Aren/NetBSD
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity

Added comment explaining why we are using a different algorithm than

Browse Source 
what was adopted in tzcode95c to solve the problem of the first char
detzcode() extracts requiring sign extention (this is needed for
machines with 64 bit longs).

Our implementation is a bit more efficent, but requires a new macro,
SIGN_EXTEND_CHAR(), and a conditional to set it appropriately).

One minor enhancement for machines without ANSI compilers that have 32
bit longs would be to define SIGN_EXTEND_CHAR() to be the identity
macro.
This commit is contained in:
jtc 1995-03-16 19:14:16 +00:00
parent 15642dad87
commit 3924aee202
1 changed files with 15 additions and 2 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

17
lib/libc/time/localtime.c
View File

@ -1,4 +1,4 @@
/* $NetBSD: localtime.c,v 1.3 1995/03/10 05:57:35 jtc Exp $ */
/* $NetBSD: localtime.c,v 1.4 1995/03/16 19:14:16 jtc Exp $ */
#ifndef lint
#ifndef NOID
@ -191,7 +191,20 @@ const char * const codep;
{
register long result;
result = (codep[0] << 24) \
/*
** The first character must be sign extended on systems with >32bit
** longs. This was solved differently in the master tzcode sources
** (the fix first appeared in tzcode95c.tar.gz). But I believe
** that this implementation is superior.
*/
#ifdef __STDC__
#define SIGN_EXTEND_CHAR(x) ((signed char) x)
#else
#define SIGN_EXTEND_CHAR(x) ((x & 0x80) ? ((~0 << 8) | x) : x)
#endif
result = (SIGN_EXTEND_CHAR(codep[0]) << 24) \
| (codep[1] & 0xff) << 16 \
| (codep[2] & 0xff) << 8
| (codep[3] & 0xff);