* libguile/r6rs-ports.c (bip_seek): Fix to allow seeking to end of port.
* test-suite/tests/r6rs-ports.test ("bytevector input port can seek to
very end"): Add tests.
* module/rnrs/base.scm (log): now takes a base argument, using the
change of base formula for logs.
* test-suite/tests/r6rs-base.test ("log (2nd arg)"): Add test cases.
* libguile/print.c (symbol_has_extended_read_syntax): Use a more
general, unicode-appropriate algorithm. Hopefully doesn't cause
any current #{}# cases to be unescaped.
(print_extended_symbol): Use more appropriate unicode algorithm, and
emit unicode hex escapes instead of our own lame escapes.
* test-suite/tests/symbols.test: Add tests.
* libguile/read.c (scm_read_extended_symbol): Interpret '\' as an escape
character. Due to some historical oddities we have to support '\'
before any character, but since we never emitted '\' in front of
"normal" characters like 'x' we can interpret "\x..;" to be an R6RS
hex escape.
* test-suite/tests/reader.test ("#{}#"): Add tests.
* libguile/print.c (symbol_has_extended_read_syntax)
(print_normal_symbol, print_extended_symbol, scm_i_print_symbol_name):
Factor scm_i_print_symbol_name into separate routines. Add comments.
There are a number of bugs here.
* module/system/repl/server.scm (run-server): Ignore SIGPIPE when we run
a server, as otherwise a rudely disconnected client could cause the
server to quit. Thanks to John Proctor for the report, and Detlev
Zundel for the debugging.
* libguile/numbers.c (scm_exact_integer_sqrt): New C procedure to
compute exact integer square root and remainder.
(scm_i_exact_integer_sqrt): New Scheme procedure `exact-integer-sqrt'
from the R6RS, imported into core guile.
* libguile/numbers.h: Add prototypes.
* module/rnrs/base.scm: Remove broken stub implementation, which would
fail badly when applied to large integers.
* doc/ref/api-data.texi: Add documentation.
* doc/ref/r6rs.texi: Change documentation for `exact-integer-sqrt' to a
stub that xrefs the core docs, as is done for other operations
available in core.
* test-suite/tests/numbers.test: Add tests.
* NEWS: Add news entries.
* module/ice-9/boot-9.scm (define-inlineable): Moved here from SRFI-9.
* module/srfi/srfi-9 (define-inlinable): Removed here.
* doc/ref/api-procedures.texi (Inlinable Procedures): Add subsection
about `define-inlinable'.
* libguile/numbers.c (mem2decimal_from_point): Use scm_divide instead of
scm_divide2real when applying a negative exponent, to preserve
exactness in case the "#e" forced exactness specifier is present.
This fixes a bug where numeric literals such as "#e1e-5" yielded
incorrect fractions.
* doc/ref/api-data.texi (Arithmetic): `floor-remainder' is equivalent to
the R5RS `modulo' when the arguments are integers. Previously,
equivalence was claimed only for exact integers. Similarly for
`truncate-quotient' and `truncate-remainder' compared with the R5RS
`quotient' and `remainder'.
* libguile/read.c (scm_read_sharp): Move the "#c..." case outside of
#if SCM_ENABLE_DEPRECATED, and to the same section which handles
"#s...", "#u..." and "#f...".
Thanks to Andreas Rottmann <a.rottmann@gmx.at> for the bug report.
* module/rnrs/arithmetic/fixnums.scm (assert-fixnum): Is now a
macro.
(assert-fixnums): New procedure checking a the elements of a list
for fixnum-ness. All callers applying `assert-fixnum' to a list
now changed to use this procedure.
* module/rnrs/arithmetic/fixnums.scm (define-fxop*): New for defining
n-ary inlinable special-casing the binary case using `case-lambda'.
All applicable procedures redefined using this macro.
* module/rnrs/arithmetic/fixnums.scm: Alias all predicates to
their non-fixnum counterparts.
* configure.ac: Check for pthread_sigmask.
* libguile/gen-scmconfig.c: Create SCM_HAVE_GC_PTHREAD_SIGMASK.
* libguile/pthread-threads.h (scm_i_pthread_sigmask): Only redirect to
GC_pthread_sigmask if GC_pthread_sigmask is present.
* test-suite/tests/foreign.test ("pointer<->string"): Add test cases.
* libguile/foreign.c (scm_string_to_pointer, scm_pointer_to_string): Add
optional encoding, and in the pointer->string case, length arguments.
* libguile/foreign.h: Update prototypes of internal functions.
Shouldn't affect ABI as they are internal.
* doc/ref/api-foreign.texi (Void Pointers and Byte Access): Update
docs.
* configure.ac: Check for sys/mman.h.
* libguile/objcodes.c (verify_cookie): Factor cookie verification out to
a helper function.
(make_objcode_from_file): Rename from make_objcode_by_mmap. If mmap
is unavailable, just read(2) to a bytevector.
* libguile/objcodes.c (make_objcode_by_mmap): MAP_PRIVATE, not
MAP_SHARED -- we don't need to update the underlying file, nor do we
need to see updates.
* libguile/strings.c (scm_encoding_error, scm_decoding_error): Use
scm_from_latin1_string for the subr and message args, as these are
internal functions, and we know their callers.
* meta/uninstalled-env.in: Our code that checked for paths already being
in the load path was not working for the last entry in the load path.
This caused the last entry to be re-added to the beginning, which also
caused relative filename canonicalization to prepend "module/" to
everything.
Terrible.
* libguile/read.c (scm_i_scan_for_encoding): Fix for coding on first
line #! and for !# immediately following the coding.
* test-suite/Makefile.am:
* test-suite/tests/coding.test: Add tests.
* module/ice-9/psyntax.scm (fx+, fx-, fx=, fx<): Given our lame lack of
an inliner, inline these manually with identifier syntax.
* module/ice-9/psyntax-pp.scm: Regenerate.
* module/ice-9/psyntax.scm (with-syntax): Allow definitions in the body,
as seems to be suggested by the R6RS.
* test-suite/tests/syncase.test ("with-syntax"): Add test.
* module/ice-9/psyntax-pp.scm: Regenerate.
* libguile/dynl.c (sysdep_dynl_value): Failure to find a symbol is not
an error, so raise our own, more appropriate error.
* test-suite/tests/foreign.test ("dynamic-pointer"): Add a test.
* libguile/continuations.c (scm_handler, c_handler)
(scm_c_with_continuation_barrier, scm_with_continuation_barrier):
Instead of calling scm_handle_by_message_noexit in the pre-unwind
handler, roll our own exception printing in the pre-unwind, and do to
exit()-on-quit in the post-unwind handler. This lets the stack unwind
at exit-time so that pending dynwinds run.
* test-suite/tests/continuations.test ("continuations"): Add a test.
GC_unregister_my_thread is only supposed to be called from
a thread other than the main thread, so, it should never be
called when the system is compiled with null threads.
* libguile/threads.c (on_thread_exit)[SCM_USE_NULL_THREADS]: don't
call GC_unregister_my_thread
This is not present in earlier versions of BDW-GC
* configure.ac: check for GC_allow_register_threads
* libguile/threads.c (scm_i_init_thread_for_guile): Only call
GC_allow_register_threads if it is present.
* libguile/control.h:
* libguile/control.c (scm_i_prompt_pop_abort_args_x): Take a VM instead
of a prompt, given that it's the VM's registers that record the abort
arguments, not the prompt registers (which actually point right below
the abort values).
* libguile/eval.c (eval):
* libguile/throw.c (pre_init_catch): Pass the vm instead of a prompt.
* configure.ac: Check for GC_pthread_exit and GC_pthread_cancel.
* libguile/gen-scmconfig.c: Write HAVE_GC_PTHREAD_CANCEL and
HAVE_GC_PTHREAD_EXIT into scmconfig.h.
* libguile/pthread-threads.h (scm_i_pthread_exit, scm_i_pthread_cancel):
Only redefine to their GC_pthread_* variants if we have those
functions, which is not the case in libgc < 7.2.
* module/language/objcode/spec.scm (decompile-value): Don't assume that
`error' will handle format strings appropriately.
* module/system/repl/command.scm (disassemble): A more human error when
you disassemble a non-procedure.
Bug reported by Andrew Horton.
* doc/ref/scheme-scripts.texi (Invoking Guile): Document -q.
* doc/ref/scheme-using.texi (Init File): New section, on .guile.
(Readline): Link to Init File.
(System Commands): Document the various REPL options, and
repl-default-option-set!.
* module/language/tree-il/compile-glil.scm (flatten): When compiling a
<prompt> in push context with an RA, after the body returns normally,
jump to that RA instead of to our POST label (which in that case does
not need to be emitted). Fixes a tail <prompt> in a push <fix>.
* test-suite/tests/control.test ("prompt in different contexts"): Add
more test cases.
* doc/ref/tools.texi (Executable Modules): Say "guile-tools modules"
instead of "executable modules". Remove obsolete statements about
not ending in .scm, being executable, and beginning with shell
script invocation sequence.
* module/scripts/README: Ditto.
It's just one variable definition, and in my opinion it confuses,
rather than helps, the overall build picture to have two names
(preinstguile and meta/guile) for the same thing.
* am/Makefile.am (am_frags): Remove pre-inst-guile.
* am/pre-inst-guile: Deleted.
* doc/ref/Makefile.am: Don't include am/pre-inst-guile.
($(snarf_doc).am, $(snarf_doc).texi): Expand $(preinstguile).
* module/Makefile.am (ice-9/psyntax-pp.scm.gen): Don't include
am/pre-inst-guile.
(ice-9/psyntax-pp.scm.gen): Expand $(preinstguile).
