If FLAT_BIGNUMS is defined, Lisp_Bignums are a vector of GMP limbs
instead of a mpz_t (i.e. pointer to limb array).
The primary difficulty was the changed return type of the bignum_val
function. The idiom "*bignum_val (foo)" was used all over the place.
For the flat representation, it seems impossible to define bignum_val
such that it returns a pointer. I changed the return type to a struct,
but that made it necessary to change the idiom to "bignum_val (foo).z".
* src/bignum.h (FLAT_BIGNUMS): Define if HAVE_MPS.
(struct Lisp_Bignums): Define alternative representation.
(struct bignum_val): New helper struct.
(bignum_val): Change return type and define different version for flat
bignums.
(bignum_integer): Define a separate version for flat bignums.
(BIGNUM_VAL): New macro.
(init_gmp_memory_functions): New. Define the prototype here for a
function that is implemented in alloc.c.
(mpz_set_integer): Update for changed xbignum_val.
* src/bignum.c (init_bignum): Call init_gmp_memory_functions.
(make_bignum_from_mpz): New helper.
(make_bignum_bits, make_bignum_str): Adapted for flat bignums.
(bignum_to_double, bignum_to_uintmax, bignum_bufsize, bignum_to_c_string)
(get_random_bignum): Adapt to changed bignum_val.
* src/alloc.c (xmalloc_for_gmp): New.
(xrealloc_for_gmp, xfree_for_gmp): Moved here from bignum.c. Call
tally_consing.
(init_gmp_memory_functions): New.
(cleanup_vector): Adapt to flat bignums.
* src/data.c (Fnatnump, arithcompare, Fminus, integer_remainder)
(Flogcount, Fash, expt_integer, Fadd1, Fsub1, Flognot): Adapt to changed
bignum_val.
* src/emacs-module.c (module_extract_big_integer): Adapt to changed
bignum_val.
* src/floatfns.c (Fabs, Flogb): Adapt to changed bignum_val.
* src/fns.c (Frandom, Ftake, Fntake, Fnthcdr, Fmemql, Feql)
(internal_equal_1, value_cmp, sxhash_bignum): Adapt to changed
bignum_val.
* src/igc.c (finalize_bignum, finalize_vector, maybe_finalize): Adapt to
flat bignums.
* src/pdumper.c (dump_bignum): Define separate version for flat bignums.
(dump_cold_bignum, dump_do_dump_relocation): Adapt to changed
bignum_val.
* src/timefns.c (ztrillion, ticks_hz_to_timespec, ticks_hz_hz_ticks)
(lispint_arith): Adapt to changed bignum_val.
Problem for mpz_get_d_rounded reported by Helmut Eller in:
https://lists.gnu.org/r/emacs-devel/2025-11/msg00795.html
* lib-src/make-docfile.c (DEFUN_pure): New constant.
(write_globals, scan_c_stream): Support "attributes: pure".
* src/bignum.h (mpz_get_d_rounded):
* src/data.c (Fsymbolp, Fmodule_function_p, Fintegerp, Fnumberp):
* src/lisp.h (bignum_to_double, bignum_to_intmax)
(bignum_to_uintmax, bignum_bufsize):
Now pure, not const, since they depend on current state.
For example, Fsymbolp now inspects symbols_with_pos_enabled,
and the bignum functions inspect bignum contents in memory.
* src/data.c (Feq):
* src/xfaces.c (Fface_attribute_relative_p):
No longer const, since they might abort when debugging.
* src/pdumper.h (pdumper_object_p, pdumper_cold_object_p)
(pdumper_find_object_type, pdumper_object_p_precise):
These are not const functions. But there is no need to declare
them to be pure, either, as they’re inline so the compiler can
figure it out.
Only allow string mutation that is certain not to require string data to
be resized and reallocated: writing bytes into a unibyte string, and
changing ASCII to ASCII in a multibyte string.
This ensures that mutation will never transform a unibyte string to
multibyte, that the size of a string in bytes never changes, and that
the byte offsets of characters remain the same. Most importantly, it
removes a long-standing obstacle to reform of string representation and
allow for future performance improvements.
* src/data.c (Faset): Disallow resizing string mutation.
* src/fns.c (clear_string_char_byte_cache):
* src/alloc.c (resize_string_data): Remove.
* test/src/data-tests.el (data-aset-string): New test.
* test/lisp/subr-tests.el (subr--subst-char-in-string):
Skip error cases.
* test/src/alloc-tests.el (aset-nbytes-change):
Remove test that is no longer relevant.
* doc/lispref/strings.texi (Modifying Strings):
* doc/lispref/sequences.texi (Array Functions):
* doc/lispref/text.texi (Substitution): Update manual.
* etc/NEWS: Announce.
* lisp/emacs-lisp/comp.el (comp--native-compile): Avoid `%s`
on objects that may not have names.
* src/data.c (Ffset): Don't compute trampoline for no-op `fset`.
1. Prefer "native-compiled" to "native compiled".
The adjective "native-compiled" with the hyphen is generally more
consistent with the typical pattern in English, especially when the
compound modifies a noun (e.g., "native-compiled code").
2. Prefer "natively compiled" to "natively-compiled".
The adverb "natively" modifies "compiled", and it is standard not
to hyphenate an adverb + adjective combination when the adverb ends
in -ly (e.g., "code that is natively compiled").
For example, note that we say "high-speed internet" but "highly
performant code".
* Makefile.in (dest):
* configure.ac (HAVE_NATIVE_COMP):
* doc/emacs/building.texi (Lisp Libraries):
* doc/lispref/compile.texi (Native Compilation)
(Native-Compilation Functions, Native-Compilation Variables):
* doc/lispref/functions.texi (What Is a Function, Declare Form):
* doc/lispref/loading.texi (How Programs Do Loading, Library Search):
* etc/NEWS:
* etc/NEWS.28:
* etc/NEWS.29:
* etc/NEWS.30:
* lisp/emacs-lisp/comp-common.el (native-comp-never-optimize-functions)
(comp-function-type-spec):
* lisp/emacs-lisp/comp-cstr.el:
* lisp/subr.el (locate-eln-file):
* src/comp.c (SETJMP_NAME, syms_of_comp):
* src/data.c (Fsubrp, Fnative_comp_function_p, Fsubr_native_lambda_list):
* src/lread.c (Fload):
* src/pdumper.c (dump_do_dump_relocation):
* test/src/comp-tests.el (lambda-return2): Avoid grammatically incorrect
variations on "natively compiled" and "native-compiled". (Bug#56727)
The case_Lisp_Int macro was originally introduced with different
definitions depending on USE_2_TAGS_FOR_INTS. However, since commit
2b57012478, we have assumed that USE_2_TAGS_FOR_INTS is always
defined, and the macro has only a single definition. As a result, the
macro is now unnecessary, and replacing it with standard C case labels
improves readability and understanding.
* src/lisp.h (case_Lisp_Int): Delete macro.
* src/alloc.c (process_mark_stack, survives_gc_p):
* src/data.c (Fcl_type_of):
* src/fns.c (value_cmp, sxhash_obj):
* src/pdumper.c (dump_object):
* src/print.c (print_object):
* src/xfaces.c (face_attr_equal_p): Remove uses of above macro.
Since the introduction of the 'calln' macro, the 'call1', 'call2', ...,
'call8' macros are just aliases for the former. This is slightly
misleading and potentially unhelpful. The number of arguments N can
also easily go out-of-synch with the used alias callN. There is no
reason not to replace these aliases with using 'calln' directly.
To reduce the risk for mistakes, the tool Coccinelle was used to make
these changes. See <https://coccinelle.gitlabpages.inria.fr/website/>.
* src/alloc.c, src/androidvfs.c, src/androidfns.c, src/buffer.c:
* src/callint.c, src/callproc.c, src/casefiddle.c, src/charset.c:
* src/chartab.c, src/cmds.c, src/coding.c, src/composite.c:
* src/data.c, src/dbusbind.c, src/dired.c, src/doc.c:
* src/emacs.c, src/eval.c, src/fileio.c, src/filelock.c:
* src/fns.c, src/frame.c, src/gtkutil.c, src/haikufns.c:
* src/haikumenu.c, src/image.c, src/insdel.c, src/intervals.c:
* src/keyboard.c, src/keymap.c, src/lisp.h, src/lread.c:
* src/minibuf.c, src/nsfns.m, src/nsselect.m, src/pgtkfns.c:
* src/pgtkselect.c, src/print.c, src/process.c, src/sort.c:
* src/syntax.c, src/textconv.c, src/textprop.c, src/undo.c:
* src/w32fns.c, src/window.c, src/xfaces.c, src/xfns.c:
* src/xmenu.c, src/xselect.c, src/xterm.c:
Replace all uses of 'call1', 'call2', ..., 'call8' with 'calln'.
* doc/lispref/functions.texi (Function Cells): Avoid talking
about the function cell being void.
* src/data.c (Fboundp, Ffmakunbound, Fsymbol_function):
Don't suggest that "void" can be considered as a kind of value.
* doc/lispref/functions.texi (Function Cells):
* src/data.c (Ffboundp, Ffmakunbound, Fsymbol_function): Update
documentation to the changes of how void functions are represented
since Emacs 24.5. (Bug#73886)
* src/data.c (coerce_marker, not_number_or_marker): New.
(arithcompare): Don't use NUMBERP to check types up-front since we
are going to branch on types anyway.
This makes comparison functions (=, /=, <, <=, >, >=, min, max) quite
a bit faster (10-20 %). Bytecode ops on fixnums are not affected,
nor is `value<`.
* src/data.c (arithcompare): Simplify the code to reduce the number of
branches. Remove the comparison code argument; instead, return the
relation encoded as bits, which can be tested cheaply. All callers
adapted.
* src/lisp.h (enum Arith_Comparison): Remove.
(Cmp_Bit_*, cmp_bits_t): New.
* src/data.c (set_internal): Signal error if a BLV with a
redirect or a forwarded symbol is being unbound.
* test/src/data-tests.el (binding-test-makunbound-built-in): New
test.
* src/lisp.h (struct Lisp_Fwd): With the predicate enum, we can now pack
the offset and the predicate into a one-word struct.
(XBUFFER_OFFSET): Use the new field name.
* src/buffer.c (DEFVAR_PER_BUFFER): Create the one-word struct.
* src/data.c (store_symval_forwarding): Use the new field name.
Using an enum instead of a symbol makes it obvious that this field is
of no concern to the GC.
* src/lisp.h (enum Lisp_Fwd_Predicate): New.
(struct Lisp_Fwd): Use it instead of a symbol.
* src/buffer.c (DEFVAR_PER_BUFFER): Create the necessary enum constant
instead of a symbol.
* src/data.c (check_fwd_predicate, check_choice): New helpers.
(store_symval_forwarding): Use it.
* src/lisp.h (struct Lisp_Buffer_Objfwd): Deleted.
(struct Lisp_Fwd): Add the fields bufoffset and bufpredicate.
Make the type a 1-byte bitfield so that the entire struct still
fits in two words.
(XBUFFER_OFFSET): Renamed from XBUFFER_OBJFWD.
* src/buffer.c (DEFVAR_PER_BUFFER, defvar_per_buffer)
(buffer_local_value): Update accordingly.
* src/data.c (do_symval_forwarding, store_symval_forwarding)
(set_internal, default_value, set_default_internal)
(Fmake_local_variable, Fkill_local_variable, Flocal_variable_): Use
XBUFFER_OFFSET.
* src/lisp.h (struct Lisp_Kboard_Objfwd): Deleted ...
(struct Lisp_Fwd): ... replaced with field kbdoffset.
(DEFVAR_KBOARD): Use new field.
* src/data.c (XKBOARD_OFFSET): Renamed from XKBOARD_OBJFWD.
(do_symval_forwarding, store_symval_forwarding
(set_default_internal): Use it .
