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

fix PR/41136: \<word search doesn't work in vi

Browse Source 
don't reuse RCHAR_T(=wchar_t)'s bits, CSI wchar_t is opaque object.
This commit is contained in:
tnozaki 2009-04-12 14:47:51 +00:00
parent 2103c225d0
commit 4bdba43474
2 changed files with 48 additions and 56 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

5
dist/nvi/common/multibyte.h vendored
View File

@ -1,4 +1,4 @@
/* $NetBSD: multibyte.h,v 1.2 2009/01/02 00:32:11 tnozaki Exp $ */ /* $NetBSD: multibyte.h,v 1.3 2009/04/12 14:47:51 tnozaki Exp $ */
#ifndef MULTIBYTE_H #ifndef MULTIBYTE_H
#define MULTIBYTE_H #define MULTIBYTE_H
@ -10,10 +10,8 @@
typedef wchar_t RCHAR_T; typedef wchar_t RCHAR_T;
typedef wchar_t CHAR_T; typedef wchar_t CHAR_T;
#if defined(__NetBSD__) #if defined(__NetBSD__)
#define RCHAR_T_MAX 0xffffffff
#define MAX_CHAR_T 0xffffffff #define MAX_CHAR_T 0xffffffff
#else #else
#define RCHAR_T_MAX WCHAR_MAX
#define MAX_CHAR_T WCHAR_MAX #define MAX_CHAR_T WCHAR_MAX
#endif #endif
typedef u_int UCHAR_T; typedef u_int UCHAR_T;
@ -31,7 +29,6 @@ typedef u_int UCHAR_T;
#else #else
typedef char RCHAR_T; typedef char RCHAR_T;
#define RCHAR_T_MAX CHAR_MAX
typedef u_char CHAR_T; typedef u_char CHAR_T;
#define MAX_CHAR_T 0xff #define MAX_CHAR_T 0xff
typedef u_char UCHAR_T; typedef u_char UCHAR_T;

99
dist/nvi/regex/engine.c vendored
View File

@ -1,4 +1,4 @@
/* $NetBSD: engine.c,v 1.5 2009/02/22 11:34:53 tnozaki Exp $ */ /* $NetBSD: engine.c,v 1.6 2009/04/12 14:47:51 tnozaki Exp $ */
/*- /*-
* Copyright (c) 1992, 1993, 1994 Henry Spencer. * Copyright (c) 1992, 1993, 1994 Henry Spencer.
@ -97,16 +97,13 @@ static RCHAR_T *dissect __P((struct match *m, RCHAR_T *start, RCHAR_T *stop, sop
static RCHAR_T *backref __P((struct match *m, RCHAR_T *start, RCHAR_T *stop, sopno startst, sopno stopst, sopno lev)); static RCHAR_T *backref __P((struct match *m, RCHAR_T *start, RCHAR_T *stop, sopno startst, sopno stopst, sopno lev));
static RCHAR_T *fast __P((struct match *m, RCHAR_T *start, RCHAR_T *stop, sopno startst, sopno stopst)); static RCHAR_T *fast __P((struct match *m, RCHAR_T *start, RCHAR_T *stop, sopno startst, sopno stopst));
static RCHAR_T *slow __P((struct match *m, RCHAR_T *start, RCHAR_T *stop, sopno startst, sopno stopst)); static RCHAR_T *slow __P((struct match *m, RCHAR_T *start, RCHAR_T *stop, sopno startst, sopno stopst));
static states step __P((struct re_guts *g, sopno start, sopno stop, states bef, int ch, states aft)); static states step __P((struct re_guts *g, sopno start, sopno stop, states bef, int flag, RCHAR_T ch, states aft));
#define BOL (OUT+1) #define BOL (1)
#define EOL (BOL+1) #define EOL (BOL+1)
#define BOLEOL (BOL+2) #define BOLEOL (BOL+2)
#define NOTHING (BOL+3) #define NOTHING (BOL+3)
#define BOW (BOL+4) #define BOW (BOL+4)
#define EOW (BOL+5) #define EOW (BOL+5)
#define CODEMAX (BOL+5) /* highest code used */
#define NONCHAR(c) ((UCHAR_T)(c) > (UCHAR_T)RCHAR_T_MAX)
#define NNONCHAR (CODEMAX-CHAR_MAX)
#ifdef REDEBUG #ifdef REDEBUG
static void print __P((struct match *m, char *caption, states st, int ch, FILE *d)); static void print __P((struct match *m, char *caption, states st, int ch, FILE *d));
#endif #endif
@ -716,15 +713,15 @@ sopno stopst;
register states fresh = m->fresh; register states fresh = m->fresh;
register states tmp = m->tmp; register states tmp = m->tmp;
register RCHAR_T *p = start; register RCHAR_T *p = start;
register int c = (start == m->beginp) ? OUT : *(start-1); register RCHAR_T c = (start == m->beginp) ? OUT : *(start-1);
register int lastc; /* previous c */ register RCHAR_T lastc; /* previous c */
register int flagch; register int flag;
register int i; register int i;
register RCHAR_T *coldp; /* last p after which no match was underway */ register RCHAR_T *coldp; /* last p after which no match was underway */
CLEAR(st); CLEAR(st);
SET1(st, startst); SET1(st, startst);
st = step(m->g, startst, stopst, st, NOTHING, st); st = step(m->g, startst, stopst, st, NOTHING, OUT, st);
ASSIGN(fresh, st); ASSIGN(fresh, st);
SP("start", st, *p); SP("start", st, *p);
coldp = NULL; coldp = NULL;
@ -736,35 +733,35 @@ sopno stopst;
coldp = p; coldp = p;
/* is there an EOL and/or BOL between lastc and c? */ /* is there an EOL and/or BOL between lastc and c? */
flagch = '\0'; flag = 0;
i = 0; i = 0;
if ( (lastc == '\n' && m->g->cflags&REG_NEWLINE) || if ( (lastc == '\n' && m->g->cflags&REG_NEWLINE) ||
(lastc == OUT && !(m->eflags&REG_NOTBOL)) ) { (lastc == OUT && !(m->eflags&REG_NOTBOL)) ) {
flagch = BOL; flag = BOL;
i = m->g->nbol; i = m->g->nbol;
} }
if ( (c == '\n' && m->g->cflags&REG_NEWLINE) || if ( (c == '\n' && m->g->cflags&REG_NEWLINE) ||
(c == OUT && !(m->eflags&REG_NOTEOL)) ) { (c == OUT && !(m->eflags&REG_NOTEOL)) ) {
flagch = (flagch == BOL) ? BOLEOL : EOL; flag = (flag == BOL) ? BOLEOL : EOL;
i += m->g->neol; i += m->g->neol;
} }
if (i != 0) { if (i != 0) {
for (; i > 0; i--) for (; i > 0; i--)
st = step(m->g, startst, stopst, st, flagch, st); st = step(m->g, startst, stopst, st, flag, OUT, st);
SP("boleol", st, c); SP("boleol", st, c);
} }
/* how about a word boundary? */ /* how about a word boundary? */
if ( (flagch == BOL || (lastc != OUT && !ISWORD(lastc))) && if ( (flag == BOL || (lastc != OUT && !ISWORD(lastc))) &&
(c != OUT && ISWORD(c)) ) { (c != OUT && ISWORD(c)) ) {
flagch = BOW; flag = BOW;
} }
if ( (lastc != OUT && ISWORD(lastc)) && if ( (lastc != OUT && ISWORD(lastc)) &&
(flagch == EOL || (c != OUT && !ISWORD(c))) ) { (flag == EOL || (c != OUT && !ISWORD(c))) ) {
flagch = EOW; flag = EOW;
} }
if (flagch == BOW || flagch == EOW) { if (flag == BOW || flag == EOW) {
st = step(m->g, startst, stopst, st, flagch, st); st = step(m->g, startst, stopst, st, flag, OUT, st);
SP("boweow", st, c); SP("boweow", st, c);
} }
@ -776,9 +773,9 @@ sopno stopst;
ASSIGN(tmp, st); ASSIGN(tmp, st);
ASSIGN(st, fresh); ASSIGN(st, fresh);
assert(c != OUT); assert(c != OUT);
st = step(m->g, startst, stopst, tmp, c, st); st = step(m->g, startst, stopst, tmp, 0, c, st);
SP("aft", st, c); SP("aft", st, c);
assert(EQ(step(m->g, startst, stopst, st, NOTHING, st), st)); assert(EQ(step(m->g, startst, stopst, st, NOTHING, OUT, st), st));
p++; p++;
} }
@ -807,9 +804,9 @@ sopno stopst;
register states empty = m->empty; register states empty = m->empty;
register states tmp = m->tmp; register states tmp = m->tmp;
register RCHAR_T *p = start; register RCHAR_T *p = start;
register int c = (start == m->beginp) ? OUT : *(start-1); register RCHAR_T c = (start == m->beginp) ? OUT : *(start-1);
register int lastc; /* previous c */ register RCHAR_T lastc; /* previous c */
register int flagch; register int flag;
register int i; register int i;
register RCHAR_T *matchp; /* last p at which a match ended */ register RCHAR_T *matchp; /* last p at which a match ended */
@ -817,7 +814,7 @@ sopno stopst;
CLEAR(st); CLEAR(st);
SET1(st, startst); SET1(st, startst);
SP("sstart", st, *p); SP("sstart", st, *p);
st = step(m->g, startst, stopst, st, NOTHING, st); st = step(m->g, startst, stopst, st, NOTHING, OUT, st);
matchp = NULL; matchp = NULL;
for (;;) { for (;;) {
/* next character */ /* next character */
@ -825,35 +822,35 @@ sopno stopst;
c = (p == m->endp) ? OUT : *p; c = (p == m->endp) ? OUT : *p;
/* is there an EOL and/or BOL between lastc and c? */ /* is there an EOL and/or BOL between lastc and c? */
flagch = '\0'; flag = 0;
i = 0; i = 0;
if ( (lastc == '\n' && m->g->cflags&REG_NEWLINE) || if ( (lastc == '\n' && m->g->cflags&REG_NEWLINE) ||
(lastc == OUT && !(m->eflags&REG_NOTBOL)) ) { (lastc == OUT && !(m->eflags&REG_NOTBOL)) ) {
flagch = BOL; flag = BOL;
i = m->g->nbol; i = m->g->nbol;
} }
if ( (c == '\n' && m->g->cflags&REG_NEWLINE) || if ( (c == '\n' && m->g->cflags&REG_NEWLINE) ||
(c == OUT && !(m->eflags&REG_NOTEOL)) ) { (c == OUT && !(m->eflags&REG_NOTEOL)) ) {
flagch = (flagch == BOL) ? BOLEOL : EOL; flag = (flag == BOL) ? BOLEOL : EOL;
i += m->g->neol; i += m->g->neol;
} }
if (i != 0) { if (i != 0) {
for (; i > 0; i--) for (; i > 0; i--)
st = step(m->g, startst, stopst, st, flagch, st); st = step(m->g, startst, stopst, st, flag, OUT, st);
SP("sboleol", st, c); SP("sboleol", st, c);
} }
/* how about a word boundary? */ /* how about a word boundary? */
if ( (flagch == BOL || (lastc != OUT && !ISWORD(lastc))) && if ( (flag == BOL || (lastc != OUT && !ISWORD(lastc))) &&
(c != OUT && ISWORD(c)) ) { (c != OUT && ISWORD(c)) ) {
flagch = BOW; flag = BOW;
} }
if ( (lastc != OUT && ISWORD(lastc)) && if ( (lastc != OUT && ISWORD(lastc)) &&
(flagch == EOL || (c != OUT && !ISWORD(c))) ) { (flag == EOL || (c != OUT && !ISWORD(c))) ) {
flagch = EOW; flag = EOW;
} }
if (flagch == BOW || flagch == EOW) { if (flag == BOW || flag == EOW) {
st = step(m->g, startst, stopst, st, flagch, st); st = step(m->g, startst, stopst, st, flag, OUT, st);
SP("sboweow", st, c); SP("sboweow", st, c);
} }
@ -867,9 +864,9 @@ sopno stopst;
ASSIGN(tmp, st); ASSIGN(tmp, st);
ASSIGN(st, empty); ASSIGN(st, empty);
assert(c != OUT); assert(c != OUT);
st = step(m->g, startst, stopst, tmp, c, st); st = step(m->g, startst, stopst, tmp, 0, c, st);
SP("saft", st, c); SP("saft", st, c);
assert(EQ(step(m->g, startst, stopst, st, NOTHING, st), st)); assert(EQ(step(m->g, startst, stopst, st, NOTHING, OUT, st), st));
p++; p++;
} }
@ -880,24 +877,22 @@ sopno stopst;
/* /*
- step - map set of states reachable before char to set reachable after - step - map set of states reachable before char to set reachable after
== static states step(register struct re_guts *g, sopno start, sopno stop, \ == static states step(register struct re_guts *g, sopno start, sopno stop, \
== register states bef, int ch, register states aft); == register states bef, int flag, RCHAR_T ch, register states aft);
== #define BOL (OUT+1) == #define BOL (1)
== #define EOL (BOL+1) == #define EOL (BOL+1)
== #define BOLEOL (BOL+2) == #define BOLEOL (BOL+2)
== #define NOTHING (BOL+3) == #define NOTHING (BOL+3)
== #define BOW (BOL+4) == #define BOW (BOL+4)
== #define EOW (BOL+5) == #define EOW (BOL+5)
== #define CODEMAX (BOL+5) // highest code used
== #define NONCHAR(c) ((c) > CHAR_MAX)
== #define NNONCHAR (CODEMAX-CHAR_MAX)
*/ */
static states static states
step(g, start, stop, bef, ch, aft) step(g, start, stop, bef, flag, ch, aft)
register struct re_guts *g; register struct re_guts *g;
sopno start; /* start state within strip */ sopno start; /* start state within strip */
sopno stop; /* state after stop state within strip */ sopno stop; /* state after stop state within strip */
register states bef; /* states reachable before */ register states bef; /* states reachable before */
int ch; /* character or NONCHAR code */ int flag; /* NONCHAR flag */
RCHAR_T ch; /* character code */
register states aft; /* states already known reachable after */ register states aft; /* states already known reachable after */
{ {
register cset *cs; register cset *cs;
@ -917,33 +912,33 @@ register states aft; /* states already known reachable after */
break; break;
case OCHAR: case OCHAR:
/* only characters can match */ /* only characters can match */
assert(!NONCHAR(ch) || ch != d); assert(!flag || ch != d);
if (ch == d) if (ch == d)
FWD(aft, bef, 1); FWD(aft, bef, 1);
break; break;
case OBOL: case OBOL:
if (ch == BOL || ch == BOLEOL) if (flag == BOL || flag == BOLEOL)
FWD(aft, bef, 1); FWD(aft, bef, 1);
break; break;
case OEOL: case OEOL:
if (ch == EOL || ch == BOLEOL) if (flag == EOL || flag == BOLEOL)
FWD(aft, bef, 1); FWD(aft, bef, 1);
break; break;
case OBOW: case OBOW:
if (ch == BOW) if (flag == BOW)
FWD(aft, bef, 1); FWD(aft, bef, 1);
break; break;
case OEOW: case OEOW:
if (ch == EOW) if (flag == EOW)
FWD(aft, bef, 1); FWD(aft, bef, 1);
break; break;
case OANY: case OANY:
if (!NONCHAR(ch)) if (!flag)
FWD(aft, bef, 1); FWD(aft, bef, 1);
break; break;
case OANYOF: case OANYOF:
cs = &g->sets[d]; cs = &g->sets[d];
if (!NONCHAR(ch) && CHIN(cs, ch)) if (!flag && CHIN(cs, ch))
FWD(aft, bef, 1); FWD(aft, bef, 1);
break; break;
case OBACK_: /* ignored here */ case OBACK_: /* ignored here */