compiled closures. Include "libguile/smob.h".
(f_eval_closure): Removed.
(scm_eval_closure_tag): New variable.
(scm_eval_closure_lookup): Renamed from eval_closure.
This function now takes a smob instead of a compiled closure.
(scm_standard_eval_closure): Create a smob instead of a compiled
closure.
(scm_init_modules): Initialize the eval closure type as a smob.
* modules.h (SCM_EVAL_CLOSURE_P): New macro.
(scm_eval_closure_tag, scm_eval_closure_lookup): Declare.
* symbols.c: Include "libguile/smob.h".
(scm_sym2vcell): Call scm_eval_closure_lookup directly if THUNK
is an eval closure.
suitable handler has been found. That way, we don't have to rely
on the user establishing a catch-all, which might be difficult for
him if he is using scm_init_guile instead of scm_boot_guile.
initializing Guile that does return to the caller.
(scm_boot_guile_1): Use scm_init_guile_1 to initialize Guile.
Do not establish a catch-all, this is no longer needed.
and give explicit dependencies for .x files that depend on
generated files. This allows parallel builds. Thanks to Matthias
Koeppe!
* Makefile.am: Added gc_os_dep.c, properties.c, properties.x,
properties.h and properties.doc in the suitable places.
about any possible false positives.
(scm_debug_newcell): don't change cell type to `allocated'.
(scm_debug_newcell2): ditto.
(scm_gc_for_newcell): ditto.
(scm_gc_mark): remove the tc16_allocated case.
(scm_init_struct): First arg to scm_make_vtable_vtable should be a
string, not a symbol. (`make-vtable-vtable' needs to append this
string to another string and then pass it through
`make-struct-layout'.)
`symbol-append'.
(symbol-append): Map `symbol->string' on
args.
(obarray-symbol-append, obarray-gensym): Simply removed. I don't
think I'll announce this in NEWS even. One of the functions never
even worked... /mdj.
(find-and-link-dynamic-module, keyword->symbol): Use
`symbol->string'.
(try-module-autoload, process-define-module): Rewrote using R5RS
semantics.
