Mercurial > hg > xemacs-beta
view src/lisp-disunion.h @ 801:2b676dc88c66
[xemacs-hg @ 2002-04-01 03:58:02 by ben]
bug fixes (e.g. ballooning on X windows)
Makefile.in.in: Try to make the Makefile notice if its source Makefile.in.in is
changed, and regenerate and run itself.
Use a bigger default SHEAP_ADJUSTMENT on Cygwin; otherwise you
can't compile under Mule if a Lisp file has changed. (can't run
temacs)
TODO.ben-mule-21-5: update.
mule/mule-cmds.el: Hash the result of mswindows-get-language-environment-from-locale,
since it's very expensive (and causes huge ballooning of memory
under X Windows, since it's called from x-get-resource).
cl-extra.el, code-files.el, files.el, simple.el, subr.el, x-faces.el: Create new string-equal-ignore-case, based on built-in
compare-strings -- compare strings ignoring case without the need
to generate garbage by calling downcase. Use it in equalp and
elsewhere.
alloc.c, bytecode.c, chartab.c, data.c, elhash.c, emacs.c, eval.c, event-Xt.c, event-unixoid.c, extents.c, file-coding.c, fileio.c, fns.c, glyphs.c, gutter.c, lisp-union.h, lisp.h, mule-charset.c, nt.c, process-unix.c, process.c, specifier.c, symbols.c, sysdep.c, sysdep.h, text.c, toolbar.c: Try to implement GC triggering based on percentage of total memory
usage. Not currently activated (percentage set to 0) because not
quite working. Add `memory-usage' primitive to return XEmacs'
idea of its memory usage.
Add primitive compare-strings, compatible with FSF 21.1 -- can
compare any part of two strings, optionally ignoring case.
Improve qxe() functions in text.c for text comparison.
Use RETURN_NOT_REACHED to try to avoid warnings about unreachable
code.
Add volatile_make_int() to fix warning in unix_send_process().
author | ben |
---|---|
date | Mon, 01 Apr 2002 03:59:04 +0000 |
parents | a5954632b187 |
children | 6728e641994e |
line wrap: on
line source
/* Fundamental definitions for XEmacs Lisp interpreter -- non-union objects. Copyright (C) 1985, 1986, 1987, 1992, 1993 Free Software Foundation, Inc. Copyright (C) 2001 Ben Wing. This file is part of XEmacs. XEmacs is free software; you can redistribute it and/or modify it under the terms of the GNU General Public License as published by the Free Software Foundation; either version 2, or (at your option) any later version. XEmacs is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License for more details. You should have received a copy of the GNU General Public License along with XEmacs; see the file COPYING. If not, write to the Free Software Foundation, Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA. */ /* Synched up with: FSF 19.30. Split out from lisp.h. */ /* This file has diverged greatly from FSF Emacs. Syncing is no longer desirable or possible */ /* Format of a non-union-type Lisp Object 3 2 1 0 bit 10987654321098765432109876543210 -------------------------------- VVVVVVVVVVVVVVVVVVVVVVVVVVVVVVTT Integers are treated specially, and look like this: 3 2 1 0 bit 10987654321098765432109876543210 -------------------------------- VVVVVVVVVVVVVVVVVVVVVVVVVVVVVVVT For integral Lisp types, i.e. integers and characters, the value bits are the Lisp object. Some people call such Lisp_Objects "immediate". The object is obtained by masking off the type bits. Bit 1 is used as a value bit by splitting the Lisp integer type into two subtypes, Lisp_Type_Int_Even and Lisp_Type_Int_Odd. By this trickery we get 31 bits for integers instead of 30. For non-integral types, the value bits of a Lisp_Object contain a pointer to a structure containing the object. The pointer is obtained by masking off the type and mark bits. All pointer-based types are coalesced under a single type called Lisp_Type_Record. The type bits for this type are required by the implementation to be 00, just like the least significant bits of word-aligned struct pointers on 32-bit hardware. This requires that all structs implementing Lisp_Objects have an alignment of at least 4 bytes. Because of this, Lisp_Object pointers don't have to be masked and are full-sized. There are no mark bits in the Lisp_Object itself (there used to be). Integers and characters don't need to be marked. All other types are lrecord-based, which means they get marked by setting the mark bit in the struct lrecord_header. Here is a brief description of the following macros: XTYPE The type bits of a Lisp_Object XPNTRVAL The value bits of a Lisp_Object storing a pointer XCHARVAL The value bits of a Lisp_Object storing a Emchar XREALINT The value bits of a Lisp_Object storing an integer, signed XUINT The value bits of a Lisp_Object storing an integer, unsigned INTP Non-zero if this Lisp_Object is an integer Qzero Lisp Integer 0 EQ Non-zero if two Lisp_Objects are identical, not merely equal. */ typedef EMACS_INT Lisp_Object; #define Lisp_Type_Int_Bit (Lisp_Type_Int_Even & Lisp_Type_Int_Odd) #define wrap_pointer_1(ptr) ((Lisp_Object) (ptr)) #define make_int(x) ((Lisp_Object) (((x) << INT_GCBITS) | Lisp_Type_Int_Bit)) #define volatile_make_int(x) make_int (x) #define make_char(x) ((Lisp_Object) (((x) << GCBITS) | Lisp_Type_Char)) #define VALMASK (((1UL << VALBITS) - 1UL) << GCTYPEBITS) #define XTYPE(x) ((enum Lisp_Type) (((EMACS_UINT)(x)) & ~VALMASK)) #define XPNTRVAL(x) (x) /* This depends on Lisp_Type_Record == 0 */ #define XCHARVAL(x) ((x) >> GCBITS) #define XREALINT(x) ((x) >> INT_GCBITS) #define XUINT(x) ((EMACS_UINT)(x) >> INT_GCBITS) #define INTP(x) ((EMACS_UINT)(x) & Lisp_Type_Int_Bit) #define INT_PLUS(x,y) ((x)+(y)-Lisp_Type_Int_Bit) #define INT_MINUS(x,y) ((x)-(y)+Lisp_Type_Int_Bit) #define INT_PLUS1(x) INT_PLUS (x, make_int (1)) #define INT_MINUS1(x) INT_MINUS (x, make_int (1)) #define Qzero make_int (0) #define Qnull_pointer ((Lisp_Object) 0) #define EQ(x,y) ((x) == (y)) /* Convert between a (void *) and a Lisp_Object, as when the Lisp_Object is passed to a toolkit callback function */ #define VOID_TO_LISP(larg,varg) ((void) ((larg) = ((Lisp_Object) (varg)))) #define CVOID_TO_LISP VOID_TO_LISP #define LISP_TO_VOID(larg) ((void *) (larg)) #define LISP_TO_CVOID(larg) ((const void *) (larg)) /* Convert a Lisp_Object into something that can't be used as an lvalue. Useful for type-checking. */ #define NON_LVALUE(larg) ((larg) + 0)