Thanks to Inge Gutheil for raising this problem.
* doc/ref/api-control.texi (Dynamic Wind): Remove doc for
scm_t_dynwind_flags, and change uses of scm_t_dynwind_flags to int.
Mention possible flags inside doc for scm_dynwind_begin instead.
Similarly for scm_t_wind_flags. In this case the doc on possible
flags is moved to after the four functions that can use them.
* libguile/dynwind.c (scm_dynwind_begin): Change scm_t_dynwind_flags
to int.
(scm_dynwind_unwind_handler): Change scm_t_wind_flags to int.
(scm_dynwind_rewind_handler): Ditto.
(scm_dynwind_unwind_handler_with_scm): Ditto.
(scm_dynwind_rewind_handler_with_scm): Ditto.
* libguile/dynwind.h (SCM_F_DYNWIND_REWINDABLE): Use #define instead
of typedef enum.
(SCM_F_WIND_EXPLICITLY): Ditto.
(function declarations): Change scm_t_dynwind_flags and
scm_t_wind_flags to int.
* test-suite/standalone/test-unwind.c (check_cont_body): Change
scm_t_dynwind_flags to int.
This fixes a bug introduced in e20d7001c3
and reported by Neil.
* libguile/eval.i.c (CEVAL)[DEVAL]: Don't duplicate ARG1 in
`debug.info->a.args' for gsubr stack frames.
(scm_apply): Likewise.
* test-suite/tests/eval.test ("stacks")["arguments of a gsubr stack
frame"]: New test.
Thanks to Scott McPeak for reporting this and providing a patch.
* libguile/socket.c (scm_to_ipv6): When address is the wrong type,
provide more information in the exception message.
(scm_to_sockaddr): scm_to_ipv6 expects just an address, not the
whole vector.
* test-suite/tests/socket.test ("AF_INET6/SOCK_STREAM"): New set of
tests.
The crux of this problem was that the thread doing a throw, and so
checking scm_i_critical_section_level, was different from the thread
that was in a critical section.
* libguile/async.h (scm_i_critical_section_level): Removed, replaced
by per-thread critical_section_level.
(SCM_CRITICAL_SECTION_START, SCM_CRITICAL_SECTION_END): Use
per-thread critical_section_level.
* libguile/continuations.c (scm_dynthrow): Check per-thread
critical_section_level.
* libguile/threads.c (guilify_self_1): Init per-thread
critical_section_level.
(scm_i_critical_section_level): Removed.
* libguile/threads.h (scm_i_thread): New critical_section_level field.
* libguile/throw.c (scm_ithrow): Check per-thread critical_section_level.
Strings returned by nl-langinfo were always being encoded using the
global setlocale and not the scm_t_locale passed to the function.
* libguile/i18n.c (scm_nl_langinfo): fix encoding of strings
Libunistring uses a function uc_locale_language to extract the
current language from the locale information. It does this by calling
setlocale. This makes incompatible with Guile functions that use the
locale_t thread-specific locale API, because the values returned by the
call to setlocale ignore the locale set by uselocale.
As a workaround, this patch extracts the language from the locale_t
structure's __names field.
A more complete solution is needed. Perhaps that solution would test
that the __names field exists in the configure step and revert to
!USE_GNU_LOCALE_API in that case.
* libguile/i18n.c (locale_language): new function that performs the
same job as uc_locale_language but is compatible with uselocale
(u32_locale_casecoll, u32_locale_tocase): replace uc_locale_language
with locale_language
* libguile/i18n.c (u32_locale_tocase, scm_char_locale_downcase)
(scm_char_locale_upcase, scm_string_locale_downcase)
(scm_string_locale_upcase): use the scm_t_uint32 type for buffers
that are used primarily for libunistring and not for Guile strings.
* libguile/generalized-arrays.c (scm_array_p_2): New, formerly
`scm_array_p ()'.
(scm_array_p): Add second argument, for compatibility with 1.8 and
earlier and to match what the doc says and what `SCM_VALIDATE_ARRAY'
expects.
* libguile/generalized-arrays.h (scm_array_p_2): New.
(scm_array_p): Adjust.
* libguile/i18n.c (u32_locale_tocase): new helper function
(scm_char_locale_downcase, scm_char_locale_upcase): add capability
for UCS-4 characters.
(scm_string_locale_downcase, scm_string_locale_upcase): modify for
the new scheme strings.
* module/system/base/compile.scm (make-compilation-module,
language-default-environment): New procedures.
(read-and-compile, compile): Have ENV default to
`(language-default-environment from)'.
(compile-and-load): Compile in `(current-module)'.
* module/system/repl/common.scm (repl-compile): Explicitly compile in
the current module so that macro definitions are visible.
* libguile/load.c (kw_env): New variable.
(do_try_autocompile): Call `compile-file' with `#:env (current-module)'.
* test-suite/tests/compiler.test ("psyntax")["compile uses a fresh module by
default", "compile-time definitions are isolated"]: New tests.
["compile in current module"]: Specify `#:env (current-module)'.
["redefinition"]: Adjust.
* test-suite/tests/bytevectors.test (c&e): Explicitly compile in the
current module so that its imports are visible.
* libguile/uniform.h (scm_array_handle_uniform_element_bit_size): New
public accessor.
* libguile/uniform.c (scm_array_handle_uniform_element_size): Better
errors in non-byte-aligned arrays.
(scm_uniform_vector_element_type, scm_uniform_vector_element_size)
(scm_c_uniform_vector_ref, scm_c_uniform_vector_set_x):
(scm_uniform_vector_to_list): Don't require byte-aligned access.
* libguile/bytevectors.c (scm_uniform_array_to_bytevector):
* libguile/arrays.c (scm_from_contiguous_typed_array): Fix for
uniform arrays whose element size is not a multiple of the byte size.
* libguile/i18n.c (RUN_IN_LOCALE_SECTION) [USE_GNU_LOCALE_API]: new
macro for this define. Should work the same as the !USE_GNU_LOCAL_API
version
(SCM_STRING_TO_U32_BUF, SCM_U32_BUF_FREE): helper macros
(compare_u32_strings, compare_u32_strings_ci): new functions that
compare scheme strings in the presence of the locale
(u32_locale_casecoll): new helper function for compare_u32_strings_ci
(compare_strings, compare_strings_ci): removed and replaced by
compare_u32_strings and compare_u32_strings_ci
(scm_string_locale_lt, scm_string_locale_gt, scm_string_locale_ci_lt)
(scm_string_locale_ci_gt, scm_string_locale_ci_eq): modified to
use new compare_u32_strings and compare_u32_strings_ci comparison
functions
(scm_char_locale_lt, scm_char_locale_gt, scm_char_locale_ci_lt)
(scm_char_locale_ci_gt, scm_char_locale_ci_eq): modified to use
new compare_u32_strings and compare_u32_strings_ci comparison
functions
From the time when a #<program> was a SMOB, really_make_boot_program
in vm.c was still using SCM_SET_SMOB_FLAGS to set the
SCM_F_PROGRAM_IS_BOOT flag - which meant that it was setting flag
1<<32 :-) which obviously was then missed by the SCM_PROGRAM_IS_BOOT
calls in stacks.c.
* libguile/programs.h (SCM_F_PROGRAM_IS_BOOT): Use a less significant
bit for this flag, now that programs use a tc7 type.
* libguile/vm.c (really_make_boot_program): Don't use
SCM_SET_SMOB_FLAGS, now that programs aren't SMOBs.
* libguile/stacks.c (stack_depth): Decrement depth count for a
macro-expansion frame. (Missing this decrement will cause a "stack
count incorrect" warning but is actually benign, as it only means
that scm_make_stack allocates a bit more space for the stack than it
really needs.)
(read_frames): Increment count of remaining unused frames, when
cutting out a macro transformer application frame. (Missing this
increment could cause the "stack count incorrect" warning, and could
make read_frames think it's filled up all the available stack frames
when there's actually still one frame unused; this wasn't benign,
because it could cause information to be missing from a stack
trace.)
* libguile/stacks.c (stack_depth): Remove unhittable else branch
apparently catering to a kind of frame that isn't an eval, apply or
void frame - in fact there are no other kinds of frames.
* libguile/Makefile.am (libpath.h): Fix pkgdatadir, pkglibdir, and
pkgincludedir entries. Add a new define, SCM_LIB_DIR.
* libguile/dynl.c (sysdep_dynl_init): Add the libdir to the libltdl
search path. Should fix
http://thread.gmane.org/gmane.lisp.guile.bugs/4289/focus=4296 -- that
is, it should allow guile to be invoked from whereever it is
installed, without munging LTDL_LIBRARY_PATH or the like variables.
There is a trick though -- during the build, we don't want to be
looking in the $libdir for loadable modules. So as with
GUILE_SYSTEM_PATH, we have GUILE_SYSTEM_LTDL_PATH.
* meta/uninstalled-env.in: Set GUILE_SYSTEM_LTDL_PATH to "" when
building.
This should also fix "condition is always true" warnings.
* libguile/vm-i-scheme.c (bv_u32_native_ref, bv_s32_native_ref,
bv_u32_native_set, bv_s32_native_set): Conditionalize use of
`BV_FIXABLE_INT_REF' vs. `BV_INT_REF' based on `SIZEOF_VOID_P'.
* libguile/threads.c (do_thread_exit): Deal with `t->mutexes' being a
weak list.
(fat_mutex_lock): Use weak-car pairs when building up `t->mutexes'.
* test-suite/tests/threads.test ("mutex-ownership")["mutex with owner
not retained (bug #27450)"]: New test.
* libguile/threads.c (scm_t_guile_ticket): Remove type.
(resume, scm_enter_guile, suspend, scm_leave_guile): Remove.
(scm_i_init_thread_for_guile): Set `t->top' to NULL, which has the
same effect as calling `scm_enter_guile ()'.
(scm_leave_guile_cleanup, scm_i_with_guile_and_parent): Remove
`scm_leave_guile ()' call.
