=================================================================== RCS file: /cvs/mandoc/chars.c,v retrieving revision 1.46 retrieving revision 1.62 diff -u -p -r1.46 -r1.62 --- mandoc/chars.c 2011/05/24 21:31:23 1.46 +++ mandoc/chars.c 2014/10/27 13:31:04 1.62 @@ -1,7 +1,7 @@ -/* $Id: chars.c,v 1.46 2011/05/24 21:31:23 kristaps Exp $ */ +/* $Id: chars.c,v 1.62 2014/10/27 13:31:04 schwarze Exp $ */ /* - * Copyright (c) 2009, 2010 Kristaps Dzonsons - * Copyright (c) 2011 Ingo Schwarze + * Copyright (c) 2009, 2010, 2011 Kristaps Dzonsons + * Copyright (c) 2011, 2014 Ingo Schwarze * * Permission to use, copy, modify, and distribute this software for any * purpose with or without fee is hereby granted, provided that the above @@ -15,17 +15,17 @@ * ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION, ARISING OUT OF * OR IN CONNECTION WITH THE USE OR PERFORMANCE OF THIS SOFTWARE. */ -#ifdef HAVE_CONFIG_H #include "config.h" -#endif +#include + #include #include -#include #include #include #include "mandoc.h" +#include "mandoc_aux.h" #include "libmandoc.h" #define PRINT_HI 126 @@ -38,7 +38,7 @@ struct ln { int unicode; }; -#define LINES_MAX 325 +#define LINES_MAX 330 #define CHAR(in, ch, code) \ { NULL, (in), (ch), (code) }, @@ -52,9 +52,10 @@ struct mchars { struct ln **htab; }; -static inline int match(const struct ln *, const char *, size_t); -static const struct ln *find(struct mchars *, const char *, size_t); +static const struct ln *find(const struct mchars *, + const char *, size_t); + void mchars_free(struct mchars *arg) { @@ -74,12 +75,11 @@ mchars_alloc(void) /* * 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 = mandoc_malloc(sizeof(struct mchars)); - htab = mandoc_calloc(PRINT_HI - PRINT_LO + 1, sizeof(struct ln **)); + htab = mandoc_calloc(PRINT_HI - PRINT_LO + 1, sizeof(struct ln *)); for (i = 0; i < LINES_MAX; i++) { hash = (int)lines[i].code[0] - PRINT_LO; @@ -98,118 +98,85 @@ mchars_alloc(void) return(tab); } - -/* - * Special character to Unicode codepoint. - */ int -mchars_spec2cp(struct mchars *arg, const char *p, size_t sz) +mchars_spec2cp(const struct mchars *arg, const char *p, size_t sz) { const struct ln *ln; ln = find(arg, p, sz); - if (NULL == ln) - return(-1); - return(ln->unicode); + return(ln != NULL ? ln->unicode : sz == 1 ? *p : 0xFFFD); } -/* - * Numbered character string to ASCII codepoint. - * This can only be a printable character (i.e., alnum, punct, space) so - * prevent the character from ruining our state (backspace, newline, and - * so on). - * If the character is illegal, returns '\0'. - */ char mchars_num2char(const char *p, size_t sz) { - int i; + int i; - if ((i = mandoc_strntou(p, sz, 10)) < 0) + if ((i = mandoc_strntoi(p, sz, 10)) < 0) return('\0'); - return(isprint(i) ? i : '\0'); + + return(i > 0 && i < 256 && isprint(i) ? i : '\0'); } -/* - * Hex character string to Unicode codepoint. - * If the character is illegal, returns '\0'. - */ int mchars_num2uc(const char *p, size_t sz) { - int i; + int i; - if ((i = mandoc_strntou(p, sz, 16)) < 0) - return('\0'); - /* FIXME: make sure we're not in a bogus range. */ - return(i > 0x80 && i <= 0x10FFFF ? i : '\0'); + if ((i = mandoc_strntoi(p, sz, 16)) < 0) + return(0xFFFD); + + /* + * XXX Code is missing here to exclude bogus ranges. + */ + + return(i <= 0x10FFFF ? i : 0xFFFD); } -/* - * Special character to string array. - */ const char * -mchars_spec2str(struct mchars *arg, const char *p, size_t sz, size_t *rsz) +mchars_spec2str(const struct mchars *arg, + const char *p, size_t sz, size_t *rsz) { const struct ln *ln; ln = find(arg, p, sz); - if (NULL == ln) - return(NULL); + if (ln == NULL) { + *rsz = 1; + return(sz == 1 ? p : NULL); + } *rsz = strlen(ln->ascii); return(ln->ascii); } +const char * +mchars_uc2str(int uc) +{ + int i; + + for (i = 0; i < LINES_MAX; i++) + if (uc == lines[i].unicode) + return(lines[i].ascii); + return(""); +} + static const struct ln * -find(struct mchars *tab, const char *p, size_t sz) +find(const struct mchars *tab, const char *p, size_t sz) { - struct ln *pp, *prev; - struct ln **htab; + const 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)) { - 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) -{ - - if (strncmp(ln->code, p, sz)) - return(0); - return('\0' == ln->code[(int)sz]); }