=================================================================== RCS file: /cvs/mandoc/mdoc.c,v retrieving revision 1.90 retrieving revision 1.109 diff -u -p -r1.90 -r1.109 --- mandoc/mdoc.c 2009/07/12 20:50:08 1.90 +++ mandoc/mdoc.c 2009/10/15 02:56:51 1.109 @@ -1,4 +1,4 @@ -/* $Id: mdoc.c,v 1.90 2009/07/12 20:50:08 kristaps Exp $ */ +/* $Id: mdoc.c,v 1.109 2009/10/15 02:56:51 kristaps Exp $ */ /* * Copyright (c) 2008, 2009 Kristaps Dzonsons * @@ -14,6 +14,8 @@ * ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION, ARISING OUT OF * OR IN CONNECTION WITH THE USE OR PERFORMANCE OF THIS SOFTWARE. */ +#include + #include #include #include @@ -25,13 +27,10 @@ const char *const __mdoc_merrnames[MERRMAX] = { "trailing whitespace", /* ETAILWS */ - "empty last list column", /* ECOLEMPTY */ - "argument-like parameter", /* EARGVPARM */ "unexpected quoted parameter", /* EQUOTPARM */ "unterminated quoted parameter", /* EQUOTTERM */ "system: malloc error", /* EMALLOC */ "argument parameter suggested", /* EARGVAL */ - "macro not callable", /* ENOCALL */ "macro disallowed in prologue", /* EBODYPROL */ "macro disallowed in body", /* EPROLBODY */ "text disallowed in prologue", /* ETEXTPROL */ @@ -72,14 +71,14 @@ const char *const __mdoc_merrnames[MERRMAX] = { "superfluous width argument", /* ENOWIDTH */ "system: utsname error", /* EUTSNAME */ "obsolete macro", /* EOBS */ - "macro-like parameter", /* EMACPARM */ "end-of-line scope violation", /* EIMPBRK */ "empty macro ignored", /* EIGNE */ "unclosed explicit scope", /* EOPEN */ "unterminated quoted phrase", /* EQUOTPHR */ "closure macro without prior context", /* ENOCTX */ - "invalid whitespace after control character", /* ESPACE */ - "no description found for library" /* ELIB */ + "no description found for library", /* ELIB */ + "bad child for parent context", /* EBADCHILD */ + "list arguments preceding type", /* ENOTYPE */ }; const char *const __mdoc_macronames[MDOC_MAX] = { @@ -117,7 +116,7 @@ const char *const __mdoc_macronames[MDOC_MAX] = { /* LINTED */ "Brc", "\%C", "Es", "En", /* LINTED */ - "Dx", "\%Q" + "Dx", "\%Q", "br", "sp" }; const char *const __mdoc_argnames[MDOC_ARG_MAX] = { @@ -129,7 +128,7 @@ const char *const __mdoc_argnames[MDOC_ARG_MAX] = { "ohang", "inset", "column", "width", "compact", "std", "filled", "words", "emphasis", - "symbolic", "nested" + "symbolic", "nested", "centered" }; const char * const *mdoc_macronames = __mdoc_macronames; @@ -144,8 +143,14 @@ static int node_append(struct mdoc *, static int parsetext(struct mdoc *, int, char *); static int parsemacro(struct mdoc *, int, char *); static int macrowarn(struct mdoc *, int, const char *); +static int pstring(struct mdoc *, int, int, + const char *, size_t); +#ifdef __linux__ +extern size_t strlcpy(char *, const char *, size_t); +#endif + const struct mdoc_node * mdoc_node(const struct mdoc *m) { @@ -229,8 +234,6 @@ mdoc_free(struct mdoc *mdoc) { mdoc_free1(mdoc); - if (mdoc->htab) - mdoc_hash_free(mdoc->htab); free(mdoc); } @@ -248,13 +251,12 @@ mdoc_alloc(void *data, int pflags, const struct mdoc_c if (cb) (void)memcpy(&p->cb, cb, sizeof(struct mdoc_cb)); + mdoc_hash_init(); + p->data = data; p->pflags = pflags; - if (NULL == (p->htab = mdoc_hash_alloc())) { - free(p); - return(NULL); - } else if (mdoc_alloc1(p)) + if (mdoc_alloc1(p)) return(p); free(p); @@ -349,7 +351,10 @@ int mdoc_macro(struct mdoc *m, int tok, int ln, int pp, int *pos, char *buf) { - + /* + * If we're in the prologue, deny "body" macros. Similarly, if + * we're in the body, deny prologue calls. + */ if (MDOC_PROLOGUE & mdoc_macros[tok].flags && MDOC_PBODY & m->flags) return(mdoc_perr(m, ln, pp, EPROLBODY)); @@ -357,9 +362,6 @@ mdoc_macro(struct mdoc *m, int tok, ! (MDOC_PBODY & m->flags)) return(mdoc_perr(m, ln, pp, EBODYPROL)); - if (1 != pp && ! (MDOC_CALLABLE & mdoc_macros[tok].flags)) - return(mdoc_perr(m, ln, pp, ENOCALL)); - return((*mdoc_macros[tok].fp)(m, tok, ln, pp, pos, buf)); } @@ -429,17 +431,17 @@ node_append(struct mdoc *mdoc, struct mdoc_node *p) static struct mdoc_node * -node_alloc(struct mdoc *mdoc, int line, +node_alloc(struct mdoc *m, int line, int pos, int tok, enum mdoc_type type) { struct mdoc_node *p; if (NULL == (p = calloc(1, sizeof(struct mdoc_node)))) { - (void)mdoc_nerr(mdoc, mdoc->last, EMALLOC); + (void)mdoc_nerr(m, m->last, EMALLOC); return(NULL); } - p->sec = mdoc->lastsec; + p->sec = m->lastsec; p->line = line; p->pos = pos; p->tok = tok; @@ -451,94 +453,127 @@ node_alloc(struct mdoc *mdoc, int line, int -mdoc_tail_alloc(struct mdoc *mdoc, int line, int pos, int tok) +mdoc_tail_alloc(struct mdoc *m, int line, int pos, int tok) { struct mdoc_node *p; - p = node_alloc(mdoc, line, pos, tok, MDOC_TAIL); + p = node_alloc(m, line, pos, tok, MDOC_TAIL); if (NULL == p) return(0); - return(node_append(mdoc, p)); + if ( ! node_append(m, p)) + return(0); + m->next = MDOC_NEXT_CHILD; + return(1); } int -mdoc_head_alloc(struct mdoc *mdoc, int line, int pos, int tok) +mdoc_head_alloc(struct mdoc *m, int line, int pos, int tok) { struct mdoc_node *p; - assert(mdoc->first); - assert(mdoc->last); + assert(m->first); + assert(m->last); - p = node_alloc(mdoc, line, pos, tok, MDOC_HEAD); + p = node_alloc(m, line, pos, tok, MDOC_HEAD); if (NULL == p) return(0); - return(node_append(mdoc, p)); + if ( ! node_append(m, p)) + return(0); + m->next = MDOC_NEXT_CHILD; + return(1); } int -mdoc_body_alloc(struct mdoc *mdoc, int line, int pos, int tok) +mdoc_body_alloc(struct mdoc *m, int line, int pos, int tok) { struct mdoc_node *p; - p = node_alloc(mdoc, line, pos, tok, MDOC_BODY); + p = node_alloc(m, line, pos, tok, MDOC_BODY); if (NULL == p) return(0); - return(node_append(mdoc, p)); + if ( ! node_append(m, p)) + return(0); + m->next = MDOC_NEXT_CHILD; + return(1); } int -mdoc_block_alloc(struct mdoc *mdoc, int line, int pos, +mdoc_block_alloc(struct mdoc *m, int line, int pos, int tok, struct mdoc_arg *args) { struct mdoc_node *p; - p = node_alloc(mdoc, line, pos, tok, MDOC_BLOCK); + p = node_alloc(m, line, pos, tok, MDOC_BLOCK); if (NULL == p) return(0); p->args = args; if (p->args) (args->refcnt)++; - return(node_append(mdoc, p)); + if ( ! node_append(m, p)) + return(0); + m->next = MDOC_NEXT_CHILD; + return(1); } int -mdoc_elem_alloc(struct mdoc *mdoc, int line, int pos, +mdoc_elem_alloc(struct mdoc *m, int line, int pos, int tok, struct mdoc_arg *args) { struct mdoc_node *p; - p = node_alloc(mdoc, line, pos, tok, MDOC_ELEM); + p = node_alloc(m, line, pos, tok, MDOC_ELEM); if (NULL == p) return(0); p->args = args; if (p->args) (args->refcnt)++; - return(node_append(mdoc, p)); + if ( ! node_append(m, p)) + return(0); + m->next = MDOC_NEXT_CHILD; + return(1); } -int -mdoc_word_alloc(struct mdoc *mdoc, - int line, int pos, const char *word) +static int +pstring(struct mdoc *m, int line, int pos, const char *p, size_t len) { - struct mdoc_node *p; + struct mdoc_node *n; + size_t sv; - p = node_alloc(mdoc, line, pos, -1, MDOC_TEXT); - if (NULL == p) - return(0); - if (NULL == (p->string = strdup(word))) { - (void)mdoc_nerr(mdoc, mdoc->last, EMALLOC); - return(0); + n = node_alloc(m, line, pos, -1, MDOC_TEXT); + if (NULL == n) + return(mdoc_nerr(m, m->last, EMALLOC)); + + n->string = malloc(len + 1); + if (NULL == n->string) { + free(n); + return(mdoc_nerr(m, m->last, EMALLOC)); } - return(node_append(mdoc, p)); + sv = strlcpy(n->string, p, len + 1); + + /* Prohibit truncation. */ + assert(sv < len + 1); + + if ( ! node_append(m, n)) + return(0); + m->next = MDOC_NEXT_SIBLING; + return(1); } +int +mdoc_word_alloc(struct mdoc *m, int line, int pos, const char *p) +{ + + return(pstring(m, line, pos, p, strlen(p))); +} + + void mdoc_node_free(struct mdoc_node *p) { @@ -574,14 +609,52 @@ mdoc_node_freelist(struct mdoc_node *p) static int parsetext(struct mdoc *m, int line, char *buf) { + int i, j; if (SEC_NONE == m->lastnamed) return(mdoc_perr(m, line, 0, ETEXTPROL)); + + /* + * If in literal mode, then pass the buffer directly to the + * back-end, as it should be preserved as a single term. + */ - if (0 == buf[0] && ! (MDOC_LITERAL & m->flags)) + if (MDOC_LITERAL & m->flags) + return(mdoc_word_alloc(m, line, 0, buf)); + + /* Disallow blank/white-space lines in non-literal mode. */ + + for (i = 0; ' ' == buf[i]; i++) + /* Skip leading whitespace. */ ; + if (0 == buf[i]) return(mdoc_perr(m, line, 0, ENOBLANK)); - if ( ! mdoc_word_alloc(m, line, 0, buf)) + /* + * Break apart a free-form line into tokens. Spaces are + * stripped out of the input. + */ + + for (j = i; buf[i]; i++) { + if (' ' != buf[i]) + continue; + + /* Escaped whitespace. */ + if (i && ' ' == buf[i] && '\\' == buf[i - 1]) + continue; + + buf[i++] = 0; + if ( ! pstring(m, line, j, &buf[j], (size_t)(i - j))) + return(0); + + for ( ; ' ' == buf[i]; i++) + /* Skip trailing whitespace. */ ; + + j = i; + if (0 == buf[i]) + break; + } + + if (j != i && ! pstring(m, line, j, &buf[j], (size_t)(i - j))) return(0); m->next = MDOC_NEXT_SIBLING; @@ -589,14 +662,16 @@ parsetext(struct mdoc *m, int line, char *buf) } + + static int macrowarn(struct mdoc *m, int ln, const char *buf) { if ( ! (MDOC_IGN_MACRO & m->pflags)) - return(mdoc_verr(m, ln, 1, + return(mdoc_verr(m, ln, 0, "unknown macro: %s%s", buf, strlen(buf) > 3 ? "..." : "")); - return(mdoc_vwarn(m, ln, 1, "unknown macro: %s%s", + return(mdoc_vwarn(m, ln, 0, "unknown macro: %s%s", buf, strlen(buf) > 3 ? "..." : "")); } @@ -608,7 +683,7 @@ macrowarn(struct mdoc *m, int ln, const char *buf) int parsemacro(struct mdoc *m, int ln, char *buf) { - int i, c; + int i, j, c; char mac[5]; /* Empty lines are ignored. */ @@ -616,33 +691,42 @@ parsemacro(struct mdoc *m, int ln, char *buf) if (0 == buf[1]) return(1); - if (' ' == buf[1]) { - i = 2; + i = 1; + + /* Accept whitespace after the initial control char. */ + + if (' ' == buf[i]) { + i++; while (buf[i] && ' ' == buf[i]) i++; if (0 == buf[i]) return(1); - return(mdoc_perr(m, ln, 1, ESPACE)); } /* Copy the first word into a nil-terminated buffer. */ - for (i = 1; i < 5; i++) { - if (0 == (mac[i - 1] = buf[i])) + for (j = 0; j < 4; j++, i++) { + if (0 == (mac[j] = buf[i])) break; else if (' ' == buf[i]) break; + + /* Check for invalid characters. */ + + if (isgraph((u_char)buf[i])) + continue; + return(mdoc_perr(m, ln, i, EPRINT)); } - mac[i - 1] = 0; + mac[j] = 0; - if (i == 5 || i <= 2) { + if (j == 4 || j < 2) { if ( ! macrowarn(m, ln, mac)) goto err; return(1); } - if (MDOC_MAX == (c = mdoc_hash_find(m->htab, mac))) { + if (MDOC_MAX == (c = mdoc_hash_find(mac))) { if ( ! macrowarn(m, ln, mac)) goto err; return(1); @@ -653,8 +737,10 @@ parsemacro(struct mdoc *m, int ln, char *buf) while (buf[i] && ' ' == buf[i]) i++; - /* Begin recursive parse sequence. */ - + /* + * Begin recursive parse sequence. Since we're at the start of + * the line, we don't need to do callable/parseable checks. + */ if ( ! mdoc_macro(m, c, ln, 1, &i, buf)) goto err; @@ -665,3 +751,5 @@ err: /* Error out. */ m->flags |= MDOC_HALT; return(0); } + +