=================================================================== RCS file: /cvs/mandoc/chars.c,v retrieving revision 1.32 retrieving revision 1.49 diff -u -p -r1.32 -r1.49 --- mandoc/chars.c 2011/01/30 16:05:37 1.32 +++ mandoc/chars.c 2011/07/22 14:15:15 1.49 @@ -1,4 +1,4 @@ -/* $Id: chars.c,v 1.32 2011/01/30 16:05:37 schwarze Exp $ */ +/* $Id: chars.c,v 1.49 2011/07/22 14:15:15 kristaps Exp $ */ /* * Copyright (c) 2009, 2010 Kristaps Dzonsons * Copyright (c) 2011 Ingo Schwarze @@ -20,12 +20,12 @@ #endif #include -#include +#include #include #include #include "mandoc.h" -#include "chars.h" +#include "libmandoc.h" #define PRINT_HI 126 #define PRINT_LO 32 @@ -35,52 +35,36 @@ struct ln { const char *code; const char *ascii; int unicode; - int type; -#define CHARS_CHAR (1 << 0) -#define CHARS_STRING (1 << 1) -#define CHARS_BOTH (CHARS_CHAR | CHARS_STRING) }; -#define LINES_MAX 351 +#define LINES_MAX 328 #define CHAR(in, ch, code) \ - { NULL, (in), (ch), (code), CHARS_CHAR }, -#define STRING(in, ch, code) \ - { NULL, (in), (ch), (code), CHARS_STRING }, -#define BOTH(in, ch, code) \ - { NULL, (in), (ch), (code), CHARS_BOTH }, + { NULL, (in), (ch), (code) }, #define CHAR_TBL_START static struct ln lines[LINES_MAX] = { #define CHAR_TBL_END }; #include "chars.in" -struct ctab { - enum chars type; +struct mchars { struct ln **htab; }; -static inline int match(const struct ln *, - const char *, size_t, int); -static const struct ln *find(struct ctab *, const char *, size_t, int); +static const struct ln *find(struct mchars *, const char *, size_t); - void -chars_free(void *arg) +mchars_free(struct mchars *arg) { - struct ctab *tab; - tab = (struct ctab *)arg; - - free(tab->htab); - free(tab); + free(arg->htab); + free(arg); } - -void * -chars_init(enum chars type) +struct mchars * +mchars_alloc(void) { - struct ctab *tab; + struct mchars *tab; struct ln **htab; struct ln *pp; int i, hash; @@ -88,22 +72,12 @@ chars_init(enum chars type) /* * Constructs a very basic chaining hashtable. The hash routine * is simply the integral value of the first character. - * Subsequent entries are chained in the order they're processed - * (they're in-line re-ordered during lookup). + * Subsequent entries are chained in the order they're processed. */ - tab = malloc(sizeof(struct ctab)); - if (NULL == tab) { - perror(NULL); - exit((int)MANDOCLEVEL_SYSERR); - } + tab = mandoc_malloc(sizeof(struct mchars)); + htab = mandoc_calloc(PRINT_HI - PRINT_LO + 1, sizeof(struct ln **)); - htab = calloc(PRINT_HI - PRINT_LO + 1, sizeof(struct ln **)); - if (NULL == htab) { - perror(NULL); - exit((int)MANDOCLEVEL_SYSERR); - } - for (i = 0; i < LINES_MAX; i++) { hash = (int)lines[i].code[0] - PRINT_LO; @@ -118,88 +92,47 @@ chars_init(enum chars type) } tab->htab = htab; - tab->type = type; return(tab); } - -/* - * Special character to Unicode codepoint. - */ int -chars_spec2cp(void *arg, const char *p, size_t sz) +mchars_spec2cp(struct mchars *arg, const char *p, size_t sz) { const struct ln *ln; - ln = find((struct ctab *)arg, p, sz, CHARS_CHAR); + ln = find(arg, p, sz); if (NULL == ln) return(-1); return(ln->unicode); } - -/* - * Reserved word to Unicode codepoint. - */ -int -chars_res2cp(void *arg, const char *p, size_t sz) +char +mchars_num2char(const char *p, size_t sz) { - const struct ln *ln; - - ln = find((struct ctab *)arg, p, sz, CHARS_STRING); - if (NULL == ln) - return(-1); - return(ln->unicode); -} - - -/* - * Numbered character to literal character, - * represented as a null-terminated string for additional safety. - */ -const char * -chars_num2char(const char *p, size_t sz) -{ int i; - static char c[2]; - if (sz > 3) - return(NULL); - i = atoi(p); - if (i < 0 || i > 255) - return(NULL); - c[0] = (char)i; - c[1] = '\0'; - return(c); + if ((i = mandoc_strntoi(p, sz, 10)) < 0) + return('\0'); + return(isprint(i) ? i : '\0'); } - -/* - * Special character to string array. - */ -const char * -chars_spec2str(void *arg, const char *p, size_t sz, size_t *rsz) +int +mchars_num2uc(const char *p, size_t sz) { - const struct ln *ln; + int i; - ln = find((struct ctab *)arg, p, sz, CHARS_CHAR); - if (NULL == ln) - return(NULL); - - *rsz = strlen(ln->ascii); - return(ln->ascii); + if ((i = mandoc_strntoi(p, sz, 16)) < 0) + return('\0'); + /* FIXME: make sure we're not in a bogus range. */ + return(i > 0x80 && i <= 0x10FFFF ? i : '\0'); } - -/* - * Reserved word to string array. - */ const char * -chars_res2str(void *arg, const char *p, size_t sz, size_t *rsz) +mchars_spec2str(struct mchars *arg, const char *p, size_t sz, size_t *rsz) { const struct ln *ln; - ln = find((struct ctab *)arg, p, sz, CHARS_STRING); + ln = find(arg, p, sz); if (NULL == ln) return(NULL); @@ -207,59 +140,23 @@ chars_res2str(void *arg, const char *p, size_t sz, siz return(ln->ascii); } - static const struct ln * -find(struct ctab *tab, const char *p, size_t sz, int type) +find(struct mchars *tab, const char *p, size_t sz) { - struct ln *pp, *prev; - struct ln **htab; + struct ln *pp; int hash; assert(p); - if (0 == sz) - return(NULL); - if (p[0] < PRINT_LO || p[0] > PRINT_HI) + if (0 == sz || p[0] < PRINT_LO || p[0] > PRINT_HI) return(NULL); - /* - * Lookup the symbol in the symbol hash. See ascii2htab for the - * hashtable specs. This dynamically re-orders the hash chain - * to optimise for repeat hits. - */ - hash = (int)p[0] - PRINT_LO; - htab = tab->htab; - if (NULL == (pp = htab[hash])) - return(NULL); + for (pp = tab->htab[hash]; pp; pp = pp->next) + if (0 == strncmp(pp->code, p, sz) && + '\0' == pp->code[(int)sz]) + return(pp); - for (prev = NULL; pp; pp = pp->next) { - if ( ! match(pp, p, sz, type)) { - prev = pp; - continue; - } - - if (prev) { - prev->next = pp->next; - pp->next = htab[hash]; - htab[hash] = pp; - } - - return(pp); - } - return(NULL); -} - - -static inline int -match(const struct ln *ln, const char *p, size_t sz, int type) -{ - - if ( ! (ln->type & type)) - return(0); - if (strncmp(ln->code, p, sz)) - return(0); - return('\0' == ln->code[(int)sz]); }