version 1.121, 2010/04/06 11:33:00 |
version 1.136, 2010/05/17 22:11:42 |
|
|
#include <string.h> |
#include <string.h> |
#include <time.h> |
#include <time.h> |
|
|
|
#include "mandoc.h" |
#include "libmdoc.h" |
#include "libmdoc.h" |
#include "libmandoc.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] = { |
const char *const __mdoc_macronames[MDOC_MAX] = { |
"Ap", "Dd", "Dt", "Os", |
"Ap", "Dd", "Dt", "Os", |
"Sh", "Ss", "Pp", "D1", |
"Sh", "Ss", "Pp", "D1", |
Line 150 static struct mdoc_node *node_alloc(struct mdoc *, int |
|
Line 96 static struct mdoc_node *node_alloc(struct mdoc *, int |
|
enum mdoct, enum mdoc_type); |
enum mdoct, enum mdoc_type); |
static int node_append(struct mdoc *, |
static int node_append(struct mdoc *, |
struct mdoc_node *); |
struct mdoc_node *); |
static int parsetext(struct mdoc *, int, char *); |
static int mdoc_ptext(struct mdoc *, int, char *, int); |
static int parsemacro(struct mdoc *, int, char *); |
static int mdoc_pmacro(struct mdoc *, int, char *, int); |
static int macrowarn(struct mdoc *, int, const char *); |
static int macrowarn(struct mdoc *, int, |
static int pstring(struct mdoc *, int, int, |
const char *, int); |
const char *, size_t); |
|
|
|
|
|
const struct mdoc_node * |
const struct mdoc_node * |
mdoc_node(const struct mdoc *m) |
mdoc_node(const struct mdoc *m) |
{ |
{ |
Line 191 mdoc_free1(struct mdoc *mdoc) |
|
Line 137 mdoc_free1(struct mdoc *mdoc) |
|
free(mdoc->meta.arch); |
free(mdoc->meta.arch); |
if (mdoc->meta.vol) |
if (mdoc->meta.vol) |
free(mdoc->meta.vol); |
free(mdoc->meta.vol); |
|
if (mdoc->meta.msec) |
|
free(mdoc->meta.msec); |
} |
} |
|
|
|
|
Line 243 mdoc_free(struct mdoc *mdoc) |
|
Line 191 mdoc_free(struct mdoc *mdoc) |
|
* Allocate volatile and non-volatile parse resources. |
* Allocate volatile and non-volatile parse resources. |
*/ |
*/ |
struct mdoc * |
struct mdoc * |
mdoc_alloc(void *data, int pflags, const struct mdoc_cb *cb) |
mdoc_alloc(void *data, int pflags, mandocmsg msg) |
{ |
{ |
struct mdoc *p; |
struct mdoc *p; |
|
|
p = mandoc_calloc(1, sizeof(struct mdoc)); |
p = mandoc_calloc(1, sizeof(struct mdoc)); |
|
|
if (cb) |
p->msg = msg; |
memcpy(&p->cb, cb, sizeof(struct mdoc_cb)); |
|
|
|
p->data = data; |
p->data = data; |
p->pflags = pflags; |
p->pflags = pflags; |
|
|
Line 280 mdoc_endparse(struct mdoc *m) |
|
Line 226 mdoc_endparse(struct mdoc *m) |
|
|
|
/* |
/* |
* Main parse routine. Parses a single line -- really just hands off to |
* 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 |
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) |
if (MDOC_HALT & m->flags) |
return(0); |
return(0); |
|
|
return('.' == *buf ? parsemacro(m, ln, buf) : |
m->flags |= MDOC_NEWLINE; |
parsetext(m, ln, buf)); |
return(('.' == buf[offs] || '\'' == buf[offs]) ? |
|
mdoc_pmacro(m, ln, buf, offs) : |
|
mdoc_ptext(m, ln, buf, offs)); |
} |
} |
|
|
|
|
int |
int |
mdoc_verr(struct mdoc *mdoc, int ln, int pos, |
mdoc_vmsg(struct mdoc *mdoc, enum mandocerr t, |
const char *fmt, ...) |
int ln, int pos, const char *fmt, ...) |
{ |
{ |
char buf[256]; |
char buf[256]; |
va_list ap; |
va_list ap; |
|
|
if (NULL == mdoc->cb.mdoc_err) |
|
return(0); |
|
|
|
va_start(ap, fmt); |
va_start(ap, fmt); |
(void)vsnprintf(buf, sizeof(buf) - 1, fmt, ap); |
vsnprintf(buf, sizeof(buf) - 1, fmt, ap); |
va_end(ap); |
va_end(ap); |
|
|
return((*mdoc->cb.mdoc_err)(mdoc->data, ln, pos, buf)); |
return((*mdoc->msg)(t, mdoc->data, ln, pos, buf)); |
} |
} |
|
|
|
|
int |
int |
mdoc_vwarn(struct mdoc *mdoc, int ln, int pos, const char *fmt, ...) |
|
{ |
|
char buf[256]; |
|
va_list ap; |
|
|
|
if (NULL == mdoc->cb.mdoc_warn) |
|
return(0); |
|
|
|
va_start(ap, fmt); |
|
(void)vsnprintf(buf, sizeof(buf) - 1, fmt, ap); |
|
va_end(ap); |
|
|
|
return((*mdoc->cb.mdoc_warn)(mdoc->data, ln, pos, buf)); |
|
} |
|
|
|
|
|
int |
|
mdoc_err(struct mdoc *m, int line, int pos, int iserr, enum merr type) |
|
{ |
|
const char *p; |
|
|
|
p = __mdoc_merrnames[(int)type]; |
|
assert(p); |
|
|
|
if (iserr) |
|
return(mdoc_verr(m, line, pos, p)); |
|
|
|
return(mdoc_vwarn(m, line, pos, p)); |
|
} |
|
|
|
|
|
int |
|
mdoc_macro(struct mdoc *m, enum mdoct tok, |
mdoc_macro(struct mdoc *m, enum mdoct tok, |
int ln, int pp, int *pos, char *buf) |
int ln, int pp, int *pos, char *buf) |
{ |
{ |
|
|
assert(tok < MDOC_MAX); |
assert(tok < MDOC_MAX); |
/* |
|
* If we're in the prologue, deny "body" macros. Similarly, if |
/* If we're in the body, deny prologue calls. */ |
* we're in the body, deny prologue calls. |
|
*/ |
|
if (MDOC_PROLOGUE & mdoc_macros[tok].flags && |
if (MDOC_PROLOGUE & mdoc_macros[tok].flags && |
MDOC_PBODY & m->flags) { |
MDOC_PBODY & m->flags) |
if ( ! mdoc_pwarn(m, ln, pp, EBODYPROL)) |
return(mdoc_pmsg(m, ln, pp, MANDOCERR_BADBODY)); |
|
|
|
/* If we're in the prologue, deny "body" macros. */ |
|
|
|
if ( ! (MDOC_PROLOGUE & mdoc_macros[tok].flags) && |
|
! (MDOC_PBODY & m->flags)) { |
|
if ( ! mdoc_pmsg(m, ln, pp, MANDOCERR_BADPROLOG)) |
return(0); |
return(0); |
/* |
|
* FIXME: do this in mdoc_action.c. |
|
*/ |
|
if (NULL == m->meta.title) |
if (NULL == m->meta.title) |
m->meta.title = mandoc_strdup("unknown"); |
m->meta.title = mandoc_strdup("unknown"); |
if (NULL == m->meta.vol) |
if (NULL == m->meta.vol) |
Line 371 mdoc_macro(struct mdoc *m, enum mdoct tok, |
|
Line 285 mdoc_macro(struct mdoc *m, enum mdoct tok, |
|
m->meta.date = time(NULL); |
m->meta.date = time(NULL); |
m->flags |= MDOC_PBODY; |
m->flags |= MDOC_PBODY; |
} |
} |
if ( ! (MDOC_PROLOGUE & mdoc_macros[tok].flags) && |
|
! (MDOC_PBODY & m->flags)) |
|
return(mdoc_perr(m, ln, pp, EBODYPROL)); |
|
|
|
return((*mdoc_macros[tok].fp)(m, tok, ln, pp, pos, buf)); |
return((*mdoc_macros[tok].fp)(m, tok, ln, pp, pos, buf)); |
} |
} |
Line 455 node_alloc(struct mdoc *m, int line, int pos, |
|
Line 366 node_alloc(struct mdoc *m, int line, int pos, |
|
p->pos = pos; |
p->pos = pos; |
p->tok = tok; |
p->tok = tok; |
p->type = type; |
p->type = type; |
|
if (MDOC_NEWLINE & m->flags) |
|
p->flags |= MDOC_LINE; |
|
m->flags &= ~MDOC_NEWLINE; |
return(p); |
return(p); |
} |
} |
|
|
Line 536 mdoc_elem_alloc(struct mdoc *m, int line, int pos, |
|
Line 449 mdoc_elem_alloc(struct mdoc *m, int line, int pos, |
|
} |
} |
|
|
|
|
static int |
int |
pstring(struct mdoc *m, int line, int pos, const char *p, size_t len) |
mdoc_word_alloc(struct mdoc *m, int line, int pos, const char *p) |
{ |
{ |
struct mdoc_node *n; |
struct mdoc_node *n; |
size_t sv; |
size_t sv, len; |
|
|
n = node_alloc(m, line, pos, -1, MDOC_TEXT); |
len = strlen(p); |
|
|
|
n = node_alloc(m, line, pos, MDOC_MAX, MDOC_TEXT); |
n->string = mandoc_malloc(len + 1); |
n->string = mandoc_malloc(len + 1); |
sv = strlcpy(n->string, p, len + 1); |
sv = strlcpy(n->string, p, len + 1); |
|
|
Line 551 pstring(struct mdoc *m, int line, int pos, const char |
|
Line 466 pstring(struct mdoc *m, int line, int pos, const char |
|
|
|
if ( ! node_append(m, n)) |
if ( ! node_append(m, n)) |
return(0); |
return(0); |
|
|
m->next = MDOC_NEXT_SIBLING; |
m->next = MDOC_NEXT_SIBLING; |
return(1); |
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 |
void |
mdoc_node_free(struct mdoc_node *p) |
mdoc_node_free(struct mdoc_node *p) |
{ |
{ |
Line 632 mdoc_node_delete(struct mdoc *m, struct mdoc_node *p) |
|
Line 540 mdoc_node_delete(struct mdoc *m, struct mdoc_node *p) |
|
* control character. |
* control character. |
*/ |
*/ |
static int |
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; |
char sv; |
|
|
|
|
/* Ignore bogus comments. */ |
|
|
|
if ('\\' == buf[offs] && |
|
'.' == buf[offs + 1] && |
|
'"' == buf[offs + 2]) |
|
return(mdoc_pmsg(m, line, offs, MANDOCERR_BADCOMMENT)); |
|
|
|
/* No text before an initial macro. */ |
|
|
if (SEC_NONE == m->lastnamed) |
if (SEC_NONE == m->lastnamed) |
return(mdoc_perr(m, line, 0, ETEXTPROL)); |
return(mdoc_pmsg(m, line, offs, MANDOCERR_NOTEXT)); |
|
|
/* |
|
* If in literal mode, then pass the buffer directly to the |
|
* back-end, as it should be preserved as a single term. |
|
*/ |
|
|
|
|
/* Literal just gets pulled in as-is. */ |
|
|
if (MDOC_LITERAL & m->flags) |
if (MDOC_LITERAL & m->flags) |
return(mdoc_word_alloc(m, line, 0, buf)); |
return(mdoc_word_alloc(m, line, offs, buf + offs)); |
|
|
/* Disallow blank/white-space lines in non-literal mode. */ |
/* Check for a blank line, which may also consist of spaces. */ |
|
|
for (i = 0; ' ' == buf[i]; i++) |
for (i = offs; ' ' == buf[i]; i++) |
/* Skip leading whitespace. */ ; |
/* Skip to first non-space. */ ; |
|
|
if ('\0' == buf[i]) { |
if ('\0' == buf[i]) { |
if ( ! mdoc_pwarn(m, line, 0, ENOBLANK)) |
if ( ! mdoc_pmsg(m, line, offs, MANDOCERR_NOBLANKLN)) |
return(0); |
return(0); |
|
|
/* |
/* |
* Assume that a `Pp' should be inserted in the case of |
* Insert a `Pp' in the case of a blank line. Technically, |
* a blank line. Technically, blank lines aren't |
* blank lines aren't allowed, but enough manuals assume this |
* allowed, but enough manuals assume this behaviour |
* behaviour that we want to work around it. |
* that we want to work around it. |
|
*/ |
*/ |
if ( ! mdoc_elem_alloc(m, line, 0, MDOC_Pp, NULL)) |
if ( ! mdoc_elem_alloc(m, line, offs, MDOC_Pp, NULL)) |
return(0); |
return(0); |
|
|
|
m->next = MDOC_NEXT_SIBLING; |
|
return(1); |
} |
} |
|
|
/* |
/* |
* Break apart a free-form line into tokens. Spaces are |
* Warn if the last un-escaped character is whitespace. Then |
* stripped out of the input. |
* strip away the remaining spaces (tabs stay!). |
*/ |
*/ |
|
|
for (j = i; buf[i]; i++) { |
i = (int)strlen(buf); |
if (' ' != buf[i]) |
assert(i); |
continue; |
|
|
|
/* Escaped whitespace. */ |
if (' ' == buf[i - 1] || '\t' == buf[i - 1]) { |
if (i && ' ' == buf[i] && '\\' == buf[i - 1]) |
if (i > 1 && '\\' != buf[i - 2]) |
continue; |
if ( ! mdoc_pmsg(m, line, i - 1, MANDOCERR_EOLNSPACE)) |
|
|
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); |
return(0); |
|
|
for ( ; ' ' == buf[i]; i++) |
for (--i; i && ' ' == buf[i]; i--) |
/* Skip trailing whitespace. */ ; |
/* Spin back to non-space. */ ; |
|
|
j = i; |
/* Jump ahead of escaped whitespace. */ |
|
i += '\\' == buf[i] ? 2 : 1; |
|
|
/* Trailing whitespace? */ |
buf[i] = '\0'; |
|
|
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))) |
/* Allocate the whole word. */ |
|
|
|
if ( ! mdoc_word_alloc(m, line, offs, buf + offs)) |
return(0); |
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); |
return(1); |
} |
} |
|
|
|
|
|
|
static int |
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)) |
int rc; |
return(mdoc_verr(m, ln, 0, |
|
"unknown macro: %s%s", |
rc = mdoc_vmsg(m, MANDOCERR_MACRO, ln, offs, |
buf, strlen(buf) > 3 ? "..." : "")); |
"unknown macro: %s%s", |
return(mdoc_vwarn(m, ln, 0, "unknown macro: %s%s", |
buf, strlen(buf) > 3 ? "..." : ""); |
buf, strlen(buf) > 3 ? "..." : "")); |
|
|
/* FIXME: logic should be in driver. */ |
|
return(MDOC_IGN_MACRO & m->pflags ? rc : 0); |
} |
} |
|
|
|
|
Line 732 macrowarn(struct mdoc *m, int ln, const char *buf) |
|
Line 642 macrowarn(struct mdoc *m, int ln, const char *buf) |
|
* character. |
* character. |
*/ |
*/ |
int |
int |
parsemacro(struct mdoc *m, int ln, char *buf) |
mdoc_pmacro(struct mdoc *m, int ln, char *buf, int offs) |
{ |
{ |
int i, j, c; |
enum mdoct tok; |
char mac[5]; |
int i, j, sv; |
|
char mac[5]; |
|
|
/* Empty lines are ignored. */ |
/* Empty lines are ignored. */ |
|
|
if ('\0' == buf[1]) |
offs++; |
|
|
|
if ('\0' == buf[offs]) |
return(1); |
return(1); |
|
|
i = 1; |
i = offs; |
|
|
/* Accept whitespace after the initial control char. */ |
/* Accept whitespace after the initial control char. */ |
|
|
Line 754 parsemacro(struct mdoc *m, int ln, char *buf) |
|
Line 667 parsemacro(struct mdoc *m, int ln, char *buf) |
|
return(1); |
return(1); |
} |
} |
|
|
|
sv = i; |
|
|
/* Copy the first word into a nil-terminated buffer. */ |
/* Copy the first word into a nil-terminated buffer. */ |
|
|
for (j = 0; j < 4; j++, i++) { |
for (j = 0; j < 4; j++, i++) { |
Line 766 parsemacro(struct mdoc *m, int ln, char *buf) |
|
Line 681 parsemacro(struct mdoc *m, int ln, char *buf) |
|
|
|
if (isgraph((u_char)buf[i])) |
if (isgraph((u_char)buf[i])) |
continue; |
continue; |
return(mdoc_perr(m, ln, i, EPRINT)); |
if ( ! mdoc_pmsg(m, ln, i, MANDOCERR_BADCHAR)) |
|
return(0); |
|
i--; |
} |
} |
|
|
mac[j] = 0; |
mac[j] = '\0'; |
|
|
if (j == 4 || j < 2) { |
if (j == 4 || j < 2) { |
if ( ! macrowarn(m, ln, mac)) |
if ( ! macrowarn(m, ln, mac, sv)) |
goto err; |
goto err; |
return(1); |
return(1); |
} |
} |
|
|
if (MDOC_MAX == (c = mdoc_hash_find(mac))) { |
if (MDOC_MAX == (tok = mdoc_hash_find(mac))) { |
if ( ! macrowarn(m, ln, mac)) |
if ( ! macrowarn(m, ln, mac, sv)) |
goto err; |
goto err; |
return(1); |
return(1); |
} |
} |
Line 788 parsemacro(struct mdoc *m, int ln, char *buf) |
|
Line 705 parsemacro(struct mdoc *m, int ln, char *buf) |
|
while (buf[i] && ' ' == buf[i]) |
while (buf[i] && ' ' == buf[i]) |
i++; |
i++; |
|
|
/* Trailing whitespace? */ |
/* |
|
* 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]) |
if ('\0' == buf[i] && ' ' == buf[i - 1]) |
if ( ! mdoc_pwarn(m, ln, i - 1, ETAILWS)) |
if ( ! mdoc_pmsg(m, ln, i - 1, MANDOCERR_EOLNSPACE)) |
goto err; |
goto err; |
|
|
/* |
/* |
* Begin recursive parse sequence. Since we're at the start of |
* Begin recursive parse sequence. Since we're at the start of |
* the line, we don't need to do callable/parseable checks. |
* the line, we don't need to do callable/parseable checks. |
*/ |
*/ |
if ( ! mdoc_macro(m, c, ln, 1, &i, buf)) |
if ( ! mdoc_macro(m, tok, ln, sv, &i, buf)) |
goto err; |
goto err; |
|
|
return(1); |
return(1); |