Mercurial > hg > xemacs-beta
view src/lisp-disunion.h @ 1330:4542b72c005e
[xemacs-hg @ 2003-03-01 07:25:26 by ben]
build patch
Makefile.in.in: Move src deletions to src/Makefile.in.in.
dump-paths.el, dumped-lisp.el: Delete. Combine stuff into setup-paths.el.
find-paths.el: Removed.
Make this file contain generic routines only. Move stuff to
compute Emacs roots to setup-paths.el.
startup.el: Removed.
Move these variables into setup-paths.el.
setup-paths.el, startup.el: Removed.
Combine all high-level code for computing the paths into
setup-paths.el. Create new function startup-find-load-path to
encapsulate all logic for computing `load-path'. Eliminate
invocation-directory and invocation-name parameters since
there is no point (false generality) -- the code references
other globals, which cannot be specified. Eliminate some code
duplicated between setup-paths.el and startup.el. Clean up
the debug-paths code and output load-path in addition.
Add logic to paths-emacs-root-p to support separated source
and build trees.
loadup.el, make-docfile.el, update-elc-2.el, update-elc.el: Rewrite to allow for separated source and build trees, as may occur
in MS Windows.
NOTE TO BUILD HACKERS:
loadup.el, make-docfile.el, update-elc.el and update-elc-2.el made two
assumptions that are no longer correct:
(1) The source and build trees are in the same place.
(2) They can make assumptions about where `.' is.
These files now compute the locations of the source and build
roots at the top of the file. *ALL* constant file names or path
snippets must now be made absolute using expand-file-name and one
of these roots.
dumped-lisp.el, packages.el: Removed.
Remove some unused lists of Lisp files. packages-hardcoded-lisp
(empty, in any case) moved to dumped-lisp.el.
startup.el: When a compiled init file is out-of-date wrt the uncompiled
version, load the uncompiled version and issue a nasty warning.
update-elc-2.el: Force touching of auto-autoloads files when REBUILD_AUTOLOADS
was set.
update-elc.el: Fix code that checks whether dumping is necessary to check against
xemacs.dmp, not xemacs.exe, when Unix and pdump.
lwlib-Xm.c: Fix compile warning.
README, config.inc.samp, xemacs.mak: -- Major reorganization and cleanup.
-- Add support for separated build tree and source tree.
-- Delete all support for X Windows building, since it's
totally bit-rotten and will never be fixed up. Instruct
people to use Cygwin if they want such support.
make-build-dir: New script to create a skeleton build tree for use with
separated build and source tree compilation.
m/acorn.h, m/alliant-2800.h, m/alliant.h, m/altos.h, m/amdahl.h, m/arm.h, m/att3b.h, m/aviion.h, m/clipper.h, m/cnvrgnt.h, m/convex.h, m/cydra5.h, m/delta.h, m/delta88k.h, m/dpx2.h, m/elxsi.h, m/ews4800r.h, m/gould.h, m/hp800.h, m/hp9000s300.h, m/i860.h, m/ibmps2-aix.h, m/ibmrs6000.h, m/ibmrt-aix.h, m/ibmrt.h, m/intel386.h, m/iris4d.h, m/iris5d.h, m/iris6d.h, m/irist.h, m/m68k.h, m/masscomp.h, m/mg1.h, m/mips-nec.h, m/mips-siemens.h, m/mips.h, m/nh3000.h, m/nh4000.h, m/ns32000.h, m/plexus.h, m/powerpc.h, m/sequent-ptx.h, m/sequent.h, m/sgi-challenge.h, m/stride.h, m/tad68k.h, m/targon31.h, m/tekxd88.h, m/template.h, m/tower32.h, m/tower32v3.h, m/ustation.h, m/wicat.h, m/xps100.h, data.c, doc.c, editfns.c, emacs.c, lrecord.h, ntheap.c, process-unix.c, sysdep.c, unexec.c: Delete all support for bit-rotten CANNOT_DUMP. Just use pdump.
Makefile.in.in: Lots o' cleanup. Use names like LISP, SRC instead of
lispdir, srcdir, for consistency with xemacs.mak and the
conventions in the rest of the file. Eliminate use of ${...}
in favor of $(...), to make it easier to move code between
this file and xemacs.mak. Fix dependency handling wrt
NEEDTODUMP to eliminate problems some people (e.g. Vin) have
been seeing with non-GNU makes. Write a long section about
the subtle but oh-so-important differences in dependency
processing between nmake, make, and GNU make. Add
unicode-encapsulate target, from xemacs.mak.
chartab.c, lrecord.h: Fix crash due to attempt to free objects across dump/undump.
author | ben |
---|---|
date | Sat, 01 Mar 2003 07:25:56 +0000 |
parents | 184461bc8de4 |
children | ae48681c47fa |
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, 2002 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 Ichar 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 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 wrap_pointer_1(ptr) ((Lisp_Object) (ptr)) DECLARE_INLINE_HEADER ( Lisp_Object make_int_verify (EMACS_INT val) ) { Lisp_Object obj = (Lisp_Object) ((val << INT_GCBITS) | Lisp_Type_Int_Bit); type_checking_assert (XREALINT (obj) == val); return obj; } #define make_int(x) ((Lisp_Object) (((x) << INT_GCBITS) | Lisp_Type_Int_Bit)) #define make_char_1(x) ((Lisp_Object) (((x) << GCBITS) | Lisp_Type_Char)) #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)) /* WARNING!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! You can only VOID_TO_LISP something that had previously been LISP_TO_VOID'd. You cannot go the other way, i.e. create a bogus Lisp_Object. If you want to stuff a void * into a Lisp_Object, use make_opaque_ptr(). */ /* Convert between a (void *) and a Lisp_Object, as when the Lisp_Object is passed to a toolkit callback function */ #define VOID_TO_LISP(varg) ((Lisp_Object) (varg)) #define LISP_TO_VOID(larg) ((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)