This partly undoes my 2011-03-30 change, which replaced int with size_t.
Back then I didn't know that the Emacs coding style prefers signed int.
Also, in the meantime I found a few more instances where arguments
were being counted with int, which may truncate counts on 64-bit
machines, or EMACS_INT, which may be unnecessarily wide.
* lisp.h (struct Lisp_Subr.function.aMANY)
(DEFUN_ARGS_MANY, internal_condition_case_n, safe_call):
Arg counts are now ptrdiff_t, not size_t.
All variadic functions and their callers changed accordingly.
(struct gcpro.nvars): Now size_t, not size_t. All uses changed.
* bytecode.c (exec_byte_code): Check maxdepth for overflow,
to avoid potential buffer overrun. Don't assume arg counts fit in 'int'.
* callint.c (Fcall_interactively): Check arg count for overflow,
to avoid potential buffer overrun. Use signed char, not 'int',
for 'varies' array, so that we needn't bother to check its size
calculation for overflow.
* editfns.c (Fformat): Use ptrdiff_t, not EMACS_INT, to count args.
* eval.c (apply_lambda):
* fns.c (Fmapconcat): Use XFASTINT, not XINT, to get args length.
(struct textprop_rec.argnum): Now ptrdiff_t, not int. All uses changed.
(mapconcat): Use ptrdiff_t, not int and EMACS_INT, to count args.
GCC 4.6.0 optimizes based on type-based alias analysis. For
example, if b is of type struct buffer * and v of type struct
Lisp_Vector *, then gcc -O2 was incorrectly assuming that &b->size
!= &v->size, and therefore "v->size = 1; b->size = 2; return
v->size;" must therefore return 1. This assumption is incorrect
for Emacs, since it type-puns struct Lisp_Vector * with many other
types. To fix this problem, this patch adds a new type struct
vector_header that documents the constraints on layout of vectors
and pseudovectors, and helps optimizing compilers not get fooled
by Emacs's type punning. It also adds the macros XSETTYPED_PVECTYPE
XSETTYPED_PSEUDOVECTOR, TYPED_PSEUDOVECTORP, for similar reasons.
* lisp.h (XVECTOR_SIZE): New convenience macro. All previous uses of
XVECTOR (foo)->size replaced to use this macro, to avoid the hassle
of writing XVECTOR (foo)->header.size.
(XVECTOR_HEADER_SIZE): New macro, for use in XSETPSEUDOVECTOR.
(XSETTYPED_PVECTYPE): New macro, specifying the name of the size
member.
(XSETPVECTYPE): Rewrite in terms of new macro.
(XSETPVECTYPESIZE): New macro, specifying both type and size.
This is a bit clearer, and further avoids the possibility of
undesirable aliasing.
(XSETTYPED_PSEUDOVECTOR): New macro, specifying the size.
(XSETPSEUDOVECTOR): Rewrite in terms of XSETTYPED_PSEUDOVECTOR
and XVECTOR_HEADER_SIZE.
(XSETSUBR): Rewrite in terms of XSETTYPED_PSEUDOVECTOR and XSIZE,
since Lisp_Subr is a special case (no "next" field).
(ASIZE): Rewrite in terms of XVECTOR_SIZE.
(struct vector_header): New type.
(TYPED_PSEUDOVECTORP): New macro, also specifying the C type of the
object, to help avoid aliasing.
(PSEUDOVECTORP): Rewrite in terms of TYPED_PSEUDOVECTORP.
(SUBRP): Likewise, since Lisp_Subr is a special case.
* lisp.h (struct Lisp_Vector, struct Lisp_Char_Table):
(struct Lisp_Sub_Char_Table, struct Lisp_Bool_Vector):
(struct Lisp_Hash_Table): Combine first two members into a single
struct vector_header member. All uses of "size" and "next" members
changed to be "header.size" and "header.next".
* buffer.h (struct buffer): Likewise.
* font.h (struct font_spec, struct font_entity, struct font): Likewise.
* frame.h (struct frame): Likewise.
* process.h (struct Lisp_Process): Likewise.
* termhooks.h (struct terminal): Likewise.
* window.c (struct save_window_data, struct saved_window): Likewise.
* window.h (struct window): Likewise.
* alloc.c (allocate_buffer, Fmake_bool_vector, allocate_pseudovector):
Use XSETPVECTYPESIZE, not XSETPVECTYPE, to avoid aliasing problems.
* buffer.c (init_buffer_once): Likewise.
* lread.c (defsubr): Use XSETTYPED_PVECTYPE, since Lisp_Subr is a
special case.
* process.c (Fformat_network_address): Use local var for size,
for brevity.
This makes it easier for human readers (and static analyzers)
to see whether these variables are used from other modules.
* alloc.c, buffer.c, bytecode.c, callint.c, casetab.c, category.c:
* ccl.c, character.c, charset.c, cmds.c, coding.c, composite.c:
* data.c, dbusbind.c, dired.c, editfns.c, eval.c, fileio.c, fns.c:
* font.c, frame.c, fringe.c, ftfont.c, image.c, keyboard.c, keymap.c:
* lread.c, macros.c, minibuf.c, print.c, process.c, search.c:
* sound.c, syntax.c, textprop.c, window.c, xdisp.c, xfaces.c, xfns.c:
* xmenu.c, xselect.c:
Declare Q* vars static if they are not used in other modules.
* ccl.h, character.h, charset.h, coding.h, composite.h, font.h:
* frame.h, intervals.h, keyboard.h, lisp.h, process.h, syntax.h:
Remove decls of unexported vars.
* keyboard.h (EVENT_HEAD_UNMODIFIED): Remove now-unused macro.
* lisp/emacs-lisp/byte-opt.el (byte-compile-inline-expand): Make it work to
inline lexbind interpreted functions into lexbind code.
(bytedecomp-bytes): Not a dynamic var any more.
(disassemble-offset): Get the bytes via an argument instead.
(byte-decompile-bytecode-1): Use push.
* lisp/emacs-lisp/bytecomp.el: Remove the bytecomp- prefix now that we use
lexical-binding.
(byte-compile-outbuffer): Rename from bytecomp-outbuffer.
* lisp/emacs-lisp/cl-macs.el (load-time-value):
* lisp/emacs-lisp/cl.el (cl-compiling-file): Adjust to new name.
* lisp/emacs-lisp/pcase.el (pcase-mutually-exclusive-predicates):
Add byte-code-function-p.
(pcase--u1): Remove left-over code from early development.
Fix case of variable shadowing in guards and predicates.
(pcase--u1): Add a new `let' pattern.
* src/image.c (parse_image_spec): Use Ffunctionp.
* src/lisp.h: Declare Ffunctionp.
* src/bytecode.c (exec_byte_code): Remove old lexical binding slot handling
and replace it with the a integer args-desc handling.
* eval.c (funcall_lambda): Adjust arglist test accordingly.
* lisp/emacs-lisp/bytecomp.el (byte-compile-arglist-signature):
Handle integer arglist descriptor.
(byte-compile-make-args-desc): Make integer arglist descriptor.
(byte-compile-lambda): Use integer arglist descriptor to mark lexical
byte-coded functions instead of an extra slot.
* lisp/help-fns.el (help-add-fundoc-usage): Don't add a dummy doc.
(help-split-fundoc): Return a nil doc if there was no actual doc.
(help-function-arglist): Generate an arglist from an integer arg-desc.
* lisp/emacs-lisp/pcase.el (pcase--memoize): Rename from pcase-memoize;
Make only the key weak.
(pcase): Change the key used in the memoization table, so it does not
always get GC'd away.
* lisp/emacs-lisp/macroexp.el (macroexpand-all-1): Slight change to the
pcase pattern to generate slightly better code.
* lisp/emacs-lisp/cconv.el: Compute freevars in cconv-analyse.
(cconv-mutated, cconv-captured): Remove.
(cconv-captured+mutated, cconv-lambda-candidates): Don't give them
a global value.
(cconv-freevars-alist): New var.
(cconv-freevars): Remove.
(cconv--lookup-let): Remove.
(cconv-closure-convert-function): Extract from cconv-closure-convert-rec.
(cconv-closure-convert-rec): Adjust to above changes.
(fboundp): New function.
(cconv-analyse-function, form): Rewrite.
* lisp/emacs-lisp/bytecomp.el (byte-compile-initial-macro-environment):
Handle declare-function here.
(byte-compile-obsolete): Remove.
(byte-compile-arglist-warn): Check late defsubst here.
(byte-compile-file-form): Simplify.
(byte-compile-file-form-defsubst): Remove.
(byte-compile-macroexpand-declare-function): Rename from
byte-compile-declare-function, turn it into a macro-expander.
(byte-compile-normal-call): Check obsolescence.
(byte-compile-quote-form): Remove.
(byte-compile-defmacro): Revert to trunk's definition which seems to
work just as well and handles `declare'.
* lisp/emacs-lisp/byte-run.el (make-obsolete): Don't modify byte-compile.
* lisp/Makefile.in (BIG_STACK_DEPTH): Increase to 1200.
(compile-onefile): Pass $(BIG_STACK_OPTS) before "-l bytecomp".
* lisp/emacs-lisp/macroexp.el: Use lexbind.
(macroexpand-all-1): Check macro obsolescence.
* lisp/vc/diff-mode.el: Use lexbind.
* lisp/follow.el (follow-calc-win-end): Simplify.
* lisp/emacs-lisp/bytecomp.el (byte-compile-lapcode): Handle new form of
`byte-constant'.
(byte-compile-close-variables, displaying-byte-compile-warnings):
Add edebug spec.
(byte-compile-toplevel-file-form): New fun, split out of
byte-compile-file-form.
(byte-compile-from-buffer): Use it to avoid applying cconv
multiple times.
(byte-compile): Only strip `function' if it's present.
(byte-compile-lambda): Add `reserved-csts' argument.
Use new lexenv arg of byte-compile-top-level.
(byte-compile-reserved-constants): New var.
(byte-compile-constants-vector): Obey it.
(byte-compile-constants-vector): Handle new `byte-constant' form.
(byte-compile-top-level): Add args `lexenv' and `reserved-csts'.
(byte-compile-form): Don't check callargs here.
(byte-compile-normal-call): Do it here instead.
(byte-compile-push-unknown-constant)
(byte-compile-resolve-unknown-constant): Remove, unused.
(byte-compile-make-closure): Use `make-byte-code' rather than `curry',
putting the environment into the "constant" pool.
(byte-compile-get-closed-var): Use special byte-constant.
* lisp/emacs-lisp/byte-opt.el (byte-optimize-form-code-walker): Handle new
intermediate special form `internal-make-vector'.
(byte-optimize-lapcode): Handle new form of `byte-constant'.
* lisp/help-fns.el (describe-function-1): Don't handle funvecs.
* lisp/emacs-lisp/macroexp.el (macroexpand-all-1): Only convert quote to
function if the content is a lambda expression, not if it's a closure.
* emacs-lisp/eieio-come.el: Remove.
* lisp/emacs-lisp/eieio.el: Don't require eieio-comp.
(defmethod): Do a bit more work to find the body and wrap it into
a function before passing it to eieio-defmethod.
(eieio-defmethod): New arg `code' for it.
* lisp/emacs-lisp/debug.el (debugger-setup-buffer): Don't hide things in
debugger backtrace.
* lisp/emacs-lisp/cl-extra.el (cl-macroexpand-all): Use backquotes, and be
more careful when quoting a function value.
* lisp/emacs-lisp/cconv.el (cconv-freevars): Accept defvar/defconst.
(cconv-closure-convert-rec): Catch stray `internal-make-closure'.
* lisp/Makefile.in (COMPILE_FIRST): Compile pcase and cconv early.
* src/eval.c (Qcurry): Remove.
(funcall_funvec): Remove.
(funcall_lambda): Move new byte-code handling to reduce impact.
Treat all args as lexical in the case of lexbind.
(Fcurry): Remove.
* src/data.c (Qfunction_vector): Remove.
(Ffunvecp): Remove.
* src/lread.c (read1): Revert to calling make_byte_code here.
(read_vector): Don't call make_byte_code any more.
* src/lisp.h (enum pvec_type): Rename back to PVEC_COMPILED.
(XSETCOMPILED): Rename back from XSETFUNVEC.
(FUNVEC_SIZE): Remove.
(FUNVEC_COMPILED_TAG_P, FUNVEC_COMPILED_P): Remove.
(COMPILEDP): Rename back from FUNVECP.
* src/fns.c (Felt): Remove unexplained FUNVEC check.
* src/doc.c (Fdocumentation): Don't handle funvec.
* src/alloc.c (make_funvec, Ffunvec): Remove.
* doc/lispref/vol2.texi (Top):
* doc/lispref/vol1.texi (Top):
* doc/lispref/objects.texi (Programming Types, Funvec Type, Type Predicates):
* doc/lispref/functions.texi (Functions, What Is a Function, FunctionCurrying):
* doc/lispref/elisp.texi (Top): Remove mentions of funvec and curry.
* lisp/emacs-lisp/byte-opt.el (byte-compile-side-effect-and-error-free-ops):
Remove interactive-p.
(byte-optimize-lapcode): Update optimizations now that stack-refs are
relative to the top rather than to the bottom.
* lisp/emacs-lisp/bytecomp.el (byte-compile-lapcode):
Turn stack-ref-0 into dup.
(byte-compile-form): Don't indirect-function since it can signal
errors.
(byte-compile-stack-ref, byte-compile-stack-set): Adjust to stack-refs
being relative to top rather than to bottom in the byte-code.
(with-output-to-temp-buffer): Remove.
(byte-compile-with-output-to-temp-buffer): Remove.
* lisp/emacs-lisp/cconv.el: Use lexical-binding.
(cconv--lookup-let): Rename from cconv-lookup-let.
(cconv-closure-convert-rec): Fix handling of captured+mutated
arguments in defun/defmacro.
* lisp/emacs-lisp/eieio-comp.el (eieio-byte-compile-file-form-defmethod):
Rename from byte-compile-file-form-defmethod.
Don't byte-compile-lambda.
(eieio-byte-compile-defmethod-param-convert): Rename from
byte-compile-defmethod-param-convert.
* lisp/emacs-lisp/eieio.el (eieio-defgeneric-form-primary-only-one):
Call byte-compile rather than byte-compile-lambda.
* src/alloc.c (Fgarbage_collect): Don't mark the byte-stack redundantly.
* src/bytecode.c (exec_byte_code): Change stack_ref and stack_set to use
offsets relative to top rather than to bottom.
* lisp/subr.el (with-output-to-temp-buffer): New macro.
* lisp/simple.el (count-words-region): Don't use interactive-p.
* lisp/emacs-lisp/lisp-mode.el (save-window-excursion): Don't touch.
* lisp/emacs-lisp/cconv.el (cconv-closure-convert-rec, cconv-analyse-form):
Don't handle save-window-excursion any more.
* lisp/emacs-lisp/bytecomp.el (interactive-p, save-window-excursion):
Don't use the byte-code any more.
(byte-compile-form): Check macro expansion was done.
(byte-compile-save-window-excursion): Remove.
* lisp/emacs-lisp/byte-opt.el (byte-optimize-form-code-walker):
Ignore save-window-excursion. Don't macroepand any more.
* src/window.c (Fsave_window_excursion): Remove. Moved to Lisp.
(syms_of_window): Don't defsubr it.
* src/window.h (Fsave_window_excursion): Don't declare it.
* src/bytecode.c (exec_byte_code): Inline Fsave_window_excursion.
* src/alloc.c (Fgarbage_collect): When using stack scanning, don't
redundantly scan byte-code stacks, catchlist, and handlerlist.
* src/bytecode.c (BYTE_MAINTAIN_TOP): New macros.
(struct byte_stack): Only define `top' and `bottom' if used.
(mark_byte_stack): Only define if used.
(BEFORE_POTENTIAL_GC, AFTER_POTENTIAL_GC): Nullify if BYTE_MAINTAIN_TOP
is not set.
(Fbyte_code): Don't set `bottom' unless BYTE_MAINTAIN_TOP is set.
* src/lisp.h (BYTE_MARK_STACK): New macro.
(mark_byte_stack): Only declare if BYTE_MARK_STACK is set.
* src/term.c (OUTPUT_IF): Use OUTPUT.
* lisp/files.el (lexical-binding): Add a safe-local-variable property.
* lisp/emacs-lisp/byte-opt.el (byte-inline-lapcode): Check how many elements
are added to the stack.
(byte-compile-splice-in-already-compiled-code): Don't touch lexical nor
byte-compile-depth now that byte-inline-lapcode does it for us.
(byte-compile-inline-expand): Don't inline dynbind byte code into
lexbind code, since it has to be done differently.
* lisp/emacs-lisp/bytecomp.el (byte-compile-arglist-warn):
Correctly extract arglist from `closure's.
(byte-compile-cl-warn): Compiler-macros are run earlier now.
(byte-compile-top-level): Bind byte-compile-lexical-environment to nil,
except for lambdas.
(byte-compile-form): Don't run the compiler-macro expander here.
(byte-compile-let): Merge with byte-compile-let*.
Don't preserve-body-value if the body's value was discarded.
* lisp/emacs-lisp/cconv.el (cconv--set-diff, cconv--set-diff-map)
(cconv--map-diff, cconv--map-diff-elem, cconv--map-diff-set): New funs.
(cconv--env-var): New constant.
(cconv-closure-convert-rec): Use it and use them. Fix a typo that
ended up forgetting to remove entries from lmenvs in `let'.
For `lambda' use the outer `fvrs' when building the closure and don't
forget to remove `vars' from the `emvrs' and `lmenvs' of the body.
* lisp/emacs-lisp/cl-macs.el (cl-byte-compile-block): Disable optimization
in lexbind, because it needs a different implementation.
* src/bytecode.c (exec_byte_code): Fix handling of &rest.
* src/eval.c (Vinternal_interpreter_environment): Remove.
(syms_of_eval): Do declare Vinternal_interpreter_environment as
a global lisp var, but unintern it to hide it.
(Fcommandp):
* src/data.c (Finteractive_form): Understand `closure's.
* src/eval.c (apply_lambda, funcall_lambda): Remove lexenv arg.
(Feval): Always eval in the empty environment.
(eval_sub): New function. Use it for all calls to Feval that should
evaluate in the lexical environment of the caller.
Pass `closure's as is to apply_lambda.
(Ffuncall): Pass `closure's as is to funcall_lambda.
(funcall_lambda): Extract lexenv for `closure's, when applicable.
Also use lexical scoping for the &rest argument, if applicable.
* src/lisp.h (eval_sub): Declare.
* src/lread.c (readevalloop): Remove `evalfun' argument.
* src/print.c (Fwith_output_to_temp_buffer):
* src/data.c (Fsetq_default): Use eval_sub.
* lisp/emacs-lisp/bytecomp.el (byte-compile-condition-case): Use push.
Don't macroexpand before evaluating in eval-and-compile, in case
`body's macro expansion uses macros and functions defined in itself.
* src/bytecode.c (exec_byte_code):
* src/eval.c (Ffunctionp): Fix up int/Lisp_Object confusions.
* eval.c (unbind_to): Don't unbind a local binding into the global
binding when the local binding disappeared. Inversely, don't unbind
a global binding into a newly created local binding.
* data.c (set_internal): Make its `buf' arg into a `where' arg so we
can specify the frame to use, when applicable. Adjust callers.
Basically, this makes the structure of buffer-local values and object
forwarding explicit in the type of Lisp_Symbols rather than use
special Lisp_Objects for that. This tends to lead to slightly more
verbose code, but is more C-like, simpler, and makes it easier to make
sure we handled all cases, among other things by letting the compiler
help us check it.
* lisp.h (enum Lisp_Misc_Type, union Lisp_Misc):
Removing forwarding objects.
(enum Lisp_Fwd_Type, enum symbol_redirect, union Lisp_Fwd): New types.
(struct Lisp_Symbol): Make the various forms of variable-forwarding
explicit rather than hiding them inside Lisp_Object "values".
(XFWDTYPE): New macro.
(XINTFWD, XBOOLFWD, XOBJFWD, XKBOARD_OBJFWD): Redefine.
(XBUFFER_LOCAL_VALUE): Remove.
(SYMBOL_VAL, SYMBOL_ALIAS, SYMBOL_BLV, SYMBOL_FWD, SET_SYMBOL_VAL)
(SET_SYMBOL_ALIAS, SET_SYMBOL_BLV, SET_SYMBOL_FWD): New macros.
(SYMBOL_VALUE, SET_SYMBOL_VALUE): Remove.
(struct Lisp_Intfwd, struct Lisp_Boolfwd, struct Lisp_Objfwd)
(struct Lisp_Buffer_Objfwd, struct Lisp_Kboard_Objfwd):
Remove the Lisp_Misc_* header.
(struct Lisp_Buffer_Local_Value): Redefine.
(BLV_FOUND, SET_BLV_FOUND, BLV_VALUE, SET_BLV_VALUE): New macros.
(struct Lisp_Misc_Any): Add filler to get the right size.
(struct Lisp_Free): Use struct Lisp_Misc_Any rather than struct
Lisp_Intfwd.
(DEFVAR_LISP, DEFVAR_LISP_NOPRO, DEFVAR_BOOL, DEFVAR_INT)
(DEFVAR_KBOARD): Allocate a forwarding object.
* data.c (do_blv_forwarding, store_blv_forwarding): New macros.
(let_shadows_global_binding_p): New function.
(union Lisp_Val_Fwd): New type.
(make_blv): New function.
(swap_in_symval_forwarding, indirect_variable, do_symval_forwarding)
(store_symval_forwarding, swap_in_global_binding, Fboundp)
(swap_in_symval_forwarding, find_symbol_value, Fset)
(let_shadows_buffer_binding_p, set_internal, default_value)
(Fset_default, Fmake_variable_buffer_local, Fmake_local_variable)
(Fkill_local_variable, Fmake_variable_frame_local)
(Flocal_variable_p, Flocal_variable_if_set_p)
(Fvariable_binding_locus):
* xdisp.c (select_frame_for_redisplay):
* lread.c (Fintern, Funintern, init_obarray, defvar_int)
(defvar_bool, defvar_lisp_nopro, defvar_lisp, defvar_kboard):
* frame.c (store_frame_param):
* eval.c (Fdefvaralias, Fuser_variable_p, specbind, unbind_to):
* bytecode.c (Fbyte_code) <varref, varset>: Adapt to the new symbol
value structure.
* buffer.c (PER_BUFFER_SYMBOL): Move from buffer.h.
(clone_per_buffer_values): Only adjust markers into the current buffer.
(reset_buffer_local_variables): PER_BUFFER_IDX is never -2.
(Fbuffer_local_value, set_buffer_internal_1)
(swap_out_buffer_local_variables):
Adapt to the new symbol value structure.
(DEFVAR_PER_BUFFER): Allocate a Lisp_Buffer_Objfwd object.
(defvar_per_buffer): Take a new arg for the fwd object.
(buffer_lisp_local_variables): Return a proper alist (different fix
for bug#4138).
* alloc.c (Fmake_symbol): Use SET_SYMBOL_VAL.
(Fgarbage_collect): Don't handle buffer_defaults specially.
(mark_object): Handle new symbol value structure rather than the old
special Lisp_Misc_* objects.
(gc_sweep) <symbols>: Free also the buffer-local-value objects.
* term.c (set_tty_color_mode):
* bidi.c (bidi_initialize): Don't access the ->value field directly.
* buffer.h (PER_BUFFER_VAR_OFFSET): Don't bother with
a buffer_local_flags.
* print.c (print_object): Get rid of impossible forwarding objects.
