mirror/guile
mirror/guile
1
Fork 0
mirror of https://git.savannah.gnu.org/git/guile.git synced 2025-04-29 19:30:36 +02:00
Code Activity

Reimplement SCM_MAKE_CHAR to evaluate its argument only once.

Browse source 
The motivation for this change is that SCM_MAKE_CHAR is sometimes passed
an expression that involves a procedure call that is not always trivial.
In other cases, the results are not guaranteed to be the same both
times, which could lead to the creation of invalid SCM objects.

* libguile/chars.h (SCM_MAKE_CHAR): Reimplement.
This commit is contained in:
Mark H Weaver 2019-04-16 23:11:42 -04:00
parent 546b0e8729
commit 63818453ad
1 changed files with 12 additions and 9 deletions
Download patch file Download diff file Expand all files Collapse all files

21
libguile/chars.h
View file

@ -3,7 +3,8 @@
#ifndef SCM_CHARS_H
#define SCM_CHARS_H
/* Copyright (C) 1995,1996,2000,2001,2004, 2006, 2008, 2009 Free Software Foundation, Inc.
/* Copyright (C) 1995, 1996, 2000, 2001, 2004, 2006, 2008, 2009, 2019
* Free Software Foundation, Inc.
*
* This library is free software; you can redistribute it and/or
* modify it under the terms of the GNU Lesser General Public License
@ -36,16 +37,18 @@ typedef scm_t_int32 scm_t_wchar;
#define SCM_CHARP(x) (SCM_ITAG8(x) == scm_tc8_char)
#define SCM_CHAR(x) ((scm_t_wchar)SCM_ITAG8_DATA(x))
/* SCM_MAKE_CHAR maps signed chars (-128 to 127) and unsigned chars (0
to 255) to Latin-1 codepoints (0 to 255) while allowing higher
/* SCM_MAKE_CHAR maps signed chars (-128 to 127) and unsigned chars
(0 to 255) to Latin-1 codepoints (0 to 255) while allowing higher
codepoints (256 to 1114111) to pass through unchanged.
This macro evaluates x twice, which may lead to side effects if not
used properly. */
#define SCM_MAKE_CHAR(x) \
((x) <= 1 \
? SCM_MAKE_ITAG8 ((scm_t_bits) (unsigned char) (x), scm_tc8_char) \
: SCM_MAKE_ITAG8 ((scm_t_bits) (x), scm_tc8_char))
To avoid evaluating X more than once, we use an arithmetic trick: we
compute (X mod 2^N) mod (2^N - 256), which is equal to the required
mapping in the range -256 .. (2^N - 257). Here, N is the number of
bits in scm_t_bits. Note that (scm_t_bits) (x) implicitly computes
(X mod 2^N), and (scm_t_bits) -256 equals (2^N - 256). GCC is able
to optimize away these operations in practice. */
#define SCM_MAKE_CHAR(x) \
(SCM_MAKE_ITAG8 ((scm_t_bits) (x) % (scm_t_bits) -256, scm_tc8_char))
#define SCM_CODEPOINT_DOTTED_CIRCLE (0x25cc)
#define SCM_CODEPOINT_SURROGATE_START (0xd800)