=================================================================== RCS file: /cvs/mandoc/roff.c,v retrieving revision 1.93 retrieving revision 1.99 diff -u -p -r1.93 -r1.99 --- mandoc/roff.c 2010/07/04 22:04:04 1.93 +++ mandoc/roff.c 2010/08/24 12:48:43 1.99 @@ -1,4 +1,4 @@ -/* $Id: roff.c,v 1.93 2010/07/04 22:04:04 schwarze Exp $ */ +/* $Id: roff.c,v 1.99 2010/08/24 12:48:43 kristaps Exp $ */ /* * Copyright (c) 2010 Kristaps Dzonsons * Copyright (c) 2010 Ingo Schwarze @@ -28,8 +28,8 @@ #include #include "mandoc.h" -#include "regs.h" #include "roff.h" +#include "libmandoc.h" #define RSTACK_MAX 128 @@ -69,6 +69,13 @@ enum roffrule { ROFFRULE_DENY }; + +struct roffstr { + char *name; /* key of symbol */ + char *string; /* current value */ + struct roffstr *next; /* next in list */ +}; + struct roff { struct roffnode *last; /* leaf of stack */ mandocmsg msg; /* err/warn/fatal messages */ @@ -76,6 +83,7 @@ struct roff { enum roffrule rstack[RSTACK_MAX]; /* stack of !`ie' rules */ int rstackpos; /* position in rstack */ struct regset *regs; /* read/writable registers */ + struct roffstr *first_string; }; struct roffnode { @@ -109,12 +117,6 @@ struct roffmac { struct roffmac *next; }; -struct roffstr { - char *name; - char *string; - struct roffstr *next; -} *first_string; - static enum rofferr roff_block(ROFF_ARGS); static enum rofferr roff_block_text(ROFF_ARGS); static enum rofferr roff_block_sub(ROFF_ARGS); @@ -124,9 +126,17 @@ static enum rofferr roff_cond(ROFF_ARGS); static enum rofferr roff_cond_text(ROFF_ARGS); static enum rofferr roff_cond_sub(ROFF_ARGS); static enum rofferr roff_ds(ROFF_ARGS); +static enum roffrule roff_evalcond(const char *, int *); +static void roff_freestr(struct roff *); +static const char *roff_getstrn(const struct roff *, + const char *, size_t); static enum rofferr roff_line(ROFF_ARGS); static enum rofferr roff_nr(ROFF_ARGS); -static enum roffrule roff_evalcond(const char *, int *); +static int roff_res(struct roff *, + char **, size_t *, int); +static void roff_setstr(struct roff *, + const char *, const char *); +static char *roff_strdup(const char *); /* See roff_hash_find() */ @@ -159,7 +169,7 @@ static void roff_free1(struct roff *); static enum rofft roff_hash_find(const char *); static void roff_hash_init(void); static void roffnode_cleanscope(struct roff *); -static int roffnode_push(struct roff *, +static void roffnode_push(struct roff *, enum rofft, int, int); static void roffnode_pop(struct roff *); static enum rofft roff_parse(const char *, int *); @@ -249,16 +259,12 @@ roffnode_pop(struct roff *r) * Push a roff node onto the instruction stack. This must later be * removed with roffnode_pop(). */ -static int +static void roffnode_push(struct roff *r, enum rofft tok, int line, int col) { struct roffnode *p; - if (NULL == (p = calloc(1, sizeof(struct roffnode)))) { - (*r->msg)(MANDOCERR_MEM, r->data, line, col, NULL); - return(0); - } - + p = mandoc_calloc(1, sizeof(struct roffnode)); p->tok = tok; p->parent = r->last; p->line = line; @@ -266,7 +272,6 @@ roffnode_push(struct roff *r, enum rofft tok, int line p->rule = p->parent ? p->parent->rule : ROFFRULE_DENY; r->last = p; - return(1); } @@ -276,7 +281,7 @@ roff_free1(struct roff *r) while (r->last) roffnode_pop(r); - roff_freestr(); + roff_freestr(r); } @@ -298,15 +303,11 @@ roff_free(struct roff *r) struct roff * -roff_alloc(struct regset *regs, const mandocmsg msg, void *data) +roff_alloc(struct regset *regs, void *data, const mandocmsg msg) { struct roff *r; - if (NULL == (r = calloc(1, sizeof(struct roff)))) { - (*msg)(MANDOCERR_MEM, data, 0, 0, NULL); - return(0); - } - + r = mandoc_calloc(1, sizeof(struct roff)); r->regs = regs; r->msg = msg; r->data = data; @@ -317,6 +318,74 @@ roff_alloc(struct regset *regs, const mandocmsg msg, v } +/* + * Pre-filter each and every line for reserved words (one beginning with + * `\*', e.g., `\*(ab'). These must be handled before the actual line + * is processed. + */ +static int +roff_res(struct roff *r, char **bufp, size_t *szp, int pos) +{ + const char *cp, *cpp, *st, *res; + int i, maxl; + size_t nsz; + char *n; + + /* LINTED */ + for (cp = &(*bufp)[pos]; (cpp = strstr(cp, "\\*")); cp++) { + cp = cpp + 2; + switch (*cp) { + case ('('): + cp++; + maxl = 2; + break; + case ('['): + cp++; + maxl = 0; + break; + default: + maxl = 1; + break; + } + + st = cp; + + for (i = 0; 0 == maxl || i < maxl; i++, cp++) { + if ('\0' == *cp) + return(1); /* Error. */ + if (0 == maxl && ']' == *cp) + break; + } + + res = roff_getstrn(r, st, (size_t)i); + + if (NULL == res) { + cp -= maxl ? 1 : 0; + continue; + } + + ROFF_DEBUG("roff: splicing reserved: [%.*s]\n", i, st); + + nsz = *szp + strlen(res) + 1; + n = mandoc_malloc(nsz); + + *n = '\0'; + + strlcat(n, *bufp, (size_t)(cpp - *bufp + 1)); + strlcat(n, res, nsz); + strlcat(n, cp + (maxl ? 0 : 1), nsz); + + free(*bufp); + + *bufp = n; + *szp = nsz; + return(0); + } + + return(1); +} + + enum rofferr roff_parseln(struct roff *r, int ln, char **bufp, size_t *szp, int pos, int *offs) @@ -325,6 +394,14 @@ roff_parseln(struct roff *r, int ln, char **bufp, int ppos; /* + * Run the reserved-word filter only if we have some reserved + * words to fill in. + */ + + if (r->first_string && ! roff_res(r, bufp, szp, pos)) + return(ROFF_RERUN); + + /* * First, if a scope is open and we're not a macro, pass the * text through the macro's filter. If a scope isn't open and * we're not a macro, just let it through. @@ -338,11 +415,8 @@ roff_parseln(struct roff *r, int ln, char **bufp, return((*roffs[t].text) (r, t, bufp, szp, ln, pos, pos, offs)); - } else if ( ! ROFF_CTL((*bufp)[pos])) { - ROFF_DEBUG("roff: pass non-scoped text: [%s]\n", - &(*bufp)[pos]); + } else if ( ! ROFF_CTL((*bufp)[pos])) return(ROFF_CONT); - } /* * If a scope is open, go to the child handler for that macro, @@ -366,11 +440,8 @@ roff_parseln(struct roff *r, int ln, char **bufp, */ ppos = pos; - if (ROFF_MAX == (t = roff_parse(*bufp, &pos))) { - ROFF_DEBUG("roff: pass non-scoped non-macro: [%s]\n", - &(*bufp)[pos]); + if (ROFF_MAX == (t = roff_parse(*bufp, &pos))) return(ROFF_CONT); - } ROFF_DEBUG("roff: intercept new-scope: %s, [%s]\n", roffs[t].name, &(*bufp)[pos]); @@ -571,8 +642,7 @@ roff_block(ROFF_ARGS) pos++; } - if ( ! roffnode_push(r, tok, ln, ppos)) - return(ROFF_ERR); + roffnode_push(r, tok, ln, ppos); if ('\0' == (*bufp)[pos]) return(ROFF_IGN); @@ -594,13 +664,8 @@ roff_block(ROFF_ARGS) if (1 == sz && '.' == (*bufp)[sv]) return(ROFF_IGN); - r->last->end = malloc(sz + 1); + r->last->end = mandoc_malloc(sz + 1); - if (NULL == r->last->end) { - (*r->msg)(MANDOCERR_MEM, r->data, ln, pos, NULL); - return(ROFF_ERR); - } - memcpy(r->last->end, *bufp + sv, sz); r->last->end[(int)sz] = '\0'; @@ -826,8 +891,7 @@ roff_cond(ROFF_ARGS) return(ROFF_ERR); } - if ( ! roffnode_push(r, tok, ln, ppos)) - return(ROFF_ERR); + roffnode_push(r, tok, ln, ppos); r->last->rule = rule; @@ -890,31 +954,41 @@ roff_cond(ROFF_ARGS) static enum rofferr roff_ds(ROFF_ARGS) { - char *name, *string, *end; + char *name, *string; + /* + * A symbol is named by the first word following the macro + * invocation up to a space. Its value is anything after the + * name's trailing whitespace and optional double-quote. Thus, + * + * [.ds foo "bar " ] + * + * will have `bar " ' as its value. + */ + name = *bufp + pos; if ('\0' == *name) return(ROFF_IGN); string = name; + /* Read until end of name. */ while (*string && ' ' != *string) string++; + + /* Nil-terminate name. */ if (*string) - *(string++) = NULL; - if (*string && '"' == *string) - string++; + *(string++) = '\0'; + + /* Read past spaces. */ while (*string && ' ' == *string) string++; - end = string; - while (*end) - end++; - if (string < end) { - end--; - if (*end == '"') - *end = '\0'; - } - roff_setstr(name, string); + /* Read passed initial double-quote. */ + if (*string && '"' == *string) + string++; + + /* The rest is the value. */ + roff_setstr(r, name, string); return(ROFF_IGN); } @@ -962,73 +1036,76 @@ roff_nr(ROFF_ARGS) } -char * -roff_setstr(const char *name, const char *string) +static char * +roff_strdup(const char *name) { - struct roffstr *n; - char *namecopy; + char *namecopy, *sv; - n = first_string; - while (n && strcmp(name, n->name)) - n = n->next; - if (n) { - free(n->string); - } else { - if (NULL == (namecopy = strdup(name))) - return(NULL); - if (NULL == (n = malloc(sizeof(struct roffstr)))) { - free(n); - return(NULL); - } - n->name = namecopy; - n->next = first_string; - first_string = n; + /* + * This isn't a nice simple mandoc_strdup() because we must + * handle roff's stupid double-escape rule. + */ + sv = namecopy = mandoc_malloc(strlen(name) + 1); + while (*name) { + if ('\\' == *name && '\\' == *(name + 1)) + name++; + *namecopy++ = *name++; } - if (string) - n->string = strdup(string); - else - n->string = NULL; - return(n->string); + + *namecopy = '\0'; + return(sv); } -char * -roff_getstr(const char *name) + +static void +roff_setstr(struct roff *r, const char *name, const char *string) { struct roffstr *n; + char *namecopy; - n = first_string; + n = r->first_string; while (n && strcmp(name, n->name)) n = n->next; - if (n) - return(n->string); - else - return(NULL); + + if (NULL == n) { + namecopy = mandoc_strdup(name); + n = mandoc_malloc(sizeof(struct roffstr)); + n->name = namecopy; + n->next = r->first_string; + r->first_string = n; + } else + free(n->string); + + /* Don't use mandoc_strdup: clean out double-escapes. */ + n->string = string ? roff_strdup(string) : NULL; + ROFF_DEBUG("roff: new symbol: [%s] = [%s]\n", name, n->string); } -char * -roff_getstrn(const char *name, size_t len) + +static const char * +roff_getstrn(const struct roff *r, const char *name, size_t len) { - struct roffstr *n; + const struct roffstr *n; - n = first_string; - while (n && (strncmp(name, n->name, len) || '\0' != n->name[len])) + n = r->first_string; + while (n && (strncmp(name, n->name, len) || '\0' != n->name[(int)len])) n = n->next; - if (n) - return(n->string); - else - return(NULL); + + return(n ? n->string : NULL); } -void -roff_freestr(void) + +static void +roff_freestr(struct roff *r) { struct roffstr *n, *nn; - for (n = first_string; n; n = nn) { + for (n = r->first_string; n; n = nn) { free(n->name); free(n->string); nn = n->next; free(n); } - first_string = NULL; + + r->first_string = NULL; }