=================================================================== RCS file: /cvs/mandoc/mdoc.c,v retrieving revision 1.96 retrieving revision 1.135 diff -u -p -r1.96 -r1.135 --- mandoc/mdoc.c 2009/07/27 19:43:02 1.96 +++ mandoc/mdoc.c 2010/05/16 10:59:36 1.135 @@ -1,4 +1,4 @@ -/* $Id: mdoc.c,v 1.96 2009/07/27 19:43:02 kristaps Exp $ */ +/* $Id: mdoc.c,v 1.135 2010/05/16 10:59:36 kristaps Exp $ */ /* * Copyright (c) 2008, 2009 Kristaps Dzonsons * @@ -14,22 +14,28 @@ * ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION, ARISING OUT OF * OR IN CONNECTION WITH THE USE OR PERFORMANCE OF THIS SOFTWARE. */ +#ifdef HAVE_CONFIG_H +#include "config.h" +#endif + +#include + #include #include #include #include #include #include +#include #include "libmdoc.h" +#include "libmandoc.h" const char *const __mdoc_merrnames[MERRMAX] = { "trailing whitespace", /* ETAILWS */ "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 */ @@ -63,21 +69,21 @@ const char *const __mdoc_merrnames[MERRMAX] = { "prologue macro out of conventional order", /* EPROLOOO */ "prologue macro repeated", /* EPROLREP */ "invalid manual section", /* EBADMSEC */ - "invalid section", /* EBADSEC */ "invalid font mode", /* EFONT */ "invalid date syntax", /* EBADDATE */ "invalid number format", /* ENUMFMT */ "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 */ + "deprecated comment style", /* EBADCOMMENT */ }; const char *const __mdoc_macronames[MDOC_MAX] = { @@ -92,11 +98,11 @@ const char *const __mdoc_macronames[MDOC_MAX] = { "Nm", "Op", "Ot", "Pa", "Rv", "St", "Va", "Vt", /* LINTED */ - "Xr", "\%A", "\%B", "\%D", + "Xr", "%A", "%B", "%D", /* LINTED */ - "\%I", "\%J", "\%N", "\%O", + "%I", "%J", "%N", "%O", /* LINTED */ - "\%P", "\%R", "\%T", "\%V", + "%P", "%R", "%T", "%V", "Ac", "Ao", "Aq", "At", "Bc", "Bf", "Bo", "Bq", "Bsx", "Bx", "Db", "Dc", @@ -113,9 +119,11 @@ const char *const __mdoc_macronames[MDOC_MAX] = { "Fr", "Ud", "Lb", "Lp", "Lk", "Mt", "Brq", "Bro", /* LINTED */ - "Brc", "\%C", "Es", "En", + "Brc", "%C", "Es", "En", /* LINTED */ - "Dx", "\%Q", "br", "sp" + "Dx", "%Q", "br", "sp", + /* LINTED */ + "%U" }; const char *const __mdoc_argnames[MDOC_ARG_MAX] = { @@ -127,29 +135,27 @@ 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; const char * const *mdoc_argnames = __mdoc_argnames; +static void mdoc_node_free(struct mdoc_node *); +static void mdoc_node_unlink(struct mdoc *, + struct mdoc_node *); static void mdoc_free1(struct mdoc *); -static int mdoc_alloc1(struct mdoc *); +static void mdoc_alloc1(struct mdoc *); static struct mdoc_node *node_alloc(struct mdoc *, int, int, - int, enum mdoc_type); + enum mdoct, enum mdoc_type); static int node_append(struct mdoc *, struct mdoc_node *); -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); +static int mdoc_ptext(struct mdoc *, int, char *, int); +static int mdoc_pmacro(struct mdoc *, int, char *, int); +static int macrowarn(struct mdoc *, int, + const char *, int); -#ifdef __linux__ -extern size_t strlcpy(char *, const char *, size_t); -#endif - const struct mdoc_node * mdoc_node(const struct mdoc *m) { @@ -174,7 +180,7 @@ mdoc_free1(struct mdoc *mdoc) { if (mdoc->first) - mdoc_node_freelist(mdoc->first); + mdoc_node_delete(mdoc, mdoc->first); if (mdoc->meta.title) free(mdoc->meta.title); if (mdoc->meta.os) @@ -185,27 +191,25 @@ mdoc_free1(struct mdoc *mdoc) free(mdoc->meta.arch); if (mdoc->meta.vol) free(mdoc->meta.vol); + if (mdoc->meta.msec) + free(mdoc->meta.msec); } /* * Allocate all volatile resources (parse tree, meta-data, fields). */ -static int +static void mdoc_alloc1(struct mdoc *mdoc) { - bzero(&mdoc->meta, sizeof(struct mdoc_meta)); + memset(&mdoc->meta, 0, sizeof(struct mdoc_meta)); mdoc->flags = 0; mdoc->lastnamed = mdoc->lastsec = SEC_NONE; - mdoc->last = calloc(1, sizeof(struct mdoc_node)); - if (NULL == mdoc->last) - return(0); - + mdoc->last = mandoc_calloc(1, sizeof(struct mdoc_node)); mdoc->first = mdoc->last; mdoc->last->type = MDOC_ROOT; mdoc->next = MDOC_NEXT_CHILD; - return(1); } @@ -215,12 +219,12 @@ mdoc_alloc1(struct mdoc *mdoc) * and the parser is ready for re-invocation on a new tree; however, * cross-parse non-volatile data is kept intact. */ -int +void mdoc_reset(struct mdoc *mdoc) { mdoc_free1(mdoc); - return(mdoc_alloc1(mdoc)); + mdoc_alloc1(mdoc); } @@ -233,8 +237,6 @@ mdoc_free(struct mdoc *mdoc) { mdoc_free1(mdoc); - if (mdoc->htab) - mdoc_hash_free(mdoc->htab); free(mdoc); } @@ -247,22 +249,17 @@ mdoc_alloc(void *data, int pflags, const struct mdoc_c { struct mdoc *p; - if (NULL == (p = calloc(1, sizeof(struct mdoc)))) - return(NULL); + p = mandoc_calloc(1, sizeof(struct mdoc)); + if (cb) - (void)memcpy(&p->cb, cb, sizeof(struct mdoc_cb)); + memcpy(&p->cb, cb, sizeof(struct mdoc_cb)); p->data = data; p->pflags = pflags; - if (NULL == (p->htab = mdoc_hash_alloc())) { - free(p); - return(NULL); - } else if (mdoc_alloc1(p)) - return(p); - - free(p); - return(NULL); + mdoc_hash_init(); + mdoc_alloc1(p); + return(p); } @@ -285,17 +282,19 @@ mdoc_endparse(struct mdoc *m) /* * Main parse routine. Parses a single line -- really just hands off to - * the macro (parsemacro()) or text parser (parsetext()). + * the macro (mdoc_pmacro()) or text parser (mdoc_ptext()). */ int -mdoc_parseln(struct mdoc *m, int ln, char *buf) +mdoc_parseln(struct mdoc *m, int ln, char *buf, int offs) { if (MDOC_HALT & m->flags) return(0); - return('.' == *buf ? parsemacro(m, ln, buf) : - parsetext(m, ln, buf)); + m->flags |= MDOC_NEWLINE; + return(('.' == buf[offs] || '\'' == buf[offs]) ? + mdoc_pmacro(m, ln, buf, offs) : + mdoc_ptext(m, ln, buf, offs)); } @@ -350,20 +349,34 @@ mdoc_err(struct mdoc *m, int line, int pos, int iserr, int -mdoc_macro(struct mdoc *m, int tok, +mdoc_macro(struct mdoc *m, enum mdoct tok, int ln, int pp, int *pos, char *buf) { + assert(tok < MDOC_MAX); + /* 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)); - if ( ! (MDOC_PROLOGUE & mdoc_macros[tok].flags) && - ! (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)); + /* If we're in the prologue, deny "body" macros. */ + if ( ! (MDOC_PROLOGUE & mdoc_macros[tok].flags) && + ! (MDOC_PBODY & m->flags)) { + if ( ! mdoc_pwarn(m, ln, pp, EBODYPROL)) + return(0); + if (NULL == m->meta.title) + m->meta.title = mandoc_strdup("unknown"); + if (NULL == m->meta.vol) + m->meta.vol = mandoc_strdup("local"); + if (NULL == m->meta.os) + m->meta.os = mandoc_strdup("local"); + if (0 == m->meta.date) + m->meta.date = time(NULL); + m->flags |= MDOC_PBODY; + } + return((*mdoc_macros[tok].fp)(m, tok, ln, pp, pos, buf)); } @@ -433,41 +446,39 @@ node_append(struct mdoc *mdoc, struct mdoc_node *p) static struct mdoc_node * -node_alloc(struct mdoc *m, int line, - int pos, int tok, enum mdoc_type type) +node_alloc(struct mdoc *m, int line, int pos, + enum mdoct tok, enum mdoc_type type) { struct mdoc_node *p; - if (NULL == (p = calloc(1, sizeof(struct mdoc_node)))) { - (void)mdoc_nerr(m, m->last, EMALLOC); - return(NULL); - } - + p = mandoc_calloc(1, sizeof(struct mdoc_node)); p->sec = m->lastsec; p->line = line; p->pos = pos; p->tok = tok; - if (MDOC_TEXT != (p->type = type)) - assert(p->tok >= 0); - + p->type = type; + if (MDOC_NEWLINE & m->flags) + p->flags |= MDOC_LINE; + m->flags &= ~MDOC_NEWLINE; return(p); } int -mdoc_tail_alloc(struct mdoc *m, int line, int pos, int tok) +mdoc_tail_alloc(struct mdoc *m, int line, int pos, enum mdoct tok) { struct mdoc_node *p; p = node_alloc(m, line, pos, tok, MDOC_TAIL); - if (NULL == p) + if ( ! node_append(m, p)) return(0); - return(node_append(m, p)); + m->next = MDOC_NEXT_CHILD; + return(1); } int -mdoc_head_alloc(struct mdoc *m, int line, int pos, int tok) +mdoc_head_alloc(struct mdoc *m, int line, int pos, enum mdoct tok) { struct mdoc_node *p; @@ -475,86 +486,80 @@ mdoc_head_alloc(struct mdoc *m, int line, int pos, int assert(m->last); p = node_alloc(m, line, pos, tok, MDOC_HEAD); - if (NULL == p) + if ( ! node_append(m, p)) return(0); - return(node_append(m, p)); + m->next = MDOC_NEXT_CHILD; + return(1); } int -mdoc_body_alloc(struct mdoc *m, int line, int pos, int tok) +mdoc_body_alloc(struct mdoc *m, int line, int pos, enum mdoct tok) { struct mdoc_node *p; p = node_alloc(m, line, pos, tok, MDOC_BODY); - if (NULL == p) + if ( ! node_append(m, p)) return(0); - return(node_append(m, p)); + m->next = MDOC_NEXT_CHILD; + return(1); } int mdoc_block_alloc(struct mdoc *m, int line, int pos, - int tok, struct mdoc_arg *args) + enum mdoct tok, struct mdoc_arg *args) { struct mdoc_node *p; 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(m, p)); + if ( ! node_append(m, p)) + return(0); + m->next = MDOC_NEXT_CHILD; + return(1); } int mdoc_elem_alloc(struct mdoc *m, int line, int pos, - int tok, struct mdoc_arg *args) + enum mdoct tok, struct mdoc_arg *args) { struct mdoc_node *p; 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(m, p)); + if ( ! node_append(m, p)) + return(0); + m->next = MDOC_NEXT_CHILD; + return(1); } -static int -pstring(struct mdoc *m, int line, int pos, const char *p, size_t len) +int +mdoc_word_alloc(struct mdoc *m, int line, int pos, const char *p) { struct mdoc_node *n; - size_t sv; + size_t sv, len; - n = node_alloc(m, line, pos, -1, MDOC_TEXT); - if (NULL == n) - return(mdoc_nerr(m, m->last, EMALLOC)); + len = strlen(p); - n->string = malloc(len + 1); - if (NULL == n->string) { - free(n); - return(mdoc_nerr(m, m->last, EMALLOC)); - } - + n = node_alloc(m, line, pos, MDOC_MAX, MDOC_TEXT); + n->string = mandoc_malloc(len + 1); sv = strlcpy(n->string, p, len + 1); /* Prohibit truncation. */ assert(sv < len + 1); - return(node_append(m, n)); -} + if ( ! node_append(m, n)) + return(0); - -int -mdoc_word_alloc(struct mdoc *m, int line, int pos, const char *p) -{ - - return(pstring(m, line, pos, p, strlen(p))); + m->next = MDOC_NEXT_SIBLING; + return(1); } @@ -562,8 +567,6 @@ void mdoc_node_free(struct mdoc_node *p) { - if (p->parent) - p->parent->nchild--; if (p->string) free(p->string); if (p->args) @@ -572,16 +575,53 @@ mdoc_node_free(struct mdoc_node *p) } -void -mdoc_node_freelist(struct mdoc_node *p) +static void +mdoc_node_unlink(struct mdoc *m, struct mdoc_node *n) { - if (p->child) - mdoc_node_freelist(p->child); - if (p->next) - mdoc_node_freelist(p->next); + /* Adjust siblings. */ + if (n->prev) + n->prev->next = n->next; + if (n->next) + n->next->prev = n->prev; + + /* Adjust parent. */ + + if (n->parent) { + n->parent->nchild--; + if (n->parent->child == n) + n->parent->child = n->prev ? n->prev : n->next; + } + + /* Adjust parse point, if applicable. */ + + if (m && m->last == n) { + if (n->prev) { + m->last = n->prev; + m->next = MDOC_NEXT_SIBLING; + } else { + m->last = n->parent; + m->next = MDOC_NEXT_CHILD; + } + } + + if (m && m->first == n) + m->first = NULL; +} + + +void +mdoc_node_delete(struct mdoc *m, struct mdoc_node *p) +{ + + while (p->child) { + assert(p->nchild); + mdoc_node_delete(m, p->child); + } assert(0 == p->nchild); + + mdoc_node_unlink(m, p); mdoc_node_free(p); } @@ -591,76 +631,96 @@ mdoc_node_freelist(struct mdoc_node *p) * control character. */ static int -parsetext(struct mdoc *m, int line, char *buf) +mdoc_ptext(struct mdoc *m, int line, char *buf, int offs) { - int i, j; + int i; + /* Ignore bogus comments. */ + + if ('\\' == buf[offs] && + '.' == buf[offs + 1] && + '"' == buf[offs + 2]) + return(mdoc_pwarn(m, line, offs, EBADCOMMENT)); + + /* No text before an initial macro. */ + if (SEC_NONE == m->lastnamed) - return(mdoc_perr(m, line, 0, ETEXTPROL)); + return(mdoc_perr(m, line, offs, ETEXTPROL)); + + /* Literal just gets pulled in as-is. */ - /* - * If in literal mode, then pass the buffer directly to the - * back-end, as it should be preserved as a single term. - */ + if (MDOC_LITERAL & m->flags) + return(mdoc_word_alloc(m, line, offs, buf + offs)); - if (MDOC_LITERAL & m->flags) { - if ( ! mdoc_word_alloc(m, line, 0, buf)) + /* Check for a blank line, which may also consist of spaces. */ + + for (i = offs; ' ' == buf[i]; i++) + /* Skip to first non-space. */ ; + + if ('\0' == buf[i]) { + if ( ! mdoc_pwarn(m, line, offs, ENOBLANK)) return(0); + + /* + * Insert a `Pp' in the case of a blank line. Technically, + * blank lines aren't allowed, but enough manuals assume this + * behaviour that we want to work around it. + */ + if ( ! mdoc_elem_alloc(m, line, offs, MDOC_Pp, NULL)) + return(0); + m->next = MDOC_NEXT_SIBLING; return(1); } - /* 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)); - - /* - * Break apart a free-form line into tokens. Spaces are - * stripped out of the input. + /* + * Warn if the last un-escaped character is whitespace. Then + * strip away the remaining spaces (tabs stay!). */ - for (j = i; buf[i]; i++) { - if (' ' != buf[i]) - continue; + i = (int)strlen(buf); + assert(i); - /* Escaped whitespace. */ - if (i && ' ' == buf[i] && '\\' == buf[i - 1]) - continue; + if (' ' == buf[i - 1] || '\t' == buf[i - 1]) { + if (i > 1 && '\\' != buf[i - 2]) + if ( ! mdoc_pwarn(m, line, i - 1, ETAILWS)) + return(0); - buf[i++] = 0; - if ( ! pstring(m, line, j, &buf[j], (size_t)(i - j))) - return(0); - m->next = MDOC_NEXT_SIBLING; + for (--i; i && ' ' == buf[i]; i--) + /* Spin back to non-space. */ ; - for ( ; ' ' == buf[i]; i++) - /* Skip trailing whitespace. */ ; + /* Jump ahead of escaped whitespace. */ + i += '\\' == buf[i] ? 2 : 1; - j = i; - if (0 == buf[i]) - break; + buf[i] = '\0'; } - if (j != i && ! pstring(m, line, j, &buf[j], (size_t)(i - j))) + /* Allocate the whole word. */ + + if ( ! mdoc_word_alloc(m, line, offs, buf + offs)) return(0); - m->next = MDOC_NEXT_SIBLING; + /* + * End-of-sentence check. If the last character is an unescaped + * EOS character, then flag the node as being the end of a + * sentence. The front-end will know how to interpret this. + */ + + assert(i); + if (mandoc_eos(buf, (size_t)i)) + m->last->flags |= MDOC_EOS; + return(1); } - - static int -macrowarn(struct mdoc *m, int ln, const char *buf) +macrowarn(struct mdoc *m, int ln, const char *buf, int offs) { if ( ! (MDOC_IGN_MACRO & m->pflags)) - return(mdoc_verr(m, ln, 0, - "unknown macro: %s%s", + return(mdoc_verr(m, ln, offs, "unknown macro: %s%s", buf, strlen(buf) > 3 ? "..." : "")); - return(mdoc_vwarn(m, ln, 0, "unknown macro: %s%s", + return(mdoc_vwarn(m, ln, offs, "unknown macro: %s%s", buf, strlen(buf) > 3 ? "..." : "")); } @@ -670,44 +730,58 @@ macrowarn(struct mdoc *m, int ln, const char *buf) * character. */ int -parsemacro(struct mdoc *m, int ln, char *buf) +mdoc_pmacro(struct mdoc *m, int ln, char *buf, int offs) { - int i, c; - char mac[5]; + enum mdoct tok; + int i, j, sv; + char mac[5]; /* Empty lines are ignored. */ - if (0 == buf[1]) + offs++; + + if ('\0' == buf[offs]) return(1); - if (' ' == buf[1]) { - i = 2; + i = offs; + + /* Accept whitespace after the initial control char. */ + + if (' ' == buf[i]) { + i++; while (buf[i] && ' ' == buf[i]) i++; - if (0 == buf[i]) + if ('\0' == buf[i]) return(1); - return(mdoc_perr(m, ln, 1, ESPACE)); } + sv = i; + /* 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 ( ! macrowarn(m, ln, mac)) + if (j == 4 || j < 2) { + if ( ! macrowarn(m, ln, mac, sv)) goto err; return(1); } - if (MDOC_MAX == (c = mdoc_hash_find(m->htab, mac))) { - if ( ! macrowarn(m, ln, mac)) + if (MDOC_MAX == (tok = mdoc_hash_find(mac))) { + if ( ! macrowarn(m, ln, mac, sv)) goto err; return(1); } @@ -717,9 +791,20 @@ parsemacro(struct mdoc *m, int ln, char *buf) while (buf[i] && ' ' == buf[i]) i++; - /* Begin recursive parse sequence. */ + /* + * Trailing whitespace. Note that tabs are allowed to be passed + * into the parser as "text", so we only warn about spaces here. + */ - if ( ! mdoc_macro(m, c, ln, 1, &i, buf)) + if ('\0' == buf[i] && ' ' == buf[i - 1]) + if ( ! mdoc_pwarn(m, ln, i - 1, ETAILWS)) + goto err; + + /* + * 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, tok, ln, sv, &i, buf)) goto err; return(1); @@ -729,3 +814,5 @@ err: /* Error out. */ m->flags |= MDOC_HALT; return(0); } + +