threads.h, coop-defs.h (coop_q_t, coop_m, coop_c):
Definitions moved to coop-defs.h.

* coop-defs.h (scm_mutex_init, scm_mutex_lock, scm_mutex_unlock,
scm_mutex_destroy, scm_cond_init, scm_cond_wait, scm_cond_signal,
scm_cond_destroy): New C interface to mutecis and cond vars.

26 years ago* iselect.c: Small fixes.
Mikael Djurfeldt [Mon, 26 Jan 1998 01:42:43 +0000 (01:42 +0000)]
* iselect.c: Small fixes.

26 years ago* error.c (scm_wta): Added support for SCM_ARG6 and SCM_ARG7.
Mikael Djurfeldt [Sun, 25 Jan 1998 12:00:47 +0000 (12:00 +0000)]
* error.c (scm_wta): Added support for SCM_ARG6 and SCM_ARG7.

26 years ago* iselect.c: Now several threads can wait on the same file
Mikael Djurfeldt [Sat, 24 Jan 1998 01:50:43 +0000 (01:50 +0000)]
* iselect.c: Now several threads can wait on the same file
descriptor.  The behaviour is compatible with OS select: All
threads waiting for the fd return with the same status.

26 years agoUpdated
Mikael Djurfeldt [Sat, 24 Jan 1998 01:49:01 +0000 (01:49 +0000)]
Updated

26 years ago* coop-threads.c, threads.h (scm_spawn_thread): New function.
Mikael Djurfeldt [Fri, 23 Jan 1998 20:51:47 +0000 (20:51 +0000)]
* coop-threads.c, threads.h (scm_spawn_thread): New function.
Can spawn a thread from application C code.

26 years ago* gh.h, gh_data.c (gh_doubles2scm, gh_doubles2dvect,
Mikael Djurfeldt [Tue, 20 Jan 1998 18:03:18 +0000 (18:03 +0000)]
* gh.h, gh_data.c (gh_doubles2scm, gh_doubles2dvect,
gh_doubles2scm): New functions.

26 years ago* gh.h, gh_data.c (gh_doubles2scm, gh_doubles2dvect,
Mikael Djurfeldt [Tue, 20 Jan 1998 17:57:01 +0000 (17:57 +0000)]
* gh.h, gh_data.c (gh_doubles2scm, gh_doubles2dvect,
gh_doubles2scm): New functions.

26 years ago*** empty log message ***
Jim Blandy [Mon, 19 Jan 1998 21:22:13 +0000 (21:22 +0000)]
*** empty log message ***

26 years agosome trivial cleanup
Mark Galassi [Fri, 16 Jan 1998 00:32:28 +0000 (00:32 +0000)]
some trivial cleanup

26 years agoFixed copyright notice on complex transcendental functions. Jerry
Jim Blandy [Mon, 5 Jan 1998 22:10:10 +0000 (22:10 +0000)]
Fixed copyright notice on complex transcendental functions.  Jerry
Hedden assigned the copyright to the FSF a while ago.

26 years agoNew file.
Tim Pierce [Mon, 5 Jan 1998 08:29:54 +0000 (08:29 +0000)]
New file.

26 years ago * socket.c (scm_bind): free soka after use.
Gary Houston [Sun, 4 Jan 1998 02:50:12 +0000 (02:50 +0000)]
* socket.c (scm_bind): free soka after use.
* stime.c (tzvar): new variable.
(setzone, restorezone, scm_localtime, scm_mktime, scm_strftime):
avoid memory leaks when allocating.

26 years ago*** empty log message ***
Jim Blandy [Sun, 4 Jan 1998 00:31:25 +0000 (00:31 +0000)]
*** empty log message ***

26 years ago* gc.c, tags.h: Doc fixes.
Jim Blandy [Sun, 4 Jan 1998 00:31:13 +0000 (00:31 +0000)]
* gc.c, tags.h: Doc fixes.

26 years ago* iselect.h: Some systems require <sys/types.h> to get the FD_SET
Jim Blandy [Sun, 4 Jan 1998 00:31:05 +0000 (00:31 +0000)]
* iselect.h: Some systems require <sys/types.h> to get the FD_SET
macro definitions.

26 years agoDoc fix.
Jim Blandy [Sun, 4 Jan 1998 00:09:54 +0000 (00:09 +0000)]
Doc fix.

26 years ago* eval.c (macro-eval!): Removed. This function was a design bug.
Mikael Djurfeldt [Fri, 2 Jan 1998 16:44:05 +0000 (16:44 +0000)]
* eval.c (macro-eval!): Removed.  This function was a design bug.
It allowed memoized code to leak out to the scheme level.  Most
things that you could do with `macro-eval!' can be done with
`local-eval'.

26 years agomacro-eval! removed
Mikael Djurfeldt [Fri, 2 Jan 1998 16:43:37 +0000 (16:43 +0000)]
macro-eval! removed

26 years ago A better fix to the SLIB identity problem -- thanks to Marius Vollmer.
Tim Pierce [Fri, 2 Jan 1998 03:36:36 +0000 (03:36 +0000)]
A better fix to the SLIB identity problem -- thanks to Marius Vollmer.
* slib.scm (identity): Unmake public.
(slib:eval): Evaluate inside `slib-module'.

26 years agoChanges for SLIB compatibility.
Tim Pierce [Fri, 26 Dec 1997 02:59:08 +0000 (02:59 +0000)]
Changes for SLIB compatibility.

26 years agoBring slib.scm up to date with current SLIB release, hopefully.
Tim Pierce [Wed, 24 Dec 1997 06:10:47 +0000 (06:10 +0000)]
Bring slib.scm up to date with current SLIB release, hopefully.

26 years ago * fports.c (scm_pipob): Use scm_generic_fgets for line i/o, since
Tim Pierce [Sat, 20 Dec 1997 13:52:09 +0000 (13:52 +0000)]
* fports.c (scm_pipob): Use scm_generic_fgets for line i/o, since
scm_fgets now depends on ftell(3) to know how many bytes were
read.  Sigh.

26 years ago * gh_data.c (gh_scm2newstr, gh_get_substr): Use RO macros for
Tim Pierce [Tue, 16 Dec 1997 05:09:50 +0000 (05:09 +0000)]
* gh_data.c (gh_scm2newstr, gh_get_substr): Use RO macros for
dealing with strings.

26 years agoboot-9.scm (read-line): Rewritten to use %read-line.
Tim Pierce [Sat, 13 Dec 1997 08:33:33 +0000 (08:33 +0000)]
boot-9.scm (read-line): Rewritten to use %read-line.
ioext.c (%read-line): Return a cons of the input line and its terminator.
fports.c, genio.c, genio.h, ports.c, ports.h: Change fgets methods
to return the string length as well as its contents.

26 years ago* filesys.c (set_element): Return file descriptor.
Mikael Djurfeldt [Mon, 8 Dec 1997 18:15:44 +0000 (18:15 +0000)]
* filesys.c (set_element): Return file descriptor.
(fill_select_type): Return the highest file descriptor.
(scm_select): Tell select about the highest file descriptor.  On
some systems the SELECT_SET_SIZE can be as much as 128 bytes.
Therefore the extra overhead for calculating the maximum fd seems
to be more than compensated.  Is this correct?  In any case,
scm_internal_select will be much faster with this info.
(scm_select, fill_select_type, set_element): Don't accept any kind
of object in the file descriptor list or vector.

26 years ago * README: using Automake 1.2d
Gary Houston [Sun, 7 Dec 1997 07:02:17 +0000 (07:02 +0000)]
* README: using Automake 1.2d
* configure.in: AC_CHECK_FUNCS: add "system".
* simpos.c (scm_system): always define: use sysmissing if not
available.  Check for HAVE_SYSTEM instead of _Windows (does
Windows lack system or does it have an unusable one?).
Check for error conditions -1 and 127.  Use SCM_DEFER_INTS.
Let the argument be optional: if not supplied, call system(NULL).
* ports.c (scm_close_port): relax the type check from OPPORTP to
PORTP; closing a closed port is allowed.

26 years ago * fports.c (scm_fgets): Return if the last char in a chunk is
Tim Pierce [Thu, 4 Dec 1997 19:05:05 +0000 (19:05 +0000)]
* fports.c (scm_fgets): Return if the last char in a chunk is
newline.  When fgets returns a string whose length is `size-1', it
is ambiguous whether a whole line was retrieved, so we must check
explicitly whether a line terminator is present.

26 years ago* filesys.c (scm_stat): Slightly optimized.
Mikael Djurfeldt [Thu, 4 Dec 1997 16:40:53 +0000 (16:40 +0000)]
* filesys.c (scm_stat): Slightly optimized.

26 years ago* print.h (SCM_COERCE_OUTPORT): Check that the object is a pair
Mikael Djurfeldt [Thu, 4 Dec 1997 16:18:59 +0000 (16:18 +0000)]
* print.h (SCM_COERCE_OUTPORT): Check that the object is a pair
before taking the CDR.

26 years ago* filesys.c (scm_stat): Coerce output port only if argument *is*
Mikael Djurfeldt [Thu, 4 Dec 1997 16:03:58 +0000 (16:03 +0000)]
* filesys.c (scm_stat): Coerce output port only if argument *is*
an output port!  (Thanks to Harald Meland.)

26 years ago*** empty log message ***
Jim Blandy [Wed, 3 Dec 1997 17:31:51 +0000 (17:31 +0000)]
*** empty log message ***

26 years ago* guile-snarf.in: Pass args through to gcc in a way that preserves
Jim Blandy [Wed, 3 Dec 1997 17:30:39 +0000 (17:30 +0000)]
* guile-snarf.in: Pass args through to gcc in a way that preserves
whitespace boundaries.  (Thanks to Greg Badros.)

26 years ago* ports.c (scm_close_port): Make sure the port is open before
Jim Blandy [Wed, 3 Dec 1997 17:30:24 +0000 (17:30 +0000)]
* ports.c (scm_close_port): Make sure the port is open before
trying to close it.

26 years agoAdd copyrights; tweak comments.
Jim Blandy [Wed, 3 Dec 1997 17:29:12 +0000 (17:29 +0000)]
Add copyrights; tweak comments.

26 years ago * md/axp.s (qt_vstart): Typo fixes, thanks to Alexander Jolk.
Tim Pierce [Tue, 2 Dec 1997 17:36:19 +0000 (17:36 +0000)]
* md/axp.s (qt_vstart): Typo fixes, thanks to Alexander Jolk.

26 years ago * stacks.c (scm_frame_procedure): Reverse the logic in the return
Tim Pierce [Tue, 2 Dec 1997 17:34:07 +0000 (17:34 +0000)]
* stacks.c (scm_frame_procedure): Reverse the logic in the return
statement.  (Thanks to Doug Evans for pointing this out.)

26 years agoacconfig.h: Add USCORE.
Tim Pierce [Mon, 1 Dec 1997 18:22:04 +0000 (18:22 +0000)]
acconfig.h: Add USCORE.
scmconfig.h.in: Regenerated.

26 years ago* coop-defs.h (struct coop_t): Renamed errno --> _errno to prevent
Mikael Djurfeldt [Sun, 30 Nov 1997 10:36:13 +0000 (10:36 +0000)]
* coop-defs.h (struct coop_t): Renamed errno --> _errno to prevent
errno macro expansion of this field name.  (errno is a C
preprocessor macro on some systems.)

26 years agoDoc fix.
Tim Pierce [Sun, 30 Nov 1997 02:35:18 +0000 (02:35 +0000)]
Doc fix.

26 years ago* init.c (scm_start_stack): Removed initialization of
Mikael Djurfeldt [Sat, 29 Nov 1997 10:49:18 +0000 (10:49 +0000)]
* init.c (scm_start_stack): Removed initialization of
scm_the_last_stack_var.

* backtrace.h: Declare scm_the_last_stack_var.

* backtrace.c: Define scm_the_last_stack_var.

* root.c (mark_root): Don't mark the_last_stack_var.

* root.h (scm_root_state): Removed the_last_stack_var.

* throw.c: Added #include "fluids.h"
(ss_handler): `the-last-stack' is now a fluid.

* (backtrace.h, backtrace.c, throw.c): Renamed the_last_stack_var
--> the_last_stack_fluid.

* backtrace.c: Added #include "fluids.h"
(scm_init_backtrace): Initialize `the-last-stack' to a fluid.
(scm_backtrace): `the-last-stack' is now a fluid.

* init.c (scm_boot_guile_1): Moved call to scm_init_backtrace
after scm_init_fluids.

26 years ago* boot-9.scm (error-catching-loop, save-stack): `the-last-stack'
Mikael Djurfeldt [Sat, 29 Nov 1997 01:11:21 +0000 (01:11 +0000)]
* boot-9.scm (error-catching-loop, save-stack): `the-last-stack'
is now a fluid.

26 years ago* backtrace.c: Added #include "fluids.h"
Mikael Djurfeldt [Sat, 29 Nov 1997 01:10:21 +0000 (01:10 +0000)]
* backtrace.c: Added #include "fluids.h"
(scm_init_backtrace): Initialize `the-last-stack' to a fluid.
(scm_backtrace): `the-last-stack' is now a fluid.

* init.c (scm_boot_guile_1): Moved call to scm_init_backtrace
after scm_init_fluids.

26 years ago * iselect.c: #ifdef USE_THREADS around thread-related includes.
Tim Pierce [Fri, 28 Nov 1997 22:58:10 +0000 (22:58 +0000)]
* iselect.c: #ifdef USE_THREADS around thread-related includes.

* dynl-dl.c (sysdep_dynl_func): Check both USCORE and
DLSYM_ADDS_USCORE to decide whether to add an underscore.

26 years agoLibtool support for find-and-link-dynamic-module
Tim Pierce [Fri, 28 Nov 1997 17:22:20 +0000 (17:22 +0000)]
Libtool support for find-and-link-dynamic-module

26 years ago* iselect.c (coop_next_runnable_thread,
Mikael Djurfeldt [Fri, 28 Nov 1997 01:11:12 +0000 (01:11 +0000)]
* iselect.c (coop_next_runnable_thread,
coop_wait_for_runnable_thread): Disable interrupts so that no
async is executed before a potential error_revive.
(scm_internal_select): Disable interrupts during the parts of the
code which manipulate the sleep queue and the file descriptors.

26 years ago* iselect.c (coop_next_runnable_thread,
Mikael Djurfeldt [Fri, 28 Nov 1997 00:50:58 +0000 (00:50 +0000)]
* iselect.c (coop_next_runnable_thread,
coop_wait_for_runnable_thread): Disable interrupts so that no
async is executed before a potential error_revive.

26 years ago* iselect.h, iselect.c: New files. Implements
Mikael Djurfeldt [Thu, 27 Nov 1997 18:06:33 +0000 (18:06 +0000)]
* iselect.h, iselect.c: New files.  Implements
scm_internal_select.  (See NEWS.)

26 years agoUpdated
Mikael Djurfeldt [Thu, 27 Nov 1997 18:05:28 +0000 (18:05 +0000)]
Updated

26 years ago* Makefile.am: Added iselect.c and iselect.h.
Mikael Djurfeldt [Thu, 27 Nov 1997 18:04:56 +0000 (18:04 +0000)]
* Makefile.am: Added iselect.c and iselect.h.

* coop.c (coop_qinit): Initialize fields used by
scm_internal_select.
(coop_qget, coop_qget, coop_tmp_queue): Made global.
(coop_next_runnable_thread): If GUILE_ISELECT enabled, use
replacement in iselect.c.
(coop_mutex_lock, coop_condition_variable_wait, coop_abort,
coop_join): If GUILE_ISELECT enabled, use
coop_wait_for_runnable_thread instead of
coop_next_runnable_thread.
(usleep, sleep): New replacements for system functions if
GUILE_ISELECT is enabled.

* coop-threads.h: Declare coop_wait_for_runnable_thread.

* coop-defs.h (coop_t): Added fields used by scm_internal_select.

* filesys.c: Added #include "iselect.h".  Moved FD-macros to
iselect.h.  Implement Scheme level `select' using
scm_internal_select.  (See NEWS.)

* genio.c (scm_getc): Block with scm_internal_select.  (See NEWS.)

* init.c: Call scm_init_iselect.

26 years agoUpdated using the correct versions of libtool and automake
Mikael Djurfeldt [Thu, 27 Nov 1997 18:04:01 +0000 (18:04 +0000)]
Updated using the correct versions of libtool and automake

26 years ago* configure.in: Added code to enable GUILE_ISELECT on systems
Mikael Djurfeldt [Thu, 27 Nov 1997 18:03:06 +0000 (18:03 +0000)]
* configure.in: Added code to enable GUILE_ISELECT on systems
which have the necessary functions (gettimeofday, select).

* acconfig.h: Added GUILE_ISELECT.

26 years ago Fix a memory leak in scm_read_line and a type cast bug in the ptob.
Tim Pierce [Thu, 27 Nov 1997 06:39:46 +0000 (06:39 +0000)]
Fix a memory leak in scm_read_line and a type cast bug in the ptob.
* fports.c (scm_fgets): Use malloc/free rather than scm_must_malloc
  and scm_must_free, since ultimately the string returned will be copied
  by scm_makfrom0str anyway.  Also, read any characters that may have
been pushed onto the port with scm_ungetc.
* ports.c (scm_generic_fgets): Same as for scm_fgets.
* ioext.c (scm_read_line): Free string after Guilifying it.
* ports.h (scm_ptobfuns): fgets method returns a char *, not a char.

26 years agogh bug fix
Anthony Green [Wed, 26 Nov 1997 03:45:23 +0000 (03:45 +0000)]
gh bug fix

26 years agogh bug fix
Anthony Green [Wed, 26 Nov 1997 03:13:00 +0000 (03:13 +0000)]
gh bug fix

26 years agoRegenerated aclocal.m4 and configure with all the necessary
Tim Pierce [Wed, 26 Nov 1997 02:16:16 +0000 (02:16 +0000)]
Regenerated aclocal.m4 and configure with all the necessary
macros (didn't get AM_PROG_LIBTOOL or GUILE_DLSYM_USCORE correct
last time).

26 years agocompleting changes to gh_repl and uniform array stuff
Mark Galassi [Tue, 25 Nov 1997 17:20:23 +0000 (17:20 +0000)]
completing changes to gh_repl and uniform array stuff

26 years agomade changes in gh_repl() and started adding the uniform array stuff to gh_
Mark Galassi [Tue, 25 Nov 1997 06:18:08 +0000 (06:18 +0000)]
made changes in gh_repl() and started adding the uniform array stuff to gh_

26 years ago*** empty log message ***
Tim Pierce [Tue, 25 Nov 1997 02:43:01 +0000 (02:43 +0000)]
*** empty log message ***

26 years ago*** empty log message ***
Tim Pierce [Tue, 25 Nov 1997 00:08:51 +0000 (00:08 +0000)]
*** empty log message ***

26 years agofixed dynl memory leak
Tim Pierce [Sat, 22 Nov 1997 19:15:09 +0000 (19:15 +0000)]
fixed dynl memory leak

26 years agoAdded support for leading dlsym underscore.
Tim Pierce [Fri, 21 Nov 1997 19:39:23 +0000 (19:39 +0000)]
Added support for leading dlsym underscore.

26 years ago*** empty log message ***
Tim Pierce [Thu, 20 Nov 1997 04:40:02 +0000 (04:40 +0000)]
*** empty log message ***

26 years ago* boot-9.scm (set-batch-mode?!, batch-mode?): initialize more
Gary Houston [Sun, 9 Nov 1997 23:36:17 +0000 (23:36 +0000)]
* boot-9.scm (set-batch-mode?!, batch-mode?): initialize more
usefully so they will work from a script.

26 years agoSmall fix of last fix...
Mikael Djurfeldt [Thu, 6 Nov 1997 17:50:19 +0000 (17:50 +0000)]
Small fix of last fix...

26 years ago* regex-posix.c (scm_free_regex_t): Return size of regex_t instead
Mikael Djurfeldt [Thu, 6 Nov 1997 15:09:30 +0000 (15:09 +0000)]
* regex-posix.c (scm_free_regex_t): Return size of regex_t instead
of 0; size_t --> scm_size_t.  Thanks to Bernard Urban.

26 years ago*** empty log message ***
Jim Blandy [Thu, 6 Nov 1997 06:15:27 +0000 (06:15 +0000)]
*** empty log message ***

26 years ago*** empty log message ***
Marius Vollmer [Fri, 31 Oct 1997 22:18:09 +0000 (22:18 +0000)]
*** empty log message ***

26 years ago* boot-9.scm (inherit-print-state): Moved definition to the
Marius Vollmer [Fri, 31 Oct 1997 22:16:24 +0000 (22:16 +0000)]
* boot-9.scm (inherit-print-state): Moved definition to the
neighborhood of the record code.

26 years agoUpdate FTP address for libtool.
Jim Blandy [Mon, 27 Oct 1997 19:09:58 +0000 (19:09 +0000)]
Update FTP address for libtool.

26 years ago*** empty log message ***
Jim Blandy [Mon, 27 Oct 1997 07:07:15 +0000 (07:07 +0000)]
*** empty log message ***

26 years ago* boot-9.scm: Revert changes to this file from Oct 23. It turns
Jim Blandy [Mon, 27 Oct 1997 07:07:03 +0000 (07:07 +0000)]
* boot-9.scm: Revert changes to this file from Oct 23.  It turns
out to interact badly with the Emacs support and the Tcl/Tk
support.  It's not a high enough priority at the moment to be
worth fixing.  I'm leaving the other readline support in, though.

26 years ago* scmconfig.h.in: Updated (HAVE_LIBTERMCAP was added when
Mikael Djurfeldt [Sun, 26 Oct 1997 16:06:23 +0000 (16:06 +0000)]
* scmconfig.h.in: Updated (HAVE_LIBTERMCAP was added when
configure.in was changed).

26 years ago* README (libtool): Tell people to use version 1.0e.
Mikael Djurfeldt [Sun, 26 Oct 1997 15:44:00 +0000 (15:44 +0000)]
* README (libtool): Tell people to use version 1.0e.

26 years ago*** empty log message ***
Jim Blandy [Sun, 26 Oct 1997 07:24:36 +0000 (07:24 +0000)]
*** empty log message ***

26 years ago* Makefile.am (modinclude_HEADERS): Include readline.h here.
Jim Blandy [Sun, 26 Oct 1997 07:24:04 +0000 (07:24 +0000)]
* Makefile.am (modinclude_HEADERS): Include readline.h here.
* Makefile.in: Regenerated.

26 years ago*** empty log message ***
Marius Vollmer [Sat, 25 Oct 1997 21:57:48 +0000 (21:57 +0000)]
*** empty log message ***

26 years ago* print.h (SCM_COERCE_OPORT): Renamed to SCM_COERCE_OUTPORT. An
Marius Vollmer [Sat, 25 Oct 1997 21:54:46 +0000 (21:54 +0000)]
* print.h (SCM_COERCE_OPORT): Renamed to SCM_COERCE_OUTPORT.  An
OPORT is an `open' port, not an output port.

26 years ago* filesys.c (scm_close, set_element, get_element, scm_chown,
Marius Vollmer [Sat, 25 Oct 1997 21:54:12 +0000 (21:54 +0000)]
* filesys.c (scm_close, set_element, get_element, scm_chown,
scm_chmod, scm_stat, scm_truncate_file, scm_fcntl, scm_fsync): Use
SCM_COERCE_OUTPORT to cope with the printstate/port magic.
* ports.c (scm_port_revealed, scm_set_port_revealed_x,
scm_close_port, scm_port_line, scm_set_port_line_x,
scm_port_column, scm_set_port_column_x, scm_port_filename,
scm_set_port_filename_x, scm_port_mode,
scm_close_all_ports_except, scm_set_current_output_port,
scm_set_current_error_port): Likewise
* ioext.c (scm_redirect_port, scm_dup_to_fdes, scm_freopen,
scm_ftell, scm_fileno, scm_isatty_p, scm_primitive_move_to_fdes):
Likewise
* posix.c (scm_ttyname, scm_tcgetpgrp, scm_tcsetpgrp): Likewise
* backtrace.c (display_backtrace_body): Likewise
* fports (scm_setvbuf): Likewise
* socket.c (scm_getsockopt, scm_setsockopt, scm_shutdown,
scm_connect, scm_bind, scm_listen, scm_accept, scm_getsockname,
scm_getpeername, scm_send, scm_sendto): Likewise
* unif.c (scm_uniform_array_write): Likewise

26 years ago*** empty log message ***
Marius Vollmer [Sat, 25 Oct 1997 21:52:05 +0000 (21:52 +0000)]
*** empty log message ***

26 years ago* Makefile.am: Include readline.scm in the list of files to be
Jim Blandy [Sat, 25 Oct 1997 18:24:01 +0000 (18:24 +0000)]
* Makefile.am: Include readline.scm in the list of files to be
installed, so Guile can find it for interactive use.
* Makefile.in: Regenerated.

26 years ago*** empty log message ***
Jim Blandy [Sat, 25 Oct 1997 06:55:55 +0000 (06:55 +0000)]
*** empty log message ***

26 years ago* Makefile.am: Call the library libqthreads.a, not libqt.a. The
Jim Blandy [Sat, 25 Oct 1997 06:54:51 +0000 (06:54 +0000)]
* Makefile.am: Call the library libqthreads.a, not libqt.a.  The
old name conflicts with the Qt user interface toolkit.
* Makefile.in: Regenerated.

26 years agoMinor problems with substring-related tag changes.
Jim Blandy [Sat, 25 Oct 1997 06:53:47 +0000 (06:53 +0000)]
Minor problems with substring-related tag changes.
* symbols.h (SCM_SUBSTRP): Don't mask off the S bit; that's
exactly what we want to leave in to detect substrings.
(SCM_ROSTRINGP, ROUCHARS): Formatting tweaks.
* tags.h: Fix diagrams and comments describing the S tag bit;
remove vestigial remarks about the D tag bit.
(SCM_TYP7, SCM_TYP7S): Rephrased for readability.
* strings.c: Formatting tweaks.

26 years ago* load.c (scm_init_load_path): Check GUILE_LOAD_PATH environment
Jim Blandy [Sat, 25 Oct 1997 06:53:11 +0000 (06:53 +0000)]
* load.c (scm_init_load_path): Check GUILE_LOAD_PATH environment
variable first; then SCHEME_LOAD_PATH, with a warning message.
(scm_parse_path): New function.
* script.c: Doc fixes.

26 years agoCall the QuickThreads library libqthreads.a, not libqt.a. The old
Jim Blandy [Sat, 25 Oct 1997 06:52:10 +0000 (06:52 +0000)]
Call the QuickThreads library libqthreads.a, not libqt.a.  The old
name conflicts with the Qt user interface toolkit.
* threads.m4 (CY_AC_WITH_THREADS): Use new library name.
* configure.in: Same.
* aclocal.m4, configure: Regenerated.