(scm_struct_i_free): New hidden struct slot. Holds destructor for
instances to this vtable.
(scm_struct_free_0): New destructor: Doesn't deallocate data.
(scm_struct_free_light): New destructor: Deallocates a light
struct (i.e. a struct without hidden slots).
(scm_struct_free_standard): New destructor: Deallocates standard
structs.
(scm_struct_free_entity): New destructor: Deallocates entity
structs.
(SCM_SET_VTABLE_DESTRUCTOR): New macro.
Changes to hidden slots:
(scm_struct_i_size): scm_struct_i_flags now shares space with
scm_struct_i_size which holds the size of light structs.
(scm_struct_i_n_words): This slot has changed meaning. Previously
it included hidden slots. Now it indicates visible slots.
(scm_alloc_struct): Clear flags.
(SCM_STRUCTF_MASK): 4 new flag positions added => 12 bits.
(struct_num, scm_struct_i_tag): Removed.
(scm_struct_vtable_tag): Base tag on the pointer to mallocated
memory.
(scm_struct_ihashq): Base hash value on pointer to struct handle.
(scm_struct_table): Weak key table with auxilliary information for
struct types. Currently used for names and wrapper classes.
(scm_struct_ihashq): Hash function for structs.
(scm_struct_create_handle): Get/create entry in scm_struct_table.
(scm_struct_vtable_name, scm_set_struct_vtable_name_x): Procedures
for accessing names of vtables. The record implementation in
boot-9.scm currently uses the setter to record the name of record
types. When the object system is initialized, it can use this
information to create wrapper classes with suitable names.
(scm_init_struct): Allocate scm_struct_table.
(scm_alloc_struct): Don't initialize scm_struct_i_tag here.
(struct tags are a finite resource and we might want to restrict
the use of tags to vtables only. E.g., Goops only uses tags for
classes.)
(scm_make_struct): Use scm_struct_entity_n_extra_words instead of
magic number 5.
(scm_struct_vtable_tag): Use scm_struct_i_tag instead of magic
number -1.
procedures if SCM_STRUCTF_ENTITY is set in vtable.
* struct.c, struct.h (scm_alloc_struct): Renamed from alloc_struct
and made global.
(scm_struct_init): Renamed from init_struct and made global.
* print.h (SCM_PRINT_STATE_P): Removed SCM_NIMP test. (NIMP
macros should by convention not test for NIMPness.)
(SCM_COERCE_OPORT): Adjust indentation.
* print.c (scm_valid_oport_value_p): Adjusted indentation; Added
SCM_NIMP test before SCM_PRINT_STATE_P.
* struct.c, struct.h, gc.c: Renamed:
scm_struct_i_layout --> scm_vtable_index_layout
scm_struct_i_vcell --> scm_vtable_index_vcell
scm_struct_i_vtable --> scm_vtable_index_vtable
scm_struct_i_printer --> scm_vtable_index_printer
scm_struct_i_vtable_offset --> scm_vtable_offset_user
* struct.c (scm_print_struct): Use new printer slot; Default
printing: Also output hex code of vtable so that type identity
will be indicated as well.
(scm_init_struct): Updated required_vtable_fields to "pruosrpw";
Removed struct_printer_var; Removed struct-vtable-offset;
(vtable-index-layout, vtable-index-vtable, vtable-index-printer,
vtable-offset-user): New constants.
* struct.h (scm_struct_i_vtable_offset): Bumped from 3 to 4.
(scm_struct_i_printer, SCM_STRUCT_PRINTER): New slot in vtables.
If this slot contains a procedure, use that to print structures of
the type represented by this vtable.
* print.c (scm_iprin1): Don't print arguments of macro
transformers. (They are always: exp env.); Bugfix: Unmemoize
transformer source with correct environment.
the Scheme variable *struct-printer*. This variable can be set by
Scheme code to override the printing of structures.
(scm_print_struct): If struct_printer is set, call it. If it is
not set, or returns #f, print the structure in the old fashion.
Include "eval.h" for scm_apply.
scm_struct_i_n_words to get the number of fields, not
-scm_struct_n_extra_words.
On the route to fancier struct printing:
* struct.c (scm_print_struct): New function to print a structure.
Include "genio.h" to support it. This function doesn't do
anything interesting right now, but I think it should be here
anyway.
* struct.h: Include "print.h" and add prototype for
scm_print_struct.
* print.c (scm_iprin1): Call scm_print_struct instead of trying to
print structures ourself.
variables aren't used uninitialized.
* error.h (scm_error, scm_syserror, scm_syserror_msg,
scm_sysmissing, scm_num_overflow, scm_out_of_range,
scm_wrong_num_args, scm_wrong_type_arg, scm_memory_error,
scm_misc_error): Tell GCC that these functions never return.
* struct.c (scm_struct_ref, scm_struct_set_x): If we can't figure
out the field type, call abort if SCM_ASSERT returns, to placate
the optimizer.
* stacks.c (scm_make_stack, scm_last_stack_frame): abort if
scm_wta ever returns. We can't handle this case anyway, and this
gives the optimizer more information.
* unif.c (scm_uniform_vector_ref, scm_array_set_x): Abort if
scm_wta ever returns.
In some cases, the code is fine, but GCC isn't smart enough to
figure that out; this usually happens when one variable is only
initialized and used when a particular condition holds true, and
we know that condition will never change within a given invocation
of the function. In this case, we simply initialize the variables
to placate the compiler, hopefully to a value which will cause a
crash if it is ever actually used.
* print.c (scm_iprin1): Initialize mw_pos.
* read.c (scm_lreadrecparen): Initialize tl2, ans2.
* throw.c (scm_ithrow): Initialize dynpair.
* unif.c (scm_uniform_vector_ref): Initialize cra.
* struct.c (init_struct): Initialize prot.
* mbstrings.c (scm_print_mb_symbol): Initialize mw_pos and inc.
required by the tagging system.
* struct.c (alloc_struct): New function.
(scm_make_struct, scm_make_vtable_vtable): Call it.
* struct.h (scm_struct_n_extra_words): Bump to 3.
(scm_struct_i_ptr): New "field".
* gc.c (scm_gc_sweep): When we need to free the data, use the
information stored by alloc_struct to find the beginning of the
block allocated to the structure, so we can free it.
scm_struct_set_x), struct.h, gc.c (scm_gc_mark): Completed Tom
Lord's implementation of structs, allowing for tail arrays as
described in the manual. Also fixed some bugs. (Both the interface
and the implementation should be improved.)
