=================================================================== RCS file: /cvs/mandoc/mdoc.c,v retrieving revision 1.184 retrieving revision 1.193 diff -u -p -r1.184 -r1.193 --- mandoc/mdoc.c 2011/03/17 11:30:23 1.184 +++ mandoc/mdoc.c 2011/07/27 12:41:02 1.193 @@ -1,4 +1,4 @@ -/* $Id: mdoc.c,v 1.184 2011/03/17 11:30:23 kristaps Exp $ */ +/* $Id: mdoc.c,v 1.193 2011/07/27 12:41:02 kristaps Exp $ */ /* * Copyright (c) 2008, 2009, 2010, 2011 Kristaps Dzonsons * Copyright (c) 2010 Ingo Schwarze @@ -28,6 +28,7 @@ #include #include +#include "mdoc.h" #include "mandoc.h" #include "libmdoc.h" #include "libmandoc.h" @@ -96,6 +97,9 @@ static struct mdoc_node *node_alloc(struct mdoc *, int enum mdoct, enum mdoc_type); static int node_append(struct mdoc *, struct mdoc_node *); +#if 0 +static int mdoc_preptext(struct mdoc *, int, char *, int); +#endif static int mdoc_ptext(struct mdoc *, int, char *, int); static int mdoc_pmacro(struct mdoc *, int, char *, int); @@ -192,15 +196,14 @@ mdoc_free(struct mdoc *mdoc) * Allocate volatile and non-volatile parse resources. */ struct mdoc * -mdoc_alloc(struct regset *regs, void *data, mandocmsg msg) +mdoc_alloc(struct roff *roff, struct mparse *parse) { struct mdoc *p; p = mandoc_calloc(1, sizeof(struct mdoc)); - p->msg = msg; - p->data = data; - p->regs = regs; + p->parse = parse; + p->roff = roff; mdoc_hash_init(); mdoc_alloc1(p); @@ -233,11 +236,11 @@ mdoc_addeqn(struct mdoc *m, const struct eqn *ep) /* No text before an initial macro. */ if (SEC_NONE == m->lastnamed) { - mdoc_pmsg(m, ep->line, ep->pos, MANDOCERR_NOTEXT); + mdoc_pmsg(m, ep->ln, ep->pos, MANDOCERR_NOTEXT); return(1); } - n = node_alloc(m, ep->line, ep->pos, MDOC_MAX, MDOC_EQN); + n = node_alloc(m, ep->ln, ep->pos, MDOC_MAX, MDOC_EQN); n->eqn = ep; if ( ! node_append(m, n)) @@ -290,34 +293,18 @@ mdoc_parseln(struct mdoc *m, int ln, char *buf, int of * whether this mode is on or off. * Note that this mode is also switched by the Sh macro. */ - if (m->regs->regs[(int)REG_nS].set) { - if (m->regs->regs[(int)REG_nS].v.u) + if (roff_regisset(m->roff, REG_nS)) { + if (roff_regget(m->roff, REG_nS)) m->flags |= MDOC_SYNOPSIS; else m->flags &= ~MDOC_SYNOPSIS; } - return(('.' == buf[offs] || '\'' == buf[offs]) ? + return(mandoc_getcontrol(buf, &offs) ? mdoc_pmacro(m, ln, buf, offs) : mdoc_ptext(m, ln, buf, offs)); } - -void -mdoc_vmsg(struct mdoc *mdoc, enum mandocerr t, - int ln, int pos, const char *fmt, ...) -{ - char buf[256]; - va_list ap; - - va_start(ap, fmt); - vsnprintf(buf, sizeof(buf) - 1, fmt, ap); - va_end(ap); - - (*mdoc->msg)(t, mdoc->data, ln, pos, buf); -} - - int mdoc_macro(MACRO_PROT_ARGS) { @@ -345,8 +332,8 @@ mdoc_macro(MACRO_PROT_ARGS) if (NULL == m->meta.os) m->meta.os = mandoc_strdup("LOCAL"); if (NULL == m->meta.date) - m->meta.date = mandoc_normdate(NULL, - m->msg, m->data, line, ppos); + m->meta.date = mandoc_normdate + (m->parse, NULL, line, ppos); m->flags |= MDOC_PBODY; } @@ -666,7 +653,60 @@ mdoc_node_delete(struct mdoc *m, struct mdoc_node *p) mdoc_node_free(p); } +#if 0 +/* + * Pre-treat a text line. + * Text lines can consist of equations, which must be handled apart from + * the regular text. + * Thus, use this function to step through a line checking if it has any + * equations embedded in it. + * This must handle multiple equations AND equations that do not end at + * the end-of-line, i.e., will re-enter in the next roff parse. + */ +static int +mdoc_preptext(struct mdoc *m, int line, char *buf, int offs) +{ + char *start, *end; + char delim; + while ('\0' != buf[offs]) { + /* Mark starting position if eqn is set. */ + start = NULL; + if ('\0' != (delim = roff_eqndelim(m->roff))) + if (NULL != (start = strchr(buf + offs, delim))) + *start++ = '\0'; + + /* Parse text as normal. */ + if ( ! mdoc_ptext(m, line, buf, offs)) + return(0); + + /* Continue only if an equation exists. */ + if (NULL == start) + break; + + /* Read past the end of the equation. */ + offs += start - (buf + offs); + assert(start == &buf[offs]); + if (NULL != (end = strchr(buf + offs, delim))) { + *end++ = '\0'; + while (' ' == *end) + end++; + } + + /* Parse the equation itself. */ + roff_openeqn(m->roff, NULL, line, offs, buf); + + /* Process a finished equation? */ + if (roff_closeeqn(m->roff)) + if ( ! mdoc_addeqn(m, roff_eqn(m->roff))) + return(0); + offs += (end - (buf + offs)); + } + + return(1); +} +#endif + /* * Parse free-form text, that is, a line that does not begin with the * control character. @@ -677,15 +717,6 @@ mdoc_ptext(struct mdoc *m, int line, char *buf, int of char *c, *ws, *end; struct mdoc_node *n; - /* Ignore bogus comments. */ - - if ('\\' == buf[offs] && - '.' == buf[offs + 1] && - '"' == buf[offs + 2]) { - mdoc_pmsg(m, line, offs, MANDOCERR_BADCOMMENT); - return(1); - } - /* No text before an initial macro. */ if (SEC_NONE == m->lastnamed) { @@ -728,11 +759,6 @@ mdoc_ptext(struct mdoc *m, int line, char *buf, int of ws = NULL; for (c = end = buf + offs; *c; c++) { switch (*c) { - case '-': - if (mandoc_hyph(buf + offs, c)) - *c = ASCII_HYPH; - ws = NULL; - break; case ' ': if (NULL == ws) ws = c; @@ -812,64 +838,57 @@ static int mdoc_pmacro(struct mdoc *m, int ln, char *buf, int offs) { enum mdoct tok; - int i, j, sv; + int i, sv; char mac[5]; struct mdoc_node *n; - /* Empty lines are ignored. */ + /* Empty post-control lines are ignored. */ - offs++; - - if ('\0' == buf[offs]) + if ('"' == buf[offs]) { + mdoc_pmsg(m, ln, offs, MANDOCERR_BADCOMMENT); return(1); + } else if ('\0' == buf[offs]) + return(1); - i = offs; + sv = offs; - /* Accept tabs/whitespace after the initial control char. */ - - if (' ' == buf[i] || '\t' == buf[i]) { - i++; - while (buf[i] && (' ' == buf[i] || '\t' == buf[i])) - i++; - if ('\0' == buf[i]) - return(1); - } - - sv = i; - /* * Copy the first word into a nil-terminated buffer. * Stop copying when a tab, space, or eoln is encountered. */ - j = 0; - while (j < 4 && '\0' != buf[i] && ' ' != buf[i] && '\t' != buf[i]) - mac[j++] = buf[i++]; - mac[j] = '\0'; + i = 0; + while (i < 4 && '\0' != buf[offs] && + ' ' != buf[offs] && '\t' != buf[offs]) + mac[i++] = buf[offs++]; - tok = (j > 1 || j < 4) ? mdoc_hash_find(mac) : MDOC_MAX; + mac[i] = '\0'; + + tok = (i > 1 || i < 4) ? mdoc_hash_find(mac) : MDOC_MAX; + if (MDOC_MAX == tok) { - mdoc_vmsg(m, MANDOCERR_MACRO, ln, sv, "%s", buf + sv - 1); + mandoc_vmsg(MANDOCERR_MACRO, m->parse, + ln, sv, "%s", buf + sv - 1); return(1); } /* Disregard the first trailing tab, if applicable. */ - if ('\t' == buf[i]) - i++; + if ('\t' == buf[offs]) + offs++; /* Jump to the next non-whitespace word. */ - while (buf[i] && ' ' == buf[i]) - i++; + while (buf[offs] && ' ' == buf[offs]) + offs++; /* * Trailing whitespace. Note that tabs are allowed to be passed * into the parser as "text", so we only warn about spaces here. */ - if ('\0' == buf[i] && ' ' == buf[i - 1]) - mdoc_pmsg(m, ln, i - 1, MANDOCERR_EOLNSPACE); + if ('\0' == buf[offs] && ' ' == buf[offs - 1]) + mdoc_pmsg(m, ln, offs - 1, MANDOCERR_EOLNSPACE); /* * If an initial macro or a list invocation, divert directly @@ -877,7 +896,7 @@ mdoc_pmacro(struct mdoc *m, int ln, char *buf, int off */ if (NULL == m->last || MDOC_It == tok || MDOC_El == tok) { - if ( ! mdoc_macro(m, tok, ln, sv, &i, buf)) + if ( ! mdoc_macro(m, tok, ln, sv, &offs, buf)) goto err; return(1); } @@ -916,7 +935,7 @@ mdoc_pmacro(struct mdoc *m, int ln, char *buf, int off /* Normal processing of a macro. */ - if ( ! mdoc_macro(m, tok, ln, sv, &i, buf)) + if ( ! mdoc_macro(m, tok, ln, sv, &offs, buf)) goto err; return(1); @@ -927,4 +946,48 @@ err: /* Error out. */ return(0); } +enum mdelim +mdoc_isdelim(const char *p) +{ + if ('\0' == p[0]) + return(DELIM_NONE); + + 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); + } + + if ('\\' != p[0]) + return(DELIM_NONE); + + if (0 == strcmp(p + 1, ".")) + return(DELIM_CLOSE); + if (0 == strcmp(p + 1, "*(Ba")) + return(DELIM_MIDDLE); + + return(DELIM_NONE); +}