=================================================================== RCS file: /cvs/mandoc/mandoc.c,v retrieving revision 1.40 retrieving revision 1.82 diff -u -p -r1.40 -r1.82 --- mandoc/mandoc.c 2011/03/17 09:16:38 1.40 +++ mandoc/mandoc.c 2014/07/06 18:37:34 1.82 @@ -1,7 +1,7 @@ -/* $Id: mandoc.c,v 1.40 2011/03/17 09:16:38 kristaps Exp $ */ +/* $Id: mandoc.c,v 1.82 2014/07/06 18:37:34 schwarze Exp $ */ /* - * Copyright (c) 2008, 2009, 2010 Kristaps Dzonsons - * Copyright (c) 2011 Ingo Schwarze + * Copyright (c) 2008, 2009, 2010, 2011 Kristaps Dzonsons + * Copyright (c) 2011, 2012, 2013, 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 @@ -23,12 +23,15 @@ #include #include +#include +#include #include #include #include #include #include "mandoc.h" +#include "mandoc_aux.h" #include "libmandoc.h" #define DATESIZE 32 @@ -36,293 +39,375 @@ static int a2time(time_t *, const char *, const char *); static char *time2a(time_t); -int -mandoc_special(char *p) + +enum mandoc_esc +mandoc_escape(const char **end, const char **start, int *sz) { - int len, i; + const char *local_start; + int local_sz; char term; - char *sv; - - len = 0; + enum mandoc_esc gly; + + /* + * When the caller doesn't provide return storage, + * use local storage. + */ + + if (NULL == start) + start = &local_start; + if (NULL == sz) + sz = &local_sz; + + /* + * Beyond the backslash, at least one input character + * is part of the escape sequence. With one exception + * (see below), that character won't be returned. + */ + + gly = ESCAPE_ERROR; + *start = ++*end; + *sz = 0; term = '\0'; - sv = p; - assert('\\' == *p); - p++; + switch ((*start)[-1]) { + /* + * First the glyphs. There are several different forms of + * these, but each eventually returns a substring of the glyph + * name. + */ + case '(': + gly = ESCAPE_SPECIAL; + *sz = 2; + break; + case '[': + gly = ESCAPE_SPECIAL; + /* + * Unicode escapes are defined in groff as \[uXXXX] to + * \[u10FFFF], where the contained value must be a valid + * Unicode codepoint. Here, however, only check whether + * it's not a zero-width escape. + */ + if ('u' == (*start)[0] && ']' != (*start)[1]) + gly = ESCAPE_UNICODE; + term = ']'; + break; + case 'C': + if ('\'' != **start) + return(ESCAPE_ERROR); + *start = ++*end; + if ('u' == (*start)[0] && '\'' != (*start)[1]) + gly = ESCAPE_UNICODE; + else + gly = ESCAPE_SPECIAL; + term = '\''; + break; - switch (*p++) { -#if 0 - case ('Z'): + /* + * Escapes taking no arguments at all. + */ + case 'd': /* FALLTHROUGH */ - case ('X'): + case 'u': + return(ESCAPE_IGNORE); + + /* + * The \z escape is supposed to output the following + * character without advancing the cursor position. + * Since we are mostly dealing with terminal mode, + * let us just skip the next character. + */ + case 'z': + return(ESCAPE_SKIPCHAR); + + /* + * Handle all triggers matching \X(xy, \Xx, and \X[xxxx], where + * 'X' is the trigger. These have opaque sub-strings. + */ + case 'F': /* FALLTHROUGH */ - case ('x'): + case 'g': /* FALLTHROUGH */ - case ('S'): + case 'k': /* FALLTHROUGH */ - case ('R'): + case 'M': /* FALLTHROUGH */ - case ('N'): + case 'm': /* FALLTHROUGH */ - case ('l'): + case 'n': /* FALLTHROUGH */ - case ('L'): + case 'V': /* FALLTHROUGH */ - case ('H'): + case 'Y': + gly = ESCAPE_IGNORE; /* FALLTHROUGH */ - case ('h'): + case 'f': + if (ESCAPE_ERROR == gly) + gly = ESCAPE_FONT; + switch (**start) { + case '(': + *start = ++*end; + *sz = 2; + break; + case '[': + *start = ++*end; + term = ']'; + break; + default: + *sz = 1; + break; + } + break; + + /* + * These escapes are of the form \X'Y', where 'X' is the trigger + * and 'Y' is any string. These have opaque sub-strings. + * The \B and \w escapes are handled in roff.c, roff_res(). + */ + case 'A': /* FALLTHROUGH */ - case ('D'): + case 'b': /* FALLTHROUGH */ - case ('C'): + case 'D': /* FALLTHROUGH */ - case ('b'): + case 'o': /* FALLTHROUGH */ - case ('B'): + case 'R': /* FALLTHROUGH */ - case ('a'): + case 'X': /* FALLTHROUGH */ - case ('A'): - if (*p++ != '\'') - return(0); - term = '\''; + case 'Z': + if ('\0' == **start) + return(ESCAPE_ERROR); + gly = ESCAPE_IGNORE; + term = **start; + *start = ++*end; break; -#endif - case ('h'): + + /* + * These escapes are of the form \X'N', where 'X' is the trigger + * and 'N' resolves to a numerical expression. + */ + case 'h': /* FALLTHROUGH */ - case ('v'): + case 'H': /* FALLTHROUGH */ - case ('s'): - if (ASCII_HYPH == *p) - *p = '-'; + case 'L': + /* FALLTHROUGH */ + case 'l': + /* FALLTHROUGH */ + case 'S': + /* FALLTHROUGH */ + case 'v': + /* FALLTHROUGH */ + case 'x': + if (strchr(" %&()*+-./0123456789:<=>", **start)) { + ++*end; + return(ESCAPE_ERROR); + } + gly = ESCAPE_IGNORE; + term = **start; + *start = ++*end; + break; - i = 0; - if ('+' == *p || '-' == *p) { - p++; - i = 1; + /* + * Special handling for the numbered character escape. + * XXX Do any other escapes need similar handling? + */ + case 'N': + if ('\0' == **start) + return(ESCAPE_ERROR); + (*end)++; + if (isdigit((unsigned char)**start)) { + *sz = 1; + return(ESCAPE_IGNORE); } + (*start)++; + while (isdigit((unsigned char)**end)) + (*end)++; + *sz = *end - *start; + if ('\0' != **end) + (*end)++; + return(ESCAPE_NUMBERED); - switch (*p++) { - case ('('): - len = 2; + /* + * Sizes get a special category of their own. + */ + case 's': + gly = ESCAPE_IGNORE; + + /* See +/- counts as a sign. */ + if ('+' == **end || '-' == **end || ASCII_HYPH == **end) + (*end)++; + + switch (**end) { + case '(': + *start = ++*end; + *sz = 2; break; - case ('['): + case '[': + *start = ++*end; term = ']'; break; - case ('\''): + case '\'': + *start = ++*end; term = '\''; break; - case ('0'): - i = 1; - /* FALLTHROUGH */ default: - len = 1; - p--; + *sz = 1; break; } - if (ASCII_HYPH == *p) - *p = '-'; - if ('+' == *p || '-' == *p) { - if (i) - return(0); - p++; - } - - /* Handle embedded numerical subexp or escape. */ + break; - if ('(' == *p) { - while (*p && ')' != *p) - if ('\\' == *p++) { - i = mandoc_special(--p); - if (0 == i) - return(0); - p += i; - } + /* + * Anything else is assumed to be a glyph. + * In this case, pass back the character after the backslash. + */ + default: + gly = ESCAPE_SPECIAL; + *start = --*end; + *sz = 1; + break; + } - if (')' == *p++) - break; + assert(ESCAPE_ERROR != gly); - return(0); - } else if ('\\' == *p) { - if (0 == (i = mandoc_special(p))) - return(0); - p += i; + /* + * Read up to the terminating character, + * paying attention to nested escapes. + */ + + if ('\0' != term) { + while (**end != term) { + switch (**end) { + case '\0': + return(ESCAPE_ERROR); + case '\\': + (*end)++; + if (ESCAPE_ERROR == + mandoc_escape(end, NULL, NULL)) + return(ESCAPE_ERROR); + break; + default: + (*end)++; + break; + } } + *sz = (*end)++ - *start; + } else { + assert(*sz > 0); + if ((size_t)*sz > strlen(*start)) + return(ESCAPE_ERROR); + *end += *sz; + } - break; -#if 0 - case ('Y'): - /* FALLTHROUGH */ - case ('V'): - /* FALLTHROUGH */ - case ('$'): - /* FALLTHROUGH */ - case ('n'): - /* FALLTHROUGH */ -#endif - case ('k'): - /* FALLTHROUGH */ - case ('M'): - /* FALLTHROUGH */ - case ('m'): - /* FALLTHROUGH */ - case ('f'): - /* FALLTHROUGH */ - case ('F'): - /* FALLTHROUGH */ - case ('*'): - switch (*p++) { - case ('('): - len = 2; + /* Run post-processors. */ + + switch (gly) { + case ESCAPE_FONT: + if (2 == *sz) { + if ('C' == **start) { + /* + * Treat constant-width font modes + * just like regular font modes. + */ + (*start)++; + (*sz)--; + } else { + if ('B' == (*start)[0] && 'I' == (*start)[1]) + gly = ESCAPE_FONTBI; + break; + } + } else if (1 != *sz) break; - case ('['): - term = ']'; + + switch (**start) { + case '3': + /* FALLTHROUGH */ + case 'B': + gly = ESCAPE_FONTBOLD; break; - default: - len = 1; - p--; + case '2': + /* FALLTHROUGH */ + case 'I': + gly = ESCAPE_FONTITALIC; break; + case 'P': + gly = ESCAPE_FONTPREV; + break; + case '1': + /* FALLTHROUGH */ + case 'R': + gly = ESCAPE_FONTROMAN; + break; } break; - case ('('): - len = 2; + case ESCAPE_SPECIAL: + if (1 == *sz && 'c' == **start) + gly = ESCAPE_NOSPACE; break; - case ('['): - term = ']'; - break; - case ('z'): - len = 1; - if ('\\' == *p) { - if (0 == (i = mandoc_special(p))) - return(0); - p += i; - return(*p ? (int)(p - sv) : 0); - } - break; - case ('o'): - /* FALLTHROUGH */ - case ('w'): - if ('\'' == *p++) { - term = '\''; - break; - } - /* FALLTHROUGH */ default: - len = 1; - p--; break; } - if (term) { - for ( ; *p && term != *p; p++) - if (ASCII_HYPH == *p) - *p = '-'; - return(*p ? (int)(p - sv) : 0); - } - - for (i = 0; *p && i < len; i++, p++) - if (ASCII_HYPH == *p) - *p = '-'; - return(i == len ? (int)(p - sv) : 0); + return(gly); } - -void * -mandoc_calloc(size_t num, size_t size) -{ - void *ptr; - - ptr = calloc(num, size); - if (NULL == ptr) { - perror(NULL); - exit((int)MANDOCLEVEL_SYSERR); - } - - return(ptr); -} - - -void * -mandoc_malloc(size_t size) -{ - void *ptr; - - ptr = malloc(size); - if (NULL == ptr) { - perror(NULL); - exit((int)MANDOCLEVEL_SYSERR); - } - - return(ptr); -} - - -void * -mandoc_realloc(void *ptr, size_t size) -{ - - ptr = realloc(ptr, size); - if (NULL == ptr) { - perror(NULL); - exit((int)MANDOCLEVEL_SYSERR); - } - - return(ptr); -} - - -char * -mandoc_strdup(const char *ptr) -{ - char *p; - - p = strdup(ptr); - if (NULL == p) { - perror(NULL); - exit((int)MANDOCLEVEL_SYSERR); - } - - return(p); -} - /* * Parse a quoted or unquoted roff-style request or macro argument. * Return a pointer to the parsed argument, which is either the original * pointer or advanced by one byte in case the argument is quoted. - * Null-terminate the argument in place. + * NUL-terminate the argument in place. * Collapse pairs of quotes inside quoted arguments. * Advance the argument pointer to the next argument, - * or to the null byte terminating the argument line. + * or to the NUL byte terminating the argument line. */ char * -mandoc_getarg(char **cpp, mandocmsg msg, void *data, int ln, int *pos) +mandoc_getarg(struct mparse *parse, char **cpp, int ln, int *pos) { char *start, *cp; int quoted, pairs, white; /* Quoting can only start with a new word. */ start = *cpp; + quoted = 0; if ('"' == *start) { quoted = 1; start++; - } else - quoted = 0; + } pairs = 0; white = 0; for (cp = start; '\0' != *cp; cp++) { - /* Move left after quoted quotes and escaped backslashes. */ + + /* + * Move the following text left + * after quoted quotes and after "\\" and "\t". + */ if (pairs) cp[-pairs] = cp[0]; + if ('\\' == cp[0]) { - if ('\\' == cp[1]) { - /* Poor man's copy mode. */ + /* + * In copy mode, translate double to single + * backslashes and backslash-t to literal tabs. + */ + switch (cp[1]) { + case 't': + cp[0] = '\t'; + /* FALLTHROUGH */ + case '\\': pairs++; cp++; - } else if (0 == quoted && ' ' == cp[1]) + break; + case ' ': /* Skip escaped blanks. */ - cp++; + if (0 == quoted) + cp++; + break; + default: + break; + } } else if (0 == quoted) { if (' ' == cp[0]) { /* Unescaped blanks end unquoted args. */ @@ -343,10 +428,10 @@ mandoc_getarg(char **cpp, mandocmsg msg, void *data, i } /* Quoted argument without a closing quote. */ - if (1 == quoted && msg) - (*msg)(MANDOCERR_BADQUOTE, data, ln, *pos, NULL); + if (1 == quoted) + mandoc_msg(MANDOCERR_BADQUOTE, parse, ln, *pos, NULL); - /* Null-terminate this argument and move to the next one. */ + /* NUL-terminate this argument and move to the next one. */ if (pairs) cp[-pairs] = '\0'; if ('\0' != *cp) { @@ -357,8 +442,8 @@ mandoc_getarg(char **cpp, mandocmsg msg, void *data, i *pos += (int)(cp - start) + (quoted ? 1 : 0); *cpp = cp; - if ('\0' == *cp && msg && (white || ' ' == cp[-1])) - (*msg)(MANDOCERR_EOLNSPACE, data, ln, *pos, NULL); + if ('\0' == *cp && (white || ' ' == cp[-1])) + mandoc_msg(MANDOCERR_EOLNSPACE, parse, ln, *pos, NULL); return(start); } @@ -371,7 +456,10 @@ a2time(time_t *t, const char *fmt, const char *p) memset(&tm, 0, sizeof(struct tm)); + pp = NULL; +#ifdef HAVE_STRPTIME pp = strptime(p, fmt, &tm); +#endif if (NULL != pp && '\0' == *pp) { *t = mktime(&tm); return(1); @@ -383,12 +471,12 @@ a2time(time_t *t, const char *fmt, const char *p) static char * time2a(time_t t) { - struct tm tm; + struct tm *tm; char *buf, *p; size_t ssz; int isz; - localtime_r(&t, &tm); + tm = localtime(&t); /* * Reserve space: @@ -398,15 +486,15 @@ time2a(time_t t) */ p = buf = mandoc_malloc(10 + 4 + 4 + 1); - if (0 == (ssz = strftime(p, 10 + 1, "%B ", &tm))) + if (0 == (ssz = strftime(p, 10 + 1, "%B ", tm))) goto fail; p += (int)ssz; - if (-1 == (isz = snprintf(p, 4 + 1, "%d, ", tm.tm_mday))) + if (-1 == (isz = snprintf(p, 4 + 1, "%d, ", tm->tm_mday))) goto fail; p += isz; - if (0 == strftime(p, 4 + 1, "%Y", &tm)) + if (0 == strftime(p, 4 + 1, "%Y", tm)) goto fail; return(buf); @@ -416,20 +504,21 @@ fail: } char * -mandoc_normdate(char *in, mandocmsg msg, void *data, int ln, int pos) +mandoc_normdate(struct mparse *parse, char *in, int ln, int pos) { char *out; time_t t; if (NULL == in || '\0' == *in || 0 == strcmp(in, "$" "Mdocdate$")) { - (*msg)(MANDOCERR_NODATE, data, ln, pos, NULL); + mandoc_msg(MANDOCERR_DATE_MISSING, parse, ln, pos, NULL); time(&t); } + else if (a2time(&t, "%Y-%m-%d", in)) + t = 0; else if (!a2time(&t, "$" "Mdocdate: %b %d %Y $", in) && - !a2time(&t, "%b %d, %Y", in) && - !a2time(&t, "%Y-%m-%d", in)) { - (*msg)(MANDOCERR_BADDATE, data, ln, pos, NULL); + !a2time(&t, "%b %d, %Y", in)) { + mandoc_msg(MANDOCERR_DATE_BAD, parse, ln, pos, in); t = 0; } out = t ? time2a(t) : NULL; @@ -437,10 +526,10 @@ mandoc_normdate(char *in, mandocmsg msg, void *data, i } int -mandoc_eos(const char *p, size_t sz, int enclosed) +mandoc_eos(const char *p, size_t sz) { - const char *q; - int found; + const char *q; + int enclosed, found; if (0 == sz) return(0); @@ -448,27 +537,27 @@ mandoc_eos(const char *p, size_t sz, int enclosed) /* * End-of-sentence recognition must include situations where * some symbols, such as `)', allow prior EOS punctuation to - * propogate outward. + * propagate outward. */ - found = 0; + enclosed = found = 0; for (q = p + (int)sz - 1; q >= p; q--) { switch (*q) { - case ('\"'): + case '\"': /* FALLTHROUGH */ - case ('\''): + case '\'': /* FALLTHROUGH */ - case (']'): + case ']': /* FALLTHROUGH */ - case (')'): + case ')': if (0 == found) enclosed = 1; break; - case ('.'): + case '.': /* FALLTHROUGH */ - case ('!'): + case '!': /* FALLTHROUGH */ - case ('?'): + case '?': found = 1; break; default: @@ -479,80 +568,33 @@ mandoc_eos(const char *p, size_t sz, int enclosed) return(found && !enclosed); } -int -mandoc_hyph(const char *start, const char *c) -{ - - /* - * Choose whether to break at a hyphenated character. We only - * do this if it's free-standing within a word. - */ - - /* Skip first/last character of buffer. */ - if (c == start || '\0' == *(c + 1)) - return(0); - /* Skip first/last character of word. */ - if ('\t' == *(c + 1) || '\t' == *(c - 1)) - return(0); - if (' ' == *(c + 1) || ' ' == *(c - 1)) - return(0); - /* Skip double invocations. */ - if ('-' == *(c + 1) || '-' == *(c - 1)) - return(0); - /* Skip escapes. */ - if ('\\' == *(c - 1)) - return(0); - - return(1); -} - /* - * Check if a string is a punctuation delimiter. This only applies to - * mdoc(7) documents, but as it's used in both front-ends and back-ends, - * it needs to go here (instead of, say, in libmdoc.h). + * Convert a string to a long that may not be <0. + * If the string is invalid, or is less than 0, return -1. */ -enum mdelim -mandoc_isdelim(const char *p) +int +mandoc_strntoi(const char *p, size_t sz, int base) { + char buf[32]; + char *ep; + long v; - if ('\0' == p[0]) - return(DELIM_NONE); + if (sz > 31) + return(-1); - if ('\0' == p[1]) - switch (p[0]) { - case('('): - /* FALLTHROUGH */ - case('['): - return(DELIM_OPEN); - case('|'): - return(DELIM_MIDDLE); - case('.'): - /* FALLTHROUGH */ - case(','): - /* FALLTHROUGH */ - case(';'): - /* FALLTHROUGH */ - case(':'): - /* FALLTHROUGH */ - case('?'): - /* FALLTHROUGH */ - case('!'): - /* FALLTHROUGH */ - case(')'): - /* FALLTHROUGH */ - case(']'): - return(DELIM_CLOSE); - default: - return(DELIM_NONE); - } + memcpy(buf, p, sz); + buf[(int)sz] = '\0'; - if ('\\' != p[0]) - return(DELIM_NONE); + errno = 0; + v = strtol(buf, &ep, base); - if (0 == strcmp(p, "\\.")) - return(DELIM_CLOSE); - if (0 == strcmp(p, "\\*(Ba")) - return(DELIM_MIDDLE); + if (buf[0] == '\0' || *ep != '\0') + return(-1); - return(DELIM_NONE); + if (v > INT_MAX) + v = INT_MAX; + if (v < INT_MIN) + v = INT_MIN; + + return((int)v); }