-/* $Id: mdoc.c,v 1.88 2009/07/06 13:04:52 kristaps Exp $ */
+/* $Id: mdoc.c,v 1.116 2010/01/07 10:24:43 kristaps Exp $ */
/*
* Copyright (c) 2008, 2009 Kristaps Dzonsons <kristaps@kth.se>
*
* 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 <sys/types.h>
+
#include <assert.h>
#include <ctype.h>
#include <stdarg.h>
#include <string.h>
#include "libmdoc.h"
+#include "libmandoc.h"
+
+const char *const __mdoc_merrnames[MERRMAX] = {
+ "trailing whitespace", /* ETAILWS */
+ "unexpected quoted parameter", /* EQUOTPARM */
+ "unterminated quoted parameter", /* EQUOTTERM */
+ "argument parameter suggested", /* EARGVAL */
+ "macro disallowed in prologue", /* EBODYPROL */
+ "macro disallowed in body", /* EPROLBODY */
+ "text disallowed in prologue", /* ETEXTPROL */
+ "blank line disallowed", /* ENOBLANK */
+ "text parameter too long", /* ETOOLONG */
+ "invalid escape sequence", /* EESCAPE */
+ "invalid character", /* EPRINT */
+ "document has no body", /* ENODAT */
+ "document has no prologue", /* ENOPROLOGUE */
+ "expected line arguments", /* ELINE */
+ "invalid AT&T argument", /* EATT */
+ "default name not yet set", /* ENAME */
+ "missing list type", /* ELISTTYPE */
+ "missing display type", /* EDISPTYPE */
+ "too many display types", /* EMULTIDISP */
+ "too many list types", /* EMULTILIST */
+ "NAME section must be first", /* ESECNAME */
+ "badly-formed NAME section", /* ENAMESECINC */
+ "argument repeated", /* EARGREP */
+ "expected boolean parameter", /* EBOOL */
+ "inconsistent column syntax", /* ECOLMIS */
+ "nested display invalid", /* ENESTDISP */
+ "width argument missing", /* EMISSWIDTH */
+ "invalid section for this manual section", /* EWRONGMSEC */
+ "section out of conventional order", /* ESECOOO */
+ "section repeated", /* ESECREP */
+ "invalid standard argument", /* EBADSTAND */
+ "multi-line arguments discouraged", /* ENOMULTILINE */
+ "multi-line arguments suggested", /* EMULTILINE */
+ "line arguments discouraged", /* ENOLINE */
+ "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 */
+ "end-of-line scope violation", /* EIMPBRK */
+ "empty macro ignored", /* EIGNE */
+ "unclosed explicit scope", /* EOPEN */
+ "unterminated quoted phrase", /* EQUOTPHR */
+ "closure macro without prior context", /* ENOCTX */
+ "no description found for library", /* ELIB */
+ "bad child for parent context", /* EBADCHILD */
+ "list arguments preceding type", /* ENOTYPE */
+};
const char *const __mdoc_macronames[MDOC_MAX] = {
"Ap", "Dd", "Dt", "Os",
"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",
"Fr", "Ud", "Lb", "Lp",
"Lk", "Mt", "Brq", "Bro",
/* LINTED */
- "Brc", "\%C", "Es", "En",
+ "Brc", "%C", "Es", "En",
+ /* LINTED */
+ "Dx", "%Q", "br", "sp",
/* LINTED */
- "Dx", "\%Q"
+ "%U"
};
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_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);
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);
const struct mdoc_node *
mdoc_node(const struct mdoc *m)
/*
* 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);
}
* 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);
}
{
mdoc_free1(mdoc);
- if (mdoc->htab)
- mdoc_hash_free(mdoc->htab);
free(mdoc);
}
{
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);
}
int
mdoc_err(struct mdoc *m, int line, int pos, int iserr, enum merr type)
{
- char *p;
-
- p = NULL;
-
- switch (type) {
- case (ENOCALL):
- p = "not callable";
- break;
- case (EPROLBODY):
- p = "macro disallowed in document body";
- break;
- case (EBODYPROL):
- p = "macro disallowed in document prologue";
- break;
- case (EMALLOC):
- p = "memory exhausted";
- break;
- case (ETEXTPROL):
- p = "text disallowed in document prologue";
- break;
- case (ENOBLANK):
- p = "blank lines disallowed in non-literal contexts";
- break;
- case (ESPACE):
- p = "whitespace disallowed after delimiter";
- break;
- case (ETOOLONG):
- p = "text argument too long";
- break;
- case (EESCAPE):
- p = "invalid escape sequence";
- break;
- case (EPRINT):
- p = "invalid character";
- break;
- case (ENESTDISP):
- p = "displays may not be nested";
- break;
- case (EBOOL):
- p = "expected boolean value";
- break;
- case (EARGREP):
- p = "argument repeated";
- break;
- case (EMULTIDISP):
- p = "multiple display types specified";
- break;
- case (EMULTILIST):
- p = "multiple list types specified";
- break;
- case (ELISTTYPE):
- p = "missing list type";
- break;
- case (EDISPTYPE):
- p = "missing display type";
- break;
- case (ESECNAME):
- p = "the NAME section must come first";
- break;
- case (ELINE):
- p = "expected line arguments";
- break;
- case (ENOPROLOGUE):
- p = "document has no prologue";
- break;
- case (ENODAT):
- p = "document has no data";
- break;
- case (ECOLMIS):
- p = "column syntax style mismatch";
- break;
- case (EATT):
- p = "expected valid AT&T symbol";
- break;
- case (ENAME):
- p = "default name not yet set";
- break;
- case (ENOWIDTH):
- p = "superfluous width argument";
- break;
- case (EMISSWIDTH):
- p = "missing width argument";
- break;
- case (EWRONGMSEC):
- p = "document section in wrong manual section";
- break;
- case (ESECOOO):
- p = "document section out of conventional order";
- break;
- case (ESECREP):
- p = "document section repeated";
- break;
- case (EBADSTAND):
- p = "unknown standard";
- break;
- case (ENAMESECINC):
- p = "NAME section contents incomplete/badly-ordered";
- break;
- case (ENOMULTILINE):
- p = "suggested no multi-line arguments";
- break;
- case (EMULTILINE):
- p = "suggested multi-line arguments";
- break;
- case (ENOLINE):
- p = "suggested no line arguments";
- break;
- case (EPROLOOO):
- p = "prologue macros out-of-order";
- break;
- case (EPROLREP):
- p = "prologue macros repeated";
- break;
- case (EARGVAL):
- p = "argument value suggested";
- break;
- case (EFONT):
- p = "invalid font mode";
- break;
- case (EBADMSEC):
- p = "inappropriate manual section";
- break;
- case (EBADSEC):
- p = "inappropriate document section";
- break;
- case (EQUOTTERM):
- p = "unterminated quoted parameter";
- break;
- case (EQUOTPARM):
- p = "unexpected quoted parameter";
- break;
- case (EARGVPARM):
- p = "argument-like parameter";
- break;
- case (ECOLEMPTY):
- p = "last list column is empty";
- break;
- case (ETAILWS):
- p = "trailing whitespace";
- break;
- case (ENUMFMT):
- p = "bad number format";
- break;
- case (EUTSNAME):
- p = "utsname";
- break;
- case (EBADDATE):
- p = "malformed date syntax";
- break;
- case (EOPEN):
- p = "explicit scope still open on exit";
- break;
- case (EQUOT):
- p = "unterminated quotation";
- break;
- case (ENOCTX):
- p = "closure has no prior context";
- break;
- case (ENOPARMS):
- p = "unexpect line arguments";
- break;
- case (EIGNE):
- p = "ignoring empty element";
- break;
- case (EIMPBRK):
- p = "crufty end-of-line scope violation";
- break;
- case (EMACPARM):
- p = "macro-like parameter";
- break;
- case (EOBS):
- p = "macro marked obsolete";
- break;
- }
+ const char *p;
+ p = __mdoc_merrnames[(int)type];
assert(p);
if (iserr)
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));
! (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));
}
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);
- return(NULL);
- }
-
- p->sec = mdoc->lastsec;
+ p = mandoc_calloc(1, sizeof(struct mdoc_node));
+ p->sec = m->lastsec;
p->line = line;
p->pos = pos;
p->tok = tok;
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);
- if (NULL == p)
+ p = node_alloc(m, line, pos, tok, MDOC_TAIL);
+ if ( ! node_append(m, p))
return(0);
- return(node_append(mdoc, p));
+ 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);
- if (NULL == p)
+ p = node_alloc(m, line, pos, tok, MDOC_HEAD);
+ if ( ! node_append(m, p))
return(0);
- return(node_append(mdoc, p));
+ 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);
- if (NULL == p)
+ p = node_alloc(m, line, pos, tok, MDOC_BODY);
+ if ( ! node_append(m, p))
return(0);
- return(node_append(mdoc, p));
+ 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);
- if (NULL == p)
- return(0);
+ p = node_alloc(m, line, pos, tok, MDOC_BLOCK);
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);
- if (NULL == p)
- return(0);
+ p = node_alloc(m, line, pos, tok, MDOC_ELEM);
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);
+ n = node_alloc(m, line, pos, -1, MDOC_TEXT);
+ n->string = mandoc_malloc(len + 1);
+ 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);
+}
- return(node_append(mdoc, p));
+
+int
+mdoc_word_alloc(struct mdoc *m, int line, int pos, const char *p)
+{
+
+ return(pstring(m, line, pos, p, strlen(p)));
}
static int
parsetext(struct mdoc *m, int line, char *buf)
{
+ int i, j;
+ char sv;
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;
+
+ sv = buf[i];
+ buf[i++] = '\0';
+
+ if ( ! pstring(m, line, j, &buf[j], (size_t)(i - j)))
+ return(0);
+
+ /* Trailing whitespace? Check at overwritten byte. */
+
+ if (' ' == sv && '\0' == buf[i])
+ if ( ! mdoc_pwarn(m, line, i - 1, ETAILWS))
+ return(0);
+
+ for ( ; ' ' == buf[i]; i++)
+ /* Skip trailing whitespace. */ ;
+
+ j = i;
+
+ /* Trailing whitespace? */
+
+ if (' ' == buf[i - 1] && '\0' == buf[i])
+ if ( ! mdoc_pwarn(m, line, i - 1, ETAILWS))
+ return(0);
+
+ 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;
}
+
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 ? "..." : ""));
}
int
parsemacro(struct mdoc *m, int ln, char *buf)
{
- int i, c;
+ int i, j, c;
char mac[5];
/* Empty lines are ignored. */
- if (0 == buf[1])
+ 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])
+ 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);
while (buf[i] && ' ' == buf[i])
i++;
- /* Begin recursive parse sequence. */
+ /* Trailing whitespace? */
+ 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, c, ln, 1, &i, buf))
goto err;
m->flags |= MDOC_HALT;
return(0);
}
+
+
git.ckatri.com / mandoc.git / blobdiff