From mboxrd@z Thu Jan 1 00:00:00 1970 Received: from smtp-2.sys.kth.se (smtp-2.sys.kth.se [130.237.32.160]) by krisdoz.my.domain (8.14.3/8.14.3) with ESMTP id p38BJNJe006192 for ; Fri, 8 Apr 2011 07:19:24 -0400 (EDT) Received: from mailscan-1.sys.kth.se (mailscan-1.sys.kth.se [130.237.32.91]) by smtp-2.sys.kth.se (Postfix) with ESMTP id ED90C14DB0D for ; Fri, 8 Apr 2011 13:19:17 +0200 (CEST) X-Virus-Scanned: by amavisd-new at kth.se Received: from smtp-2.sys.kth.se ([130.237.32.160]) by mailscan-1.sys.kth.se (mailscan-1.sys.kth.se [130.237.32.91]) (amavisd-new, port 10024) with LMTP id wMy7MgfShA0Q for ; Fri, 8 Apr 2011 13:19:13 +0200 (CEST) X-KTH-Auth: kristaps [193.10.49.5] X-KTH-mail-from: kristaps@bsd.lv X-KTH-rcpt-to: tech@mdocml.bsd.lv Received: from [172.16.18.84] (unknown [193.10.49.5]) by smtp-2.sys.kth.se (Postfix) with ESMTP id 5BB1B14C113 for ; Fri, 8 Apr 2011 13:19:11 +0200 (CEST) Message-ID: <4D9EEF2F.2030307@bsd.lv> Date: Fri, 08 Apr 2011 13:19:11 +0200 From: Kristaps Dzonsons User-Agent: Mozilla/5.0 (X11; U; Linux x86_64; en-US; rv:1.9.1.16) Gecko/20110303 Icedove/3.0.11 X-Mailinglist: mdocml-tech Reply-To: tech@mdocml.bsd.lv MIME-Version: 1.0 To: tech@mdocml.bsd.lv Subject: Re: Unifying the escape-sequence parser. References: <4D9DC396.9010504@bsd.lv> In-Reply-To: <4D9DC396.9010504@bsd.lv> Content-Type: multipart/mixed; boundary="------------020603040701080702050109" This is a multi-part message in MIME format. --------------020603040701080702050109 Content-Type: text/plain; charset=ISO-8859-1; format=flowed Content-Transfer-Encoding: 7bit > Enclosed are my efforts to unify the escaped-sequence functions in out.c > (a2roffdeco()) and mandoc.c (mandoc_special()). > > This handles, as far as I can see, all syntaxes of the groff(7) escapes. > > When called during libmandoc validation, it will check for GLYPH_ERROR > and be followed by a search-and-replace of ASCII_HYPH for `-' in the > substring. When invoked from term.c or html.c, it will switch on the > returned type and substring value. > > This will clear up a nice big chunk of code, but it's a pretty delicate > area, so please look it over! > > If you compile this file, you can test escapes by running, e.g., > > % ./a.out s+\'\(\\f\[asdf\]\)\'123 > > to see the values of "start" and "end". Step 2. This finishes off the new escape-sequence parser and puts it into mandoc_escape (mandoc.c, mandoc.h), then makes it the underlying engine for a2roffdeco (out.c) (requiring a tiny change to term.c and html.c for bailing out on bad sequences) and roff_res (roff.c). Now all escape-sequences are being parsed with the same engine! This logic was being repeated in THREE different places, earlier (mandoc.c for validation, out.c for output, and roff.c for predefined escapes). I've run this over all manuals I know of without problems, but it can really use a close look-over with border cases. The next step is to clean out the out.c code, completely removing enum roffdeco (putting that logic into mandoc.c, perhaps). Thoughts? Kristaps --------------020603040701080702050109 Content-Type: text/plain; name="patch.escapes.txt" Content-Transfer-Encoding: 7bit Content-Disposition: attachment; filename="patch.escapes.txt" Index: html.c =================================================================== RCS file: /usr/vhosts/mdocml.bsd.lv/cvs/mdocml/html.c,v retrieving revision 1.131 diff -u -r1.131 html.c --- html.c 22 Mar 2011 14:05:45 -0000 1.131 +++ html.c 8 Apr 2011 11:16:10 -0000 @@ -337,7 +337,8 @@ break; seq = ++p; - len = a2roffdeco(&deco, &seq, &sz); + if (0 == (len = a2roffdeco(&deco, &seq, &sz))) + break; switch (deco) { case (DECO_NUMBERED): Index: libmandoc.h =================================================================== RCS file: /usr/vhosts/mdocml.bsd.lv/cvs/mdocml/libmandoc.h,v retrieving revision 1.17 diff -u -r1.17 libmandoc.h --- libmandoc.h 28 Mar 2011 23:52:13 -0000 1.17 +++ libmandoc.h 8 Apr 2011 11:16:10 -0000 @@ -73,7 +73,6 @@ int, int, const char *); void mandoc_vmsg(enum mandocerr, struct mparse *, int, int, const char *, ...); -int mandoc_special(char *); char *mandoc_strdup(const char *); char *mandoc_getarg(struct mparse *, char **, int, int *); char *mandoc_normdate(struct mparse *, char *, int, int); Index: man_validate.c =================================================================== RCS file: /usr/vhosts/mdocml.bsd.lv/cvs/mdocml/man_validate.c,v retrieving revision 1.67 diff -u -r1.67 man_validate.c --- man_validate.c 22 Mar 2011 15:30:30 -0000 1.67 +++ man_validate.c 8 Apr 2011 11:16:11 -0000 @@ -54,7 +54,7 @@ static int check_part(CHKARGS); static int check_root(CHKARGS); static int check_sec(CHKARGS); -static int check_text(CHKARGS); +static void check_text(CHKARGS); static int post_AT(CHKARGS); static int post_fi(CHKARGS); @@ -151,7 +151,8 @@ switch (m->last->type) { case (MAN_TEXT): - return(check_text(m, m->last)); + check_text(m, m->last); + return(1); case (MAN_ROOT): return(check_root(m, m->last)); case (MAN_EQN): @@ -204,43 +205,48 @@ return(1); } - -static int +static void check_text(CHKARGS) { - char *p; - int pos, c; + char *p, *pp, *cpp; + int pos; size_t sz; - for (p = n->string, pos = n->pos + 1; *p; p++, pos++) { - sz = strcspn(p, "\t\\"); - p += (int)sz; + p = n->string; + pos = n->pos + 1; - if ('\0' == *p) - break; + while ('\0' != *p) { + sz = strcspn(p, "\t\\"); + p += (int)sz; pos += (int)sz; if ('\t' == *p) { - if (MAN_LITERAL & m->flags) - continue; - man_pmsg(m, n->line, pos, MANDOCERR_BADTAB); + if ( ! (MAN_LITERAL & m->flags)) + man_pmsg(m, n->line, pos, MANDOCERR_BADTAB); + p++; + pos++; continue; - } + } else if ('\0' == *p) + break; - /* Check the special character. */ + pos++; + pp = ++p; - c = mandoc_special(p); - if (c) { - p += c - 1; - pos += c - 1; - } else + if (ESCAPE_ERROR == mandoc_escape + ((const char **)&pp, NULL, NULL)) { man_pmsg(m, n->line, pos, MANDOCERR_BADESCAPE); - } + break; + } - return(1); -} + cpp = p; + while (NULL != (cpp = memchr(cpp, ASCII_HYPH, pp - cpp))) + *cpp = '-'; + pos += pp - p; + p = pp; + } +} #define INEQ_DEFINE(x, ineq, name) \ static int \ Index: mandoc.c =================================================================== RCS file: /usr/vhosts/mdocml.bsd.lv/cvs/mdocml/mandoc.c,v retrieving revision 1.44 diff -u -r1.44 mandoc.c --- mandoc.c 28 Mar 2011 23:52:13 -0000 1.44 +++ mandoc.c 8 Apr 2011 11:16:11 -0000 @@ -35,198 +35,315 @@ static int a2time(time_t *, const char *, const char *); static char *time2a(time_t); +static int numescape(const char *); -int -mandoc_special(char *p) +/* + * Pass over recursive numerical expressions. This context of this + * function is important: it's only called within character-terminating + * escapes (e.g., \s[xxxyyy]), so all we need to do is handle initial + * recursion: we don't care about what's in these blocks. + * This returns the number of characters skipped or -1 if an error + * occurs (the caller should bail). + */ +static int +numescape(const char *start) { - int len, i; - char term; - char *sv; - - len = 0; + int i; + size_t sz; + const char *cp; + + i = 0; + + /* The expression consists of a subexpression. */ + + if ('\\' == start[i]) { + cp = &start[++i]; + /* + * Read past the end of the subexpression. + * Bail immediately on errors. + */ + if (ESCAPE_ERROR == mandoc_escape(&cp, NULL, NULL)) + return(-1); + return(i + cp - &start[i]); + } + + if ('(' != start[i++]) + return(0); + + /* + * A parenthesised subexpression. Read until the closing + * parenthesis, making sure to handle any nested subexpressions + * that might ruin our parse. + */ + + while (')' != start[i]) { + sz = strcspn(&start[i], ")\\"); + i += (int)sz; + + if ('\0' == start[i]) + return(-1); + else if ('\\' != start[i]) + continue; + + cp = &start[++i]; + if (ESCAPE_ERROR == mandoc_escape(&cp, NULL, NULL)) + return(-1); + i += cp - &start[i]; + } + + /* Read past the terminating ')'. */ + return(++i); +} + +/* + * Handle an escaped sequeence. This should be called with any + * string subsequent a `\'. Pass a pointer to this substring as "end"; + * it will be set to the supremum of the parsed escape sequence. If + * this returns ESCAPE_ERROR, the string is bogus and should be thrown + * away. If not ESCAPE_ERROR or ESCAPE_IGNORE, "start" is set to the + * first relevant character of the substring (font, glyph, whatever) of + * length sz. Both "start" and "sz" may be NULL. + */ +enum mandoc_esc +mandoc_escape(const char **end, const char **start, int *sz) +{ + char c, term, numeric; + int i, lim, ssz; + const char *cp; + enum mandoc_esc gly; + + cp = *end; + if (start) + *start = cp; + i = 0; + gly = ESCAPE_ERROR; term = '\0'; - sv = p; + numeric = 0; - assert('\\' == *p); - p++; + switch ((c = cp[i++])) { + /* + * First the glyphs. There are several different forms of + * these, but each eventually returns a substring of the glyph + * name. + */ + case ('('): + gly = ESCAPE_SPECIAL; + lim = 2; + break; + case ('['): + gly = ESCAPE_SPECIAL; + term = ']'; + break; + case ('C'): + if ('\'' != cp[i]) + return(ESCAPE_ERROR); + gly = ESCAPE_SPECIAL; + term = '\''; + break; - switch (*p++) { -#if 0 - case ('Z'): - /* FALLTHROUGH */ - case ('X'): - /* FALLTHROUGH */ - case ('x'): - /* FALLTHROUGH */ - case ('S'): - /* FALLTHROUGH */ - case ('R'): - /* FALLTHROUGH */ - case ('N'): - /* FALLTHROUGH */ - case ('l'): - /* FALLTHROUGH */ - case ('L'): - /* FALLTHROUGH */ - case ('H'): + /* + * Handle all triggers matching \X(xy, \Xx, and \X[xxxx], where + * 'X' is the trigger. These have opaque sub-strings. + */ + case ('g'): /* FALLTHROUGH */ - case ('h'): + case ('k'): /* FALLTHROUGH */ - case ('D'): + case ('M'): /* FALLTHROUGH */ - case ('C'): + case ('m'): /* FALLTHROUGH */ - case ('b'): + case ('n'): /* FALLTHROUGH */ - case ('B'): + case ('V'): /* FALLTHROUGH */ - case ('a'): + case ('Y'): + if (ESCAPE_ERROR == gly) + gly = ESCAPE_IGNORE; /* FALLTHROUGH */ - case ('A'): - if (*p++ != '\'') - return(0); - term = '\''; - break; -#endif - case ('h'): + case ('*'): + if (ESCAPE_ERROR == gly) + gly = ESCAPE_PREDEF; /* FALLTHROUGH */ - case ('v'): + case ('F'): + if (ESCAPE_ERROR == gly) + gly = ESCAPE_FONTFAM; /* FALLTHROUGH */ - case ('s'): - if (ASCII_HYPH == *p) - *p = '-'; + case ('f'): + if (ESCAPE_ERROR == gly) + gly = ESCAPE_FONT; - i = 0; - if ('+' == *p || '-' == *p) { - p++; - i = 1; - } + if (start) + *start = &cp[i]; - switch (*p++) { + switch (cp[i++]) { case ('('): - len = 2; + lim = 2; break; case ('['): term = ']'; break; - case ('\''): - term = '\''; - break; - case ('0'): - i = 1; - /* FALLTHROUGH */ default: - len = 1; - p--; + lim = 1; + i--; break; } + break; - if (ASCII_HYPH == *p) - *p = '-'; - if ('+' == *p || '-' == *p) { - if (i) - return(0); - p++; - } - - /* Handle embedded numerical subexp or escape. */ - - if ('(' == *p) { - while (*p && ')' != *p) - if ('\\' == *p++) { - i = mandoc_special(--p); - if (0 == i) - return(0); - p += i; - } - - if (')' == *p++) - break; - - return(0); - } else if ('\\' == *p) { - if (0 == (i = mandoc_special(p))) - return(0); - p += i; - } - + /* + * These escapes are of the form \X'Y', where 'X' is the trigger + * and 'Y' is any string. These have opaque sub-strings. + */ + case ('A'): + /* FALLTHROUGH */ + case ('b'): + /* FALLTHROUGH */ + case ('D'): + /* FALLTHROUGH */ + case ('o'): + /* FALLTHROUGH */ + case ('R'): + /* FALLTHROUGH */ + case ('X'): + /* FALLTHROUGH */ + case ('Z'): + if ('\'' != cp[i++]) + return(ESCAPE_ERROR); + gly = ESCAPE_IGNORE; + term = '\''; break; -#if 0 - case ('Y'): + + /* + * These escapes are of the form \X'N', where 'X' is the trigger + * and 'N' resolves to a numerical expression. + */ + case ('B'): /* FALLTHROUGH */ - case ('V'): + case ('h'): /* FALLTHROUGH */ - case ('$'): + case ('H'): /* FALLTHROUGH */ - case ('n'): + case ('L'): /* FALLTHROUGH */ -#endif - case ('k'): + case ('l'): /* FALLTHROUGH */ - case ('M'): + case ('N'): + if (ESCAPE_ERROR == gly) + gly = ESCAPE_NUMBERED; /* FALLTHROUGH */ - case ('m'): + case ('S'): /* FALLTHROUGH */ - case ('f'): + case ('v'): /* FALLTHROUGH */ - case ('F'): + case ('w'): /* FALLTHROUGH */ - case ('*'): - switch (*p++) { + case ('x'): + if (ESCAPE_ERROR == gly) + gly = ESCAPE_IGNORE; + if ('\'' != cp[i++]) + return(ESCAPE_ERROR); + term = numeric = '\''; + break; + + /* + * Sizes get a special category of their own. + */ + case ('s'): + gly = ESCAPE_IGNORE; + + if (start) + *start = &cp[i]; + + /* See +/- counts as a sign. */ + c = cp[i]; + if ('+' == c || '-' == c || ASCII_HYPH == c) + ++i; + + switch (cp[i++]) { case ('('): - len = 2; + lim = 2; break; case ('['): - term = ']'; + term = numeric = ']'; + break; + case ('\''): + term = numeric = '\''; break; default: - len = 1; - p--; + lim = 1; + i--; break; } + + /* See +/- counts as a sign. */ + c = cp[i]; + if ('+' == c || '-' == c || ASCII_HYPH == c) + ++i; + break; - case ('('): - len = 2; - 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 */ + + /* + * Anything else is assumed to be a glyph. + */ default: - len = 1; - p--; + gly = ESCAPE_SPECIAL; + lim = 1; + i--; break; } - if (term) { - for ( ; *p && term != *p; p++) - if (ASCII_HYPH == *p) - *p = '-'; - return(*p ? (int)(p - sv) : 0); + assert(ESCAPE_ERROR != gly); + + if (start) + *start = &cp[i]; + + /* + * If a terminating block has been specified, we need to + * handle the case of recursion, which could have their + * own terminating blocks that mess up our parse. This, by the + * way, means that the "start" and "size" values will be + * effectively meaningless. + */ + + ssz = 0; + if (numeric && -1 == (ssz = numescape(&cp[i]))) + return(ESCAPE_ERROR); + + i += ssz; + + /* + * We have a character terminator. Try to read up to that + * character. If we can't (i.e., we hit the nil), then return + * an error; if we can, calculate our length, read past the + * terminating character, and exit. + */ + + if ('\0' != term) { + *end = strchr(&cp[i], term); + if ('\0' == *end) + return(ESCAPE_ERROR); + if (sz) + *sz = *end - &cp[i]; + (*end)++; + return(gly); } - for (i = 0; *p && i < len; i++, p++) - if (ASCII_HYPH == *p) - *p = '-'; - return(i == len ? (int)(p - sv) : 0); -} + assert(lim > 0); + /* + * We have a numeric limit. If the string is shorter than that, + * stop and return an error. Else adjust our endpoint, length, + * and return the current glyph. + */ + + if ((size_t)lim > strlen(&cp[i])) + return(ESCAPE_ERROR); + + if (sz) + *sz = lim; + *end = &cp[i] + lim; + return(gly); +} void * mandoc_calloc(size_t num, size_t size) Index: mandoc.h =================================================================== RCS file: /usr/vhosts/mdocml.bsd.lv/cvs/mdocml/mandoc.h,v retrieving revision 1.69 diff -u -r1.69 mandoc.h --- mandoc.h 28 Mar 2011 21:49:42 -0000 1.69 +++ mandoc.h 8 Apr 2011 11:16:11 -0000 @@ -288,6 +288,16 @@ MPARSE_MAN /* assume -man */ }; +enum mandoc_esc { + ESCAPE_ERROR = 0, + ESCAPE_IGNORE, /* escape to be ignored */ + ESCAPE_SPECIAL, /* a regular special character */ + ESCAPE_PREDEF, /* a predefined special character */ + ESCAPE_FONT, /* a font mode */ + ESCAPE_FONTFAM, /* a font family */ + ESCAPE_NUMBERED /* a numbered glyph */ +}; + typedef void (*mandocmsg)(enum mandocerr, enum mandoclevel, const char *, int, int, const char *); @@ -309,6 +319,8 @@ void *mandoc_calloc(size_t, size_t); void *mandoc_malloc(size_t); void *mandoc_realloc(void *, size_t); + +enum mandoc_esc mandoc_escape(const char **, const char **, int *); __END_DECLS Index: out.c =================================================================== RCS file: /usr/vhosts/mdocml.bsd.lv/cvs/mdocml/out.c,v retrieving revision 1.39 diff -u -r1.39 out.c --- out.c 17 Mar 2011 08:49:34 -0000 1.39 +++ out.c 8 Apr 2011 11:16:11 -0000 @@ -178,237 +178,70 @@ int a2roffdeco(enum roffdeco *d, const char **word, size_t *sz) { - int i, j, lim; - char term, c; - const char *wp; - enum roffdeco dd; + const char *cp, *start; + int ssz; + enum mandoc_esc esc; *d = DECO_NONE; - lim = i = 0; - term = '\0'; - wp = *word; - switch ((c = wp[i++])) { - case ('('): + cp = start = *word; + + esc = mandoc_escape(&cp, word, &ssz); + + switch (esc) { + case (ESCAPE_ERROR): + return(0); + case (ESCAPE_IGNORE): + break; + case (ESCAPE_NUMBERED): + *d = DECO_NUMBERED; + break; + case (ESCAPE_FONT): + *d = DECO_FONT; + break; + case (ESCAPE_FONTFAM): + *d = DECO_FFONT; + break; + case (ESCAPE_SPECIAL): *d = DECO_SPECIAL; - lim = 2; break; - case ('F'): - /* FALLTHROUGH */ - case ('f'): - *d = 'F' == c ? DECO_FFONT : DECO_FONT; - - switch (wp[i++]) { - case ('('): - lim = 2; - break; - case ('['): - term = ']'; - break; + case (ESCAPE_PREDEF): + *d = DECO_RESERVED; + break; + } + + assert(ssz >= 0); + *sz = (size_t)ssz; + ssz = cp - start; + + if (1 == *sz && (DECO_FONT == *d || DECO_FFONT == *d)) + switch (**word) { case ('3'): /* FALLTHROUGH */ case ('B'): *d = DECO_BOLD; - return(i); + break; case ('2'): /* FALLTHROUGH */ case ('I'): *d = DECO_ITALIC; - return(i); + break; case ('P'): *d = DECO_PREVIOUS; - return(i); + break; case ('1'): /* FALLTHROUGH */ case ('R'): *d = DECO_ROMAN; - return(i); - default: - i--; - lim = 1; - break; - } - break; - case ('k'): - /* FALLTHROUGH */ - case ('M'): - /* FALLTHROUGH */ - case ('m'): - /* FALLTHROUGH */ - case ('*'): - if ('*' == c) - *d = DECO_RESERVED; - - switch (wp[i++]) { - case ('('): - lim = 2; - break; - case ('['): - term = ']'; - break; - default: - i--; - lim = 1; - break; - } - break; - - case ('N'): - - /* - * Sequence of characters: backslash, 'N' (i = 0), - * starting delimiter (i = 1), character number (i = 2). - */ - - *word = wp + 2; - *sz = 0; - - /* - * Cannot use a digit as a starting delimiter; - * but skip the digit anyway. - */ - - if (isdigit((int)wp[1])) - return(2); - - /* - * Any non-digit terminates the character number. - * That is, the terminating delimiter need not - * match the starting delimiter. - */ - - for (i = 2; isdigit((int)wp[i]); i++) - (*sz)++; - - /* - * This is only a numbered character - * if the character number has at least one digit. - */ - - if (*sz) - *d = DECO_NUMBERED; - - /* - * Skip the terminating delimiter, even if it does not - * match, and even if there is no character number. - */ - - return(++i); - - case ('h'): - /* FALLTHROUGH */ - case ('v'): - /* FALLTHROUGH */ - case ('s'): - j = 0; - if ('+' == wp[i] || '-' == wp[i]) { - i++; - j = 1; - } - - switch (wp[i++]) { - case ('('): - lim = 2; - break; - case ('['): - term = ']'; break; - case ('\''): - term = '\''; - break; - case ('0'): - j = 1; - /* FALLTHROUGH */ default: - i--; - lim = 1; - break; - } - - if ('+' == wp[i] || '-' == wp[i]) { - if (j) - return(i); - i++; - } - - /* Handle embedded numerical subexp or escape. */ - - if ('(' == wp[i]) { - while (wp[i] && ')' != wp[i]) - if ('\\' == wp[i++]) { - /* Handle embedded escape. */ - *word = &wp[i]; - i += a2roffdeco(&dd, word, sz); - } - - if (')' == wp[i++]) - break; - - *d = DECO_NONE; - return(i - 1); - } else if ('\\' == wp[i]) { - *word = &wp[++i]; - i += a2roffdeco(&dd, word, sz); - } - - break; - case ('['): - *d = DECO_SPECIAL; - term = ']'; - break; - case ('c'): - *d = DECO_NOSPACE; - return(i); - case ('z'): - *d = DECO_NONE; - if ('\\' == wp[i]) { - *word = &wp[++i]; - return(i + a2roffdeco(&dd, word, sz)); - } else - lim = 1; - break; - case ('o'): - /* FALLTHROUGH */ - case ('w'): - if ('\'' == wp[i++]) { - term = '\''; break; - } - /* FALLTHROUGH */ - default: - *d = DECO_SSPECIAL; - i--; - lim = 1; - break; - } - - assert(term || lim); - *word = &wp[i]; - - if (term) { - j = i; - while (wp[i] && wp[i] != term) - i++; - if ('\0' == wp[i]) { - *d = DECO_NONE; - return(i); } - assert(i >= j); - *sz = (size_t)(i - j); - - return(i + 1); - } - - assert(lim > 0); - *sz = (size_t)lim; - - for (j = 0; wp[i] && j < lim; j++) - i++; - if (j < lim) - *d = DECO_NONE; + if (1 == *sz && DECO_SPECIAL == *d) + *d = 'c' == **word ? DECO_NOSPACE : DECO_SSPECIAL; - return(i); + return(ssz); } /* Index: read.c =================================================================== RCS file: /usr/vhosts/mdocml.bsd.lv/cvs/mdocml/read.c,v retrieving revision 1.11 diff -u -r1.11 read.c --- read.c 4 Apr 2011 23:04:38 -0000 1.11 +++ read.c 8 Apr 2011 11:16:11 -0000 @@ -142,7 +142,7 @@ "tab in non-literal context", "end of line whitespace", "bad comment style", - "unknown escape sequence", + "bad escape sequence", "unterminated quoted string", "generic error", Index: roff.c =================================================================== RCS file: /usr/vhosts/mdocml.bsd.lv/cvs/mdocml/roff.c,v retrieving revision 1.131 diff -u -r1.131 roff.c --- roff.c 5 Apr 2011 22:22:33 -0000 1.131 +++ roff.c 8 Apr 2011 11:16:11 -0000 @@ -382,9 +382,10 @@ const char *stnam; /* start of the name, after "[(*" */ const char *cp; /* end of the name, e.g. before ']' */ const char *res; /* the string to be substituted */ - int i, maxl; + int i; size_t nsz; char *n; + enum mandoc_esc esc; /* Search for a leading backslash and save a pointer to it. */ @@ -392,48 +393,11 @@ while (NULL != (cp = strchr(cp, '\\'))) { stesc = cp++; - /* - * The second character must be an asterisk. - * If it isn't, skip it anyway: It is escaped, - * so it can't start another escape sequence. - */ - - if ('\0' == *cp) - return(1); - if ('*' != *cp++) - continue; - - /* - * The third character decides the length - * of the name of the string. - * Save a pointer to the name. - */ - - switch (*cp) { - case ('\0'): - return(1); - case ('('): - cp++; - maxl = 2; - break; - case ('['): - cp++; - maxl = 0; - break; - default: - maxl = 1; + esc = mandoc_escape(&cp, &stnam, &i); + if (ESCAPE_ERROR == esc) break; - } - stnam = cp; - - /* Advance to the end of the name. */ - - for (i = 0; 0 == maxl || i < maxl; i++, cp++) { - if ('\0' == *cp) - return(1); /* Error. */ - if (0 == maxl && ']' == *cp) - break; - } + else if (ESCAPE_PREDEF != esc) + continue; /* * Retrieve the replacement string; if it is @@ -442,10 +406,8 @@ res = roff_getstrn(r, stnam, (size_t)i); - if (NULL == res) { - cp -= maxl ? 1 : 0; + if (NULL == res) continue; - } /* Replace the escape sequence by the string. */ @@ -454,7 +416,7 @@ strlcpy(n, *bufp, (size_t)(stesc - *bufp + 1)); strlcat(n, res, nsz); - strlcat(n, cp + (maxl ? 0 : 1), nsz); + strlcat(n, cp, nsz); free(*bufp); Index: term.c =================================================================== RCS file: /usr/vhosts/mdocml.bsd.lv/cvs/mdocml/term.c,v retrieving revision 1.183 diff -u -r1.183 term.c --- term.c 4 Apr 2011 21:14:12 -0000 1.183 +++ term.c 8 Apr 2011 11:16:11 -0000 @@ -457,6 +457,7 @@ term_word(struct termp *p, const char *word) { const char *seq; + int sz; size_t ssz; enum roffdeco deco; @@ -487,7 +488,9 @@ continue; seq = ++word; - word += a2roffdeco(&deco, &seq, &ssz); + if (0 == (sz = a2roffdeco(&deco, &seq, &ssz))) + break; + word += sz; switch (deco) { case (DECO_NUMBERED): --------------020603040701080702050109-- -- To unsubscribe send an email to tech+unsubscribe@mdocml.bsd.lv