static void
append_delims(struct roff_man *mdoc, int line, int *pos, char *buf)
{
char *p;
int la;
if (buf[*pos] == '\0')
return;
for (;;) {
la = *pos;
if (mdoc_args(mdoc, line, pos, buf, TOKEN_NONE, &p) ==
ARGS_EOLN)
break;
dword(mdoc, line, la, p, DELIM_MAX, 1);
/*
* If we encounter end-of-sentence symbols, then trigger
* the double-space.
*
* XXX: it's easy to allow this to propagate outward to
* the last symbol, such that `. )' will cause the
* correct double-spacing. However, (1) groff isn't
* smart enough to do this and (2) it would require
* knowing which symbols break this behaviour, for
* example, `. ;' shouldn't propagate the double-space.
*/
if (mandoc_eos(p, strlen(p)))
mdoc->last->flags |= MDOC_EOS;
}
}
static int
append_delims(struct mdoc *m, int line, int *pos, char *buf)
{
int la;
enum margserr ac;
char *p;
if ('\0' == buf[*pos])
return(1);
for (;;) {
la = *pos;
ac = mdoc_zargs(m, line, pos, buf, ARGS_NOWARN, &p);
if (ARGS_ERROR == ac)
return(0);
else if (ARGS_EOLN == ac)
break;
assert(DELIM_NONE != mdoc_isdelim(p));
if ( ! mdoc_word_alloc(m, line, la, p))
return(0);
/*
* If we encounter end-of-sentence symbols, then trigger
* the double-space.
*
* XXX: it's easy to allow this to propogate outward to
* the last symbol, such that `. )' will cause the
* correct double-spacing. However, (1) groff isn't
* smart enough to do this and (2) it would require
* knowing which symbols break this behaviour, for
* example, `. ;' shouldn't propogate the double-space.
*/
if (mandoc_eos(p, strlen(p)))
m->last->flags |= MDOC_EOS;
}
return(1);
}
static int
man_ptext(struct man *man, int line, char *buf, int offs)
{
int i;
/* Literal free-form text whitespace is preserved. */
if (MAN_LITERAL & man->flags) {
if ( ! man_word_alloc(man, line, offs, buf + offs))
return(0);
return(man_descope(man, line, offs));
}
for (i = offs; ' ' == buf[i]; i++)
/* Skip leading whitespace. */ ;
/*
* Blank lines are ignored right after headings
* but add a single vertical space elsewhere.
*/
if ('\0' == buf[i]) {
/* Allocate a blank entry. */
if (MAN_SH != man->last->tok &&
MAN_SS != man->last->tok) {
if ( ! man_elem_alloc(man, line, offs, MAN_sp))
return(0);
man->next = MAN_NEXT_SIBLING;
}
return(1);
}
/*
* Warn if the last un-escaped character is whitespace. Then
* strip away the remaining spaces (tabs stay!).
*/
i = (int)strlen(buf);
assert(i);
if (' ' == buf[i - 1] || '\t' == buf[i - 1]) {
if (i > 1 && '\\' != buf[i - 2])
man_pmsg(man, line, i - 1, MANDOCERR_EOLNSPACE);
for (--i; i && ' ' == buf[i]; i--)
/* Spin back to non-space. */ ;
/* Jump ahead of escaped whitespace. */
i += '\\' == buf[i] ? 2 : 1;
buf[i] = '\0';
}
if ( ! man_word_alloc(man, line, offs, buf + offs))
return(0);
/*
* 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, 0))
man->last->flags |= MAN_EOS;
return(man_descope(man, line, offs));
}
/*
* Parse free-form text, that is, a line that does not begin with the
* control character.
*/
static int
mdoc_ptext(struct mdoc *m, int line, char *buf, int offs)
{
char *c, *ws, *end;
struct mdoc_node *n;
/* Ignore bogus comments. */
if ('\\' == buf[offs] &&
'.' == buf[offs + 1] &&
'"' == buf[offs + 2]) {
mdoc_pmsg(m, line, offs, MANDOCERR_BADCOMMENT);
return(1);
}
/* No text before an initial macro. */
if (SEC_NONE == m->lastnamed) {
mdoc_pmsg(m, line, offs, MANDOCERR_NOTEXT);
return(1);
}
assert(m->last);
n = m->last;
/*
* Divert directly to list processing if we're encountering a
* columnar MDOC_BLOCK with or without a prior MDOC_BLOCK entry
* (a MDOC_BODY means it's already open, in which case we should
* process within its context in the normal way).
*/
if (MDOC_Bl == n->tok && MDOC_BODY == n->type &&
LIST_column == n->norm->Bl.type) {
/* `Bl' is open without any children. */
m->flags |= MDOC_FREECOL;
return(mdoc_macro(m, MDOC_It, line, offs, &offs, buf));
}
if (MDOC_It == n->tok && MDOC_BLOCK == n->type &&
NULL != n->parent &&
MDOC_Bl == n->parent->tok &&
LIST_column == n->parent->norm->Bl.type) {
/* `Bl' has block-level `It' children. */
m->flags |= MDOC_FREECOL;
return(mdoc_macro(m, MDOC_It, line, offs, &offs, buf));
}
/*
* Search for the beginning of unescaped trailing whitespace (ws)
* and for the first character not to be output (end).
*/
/* FIXME: replace with strcspn(). */
ws = NULL;
for (c = end = buf + offs; *c; c++) {
switch (*c) {
case '-':
if (mandoc_hyph(buf + offs, c))
*c = ASCII_HYPH;
ws = NULL;
break;
case ' ':
if (NULL == ws)
ws = c;
continue;
case '\t':
/*
* Always warn about trailing tabs,
* even outside literal context,
* where they should be put on the next line.
*/
if (NULL == ws)
ws = c;
/*
* Strip trailing tabs in literal context only;
* outside, they affect the next line.
*/
if (MDOC_LITERAL & m->flags)
continue;
break;
case '\\':
/* Skip the escaped character, too, if any. */
if (c[1])
c++;
/* FALLTHROUGH */
default:
ws = NULL;
break;
}
end = c + 1;
}
*end = '\0';
if (ws)
mdoc_pmsg(m, line, (int)(ws-buf), MANDOCERR_EOLNSPACE);
if ('\0' == buf[offs] && ! (MDOC_LITERAL & m->flags)) {
mdoc_pmsg(m, line, (int)(c-buf), MANDOCERR_NOBLANKLN);
/*
* Insert a `sp' 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_sp, NULL))
return(0);
m->next = MDOC_NEXT_SIBLING;
return(1);
}
if ( ! mdoc_word_alloc(m, line, offs, buf+offs))
return(0);
if (MDOC_LITERAL & m->flags)
return(1);
/*
* 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(buf < end);
if (mandoc_eos(buf+offs, (size_t)(end-buf-offs), 0))
m->last->flags |= MDOC_EOS;
return(1);
}
static int
man_ptext(struct man *m, int line, char *buf, int offs)
{
int i;
/* Literal free-form text whitespace is preserved. */
if (MAN_LITERAL & m->flags) {
if ( ! man_word_alloc(m, line, offs, buf + offs))
return(0);
return(man_descope(m, line, offs));
}
/* Pump blank lines directly into the backend. */
for (i = offs; ' ' == buf[i]; i++)
/* Skip leading whitespace. */ ;
if ('\0' == buf[i]) {
/* Allocate a blank entry. */
if ( ! man_word_alloc(m, line, offs, ""))
return(0);
return(man_descope(m, line, offs));
}
/*
* Warn if the last un-escaped character is whitespace. Then
* strip away the remaining spaces (tabs stay!).
*/
i = (int)strlen(buf);
assert(i);
if (' ' == buf[i - 1] || '\t' == buf[i - 1]) {
if (i > 1 && '\\' != buf[i - 2])
man_pmsg(m, line, i - 1, MANDOCERR_EOLNSPACE);
for (--i; i && ' ' == buf[i]; i--)
/* Spin back to non-space. */ ;
/* Jump ahead of escaped whitespace. */
i += '\\' == buf[i] ? 2 : 1;
buf[i] = '\0';
}
if ( ! man_word_alloc(m, line, offs, buf + offs))
return(0);
/*
* 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, 0))
m->last->flags |= MAN_EOS;
return(man_descope(m, line, offs));
}
static int
blk_part_imp(MACRO_PROT_ARGS)
{
int la, nl;
enum mdoct ntok;
enum margserr ac;
char *p;
struct mdoc_node *blk; /* saved block context */
struct mdoc_node *body; /* saved body context */
struct mdoc_node *n;
nl = MDOC_NEWLINE & m->flags;
/*
* A macro that spans to the end of the line. This is generally
* (but not necessarily) called as the first macro. The block
* has a head as the immediate child, which is always empty,
* followed by zero or more opening punctuation nodes, then the
* body (which may be empty, depending on the macro), then zero
* or more closing punctuation nodes.
*/
if ( ! mdoc_block_alloc(m, line, ppos, tok, NULL))
return(0);
blk = m->last;
if ( ! mdoc_head_alloc(m, line, ppos, tok))
return(0);
if ( ! rew_sub(MDOC_HEAD, m, tok, line, ppos))
return(0);
/*
* Open the body scope "on-demand", that is, after we've
* processed all our the leading delimiters (open parenthesis,
* etc.).
*/
for (body = NULL; ; ) {
la = *pos;
ac = mdoc_args(m, line, pos, buf, tok, &p);
if (ARGS_ERROR == ac)
return(0);
if (ARGS_EOLN == ac)
break;
if (ARGS_PUNCT == ac)
break;
if (NULL == body && ARGS_QWORD != ac &&
DELIM_OPEN == mdoc_isdelim(p)) {
if ( ! dword(m, line, la, p, DELIM_OPEN))
return(0);
continue;
}
if (NULL == body) {
if ( ! mdoc_body_alloc(m, line, ppos, tok))
return(0);
body = m->last;
}
ntok = ARGS_QWORD == ac ? MDOC_MAX : lookup(tok, p);
if (MDOC_MAX == ntok) {
if ( ! dword(m, line, la, p, DELIM_MAX))
return(0);
continue;
}
if ( ! mdoc_macro(m, ntok, line, la, pos, buf))
return(0);
break;
}
/* Clean-ups to leave in a consistent state. */
if (NULL == body) {
if ( ! mdoc_body_alloc(m, line, ppos, tok))
return(0);
body = m->last;
}
for (n = body->child; n && n->next; n = n->next)
/* Do nothing. */ ;
/*
* End of sentence spacing: if the last node is a text node and
* has a trailing period, then mark it as being end-of-sentence.
*/
if (n && MDOC_TEXT == n->type && n->string)
if (mandoc_eos(n->string, strlen(n->string), 1))
n->flags |= MDOC_EOS;
/* Up-propagate the end-of-space flag. */
if (n && (MDOC_EOS & n->flags)) {
body->flags |= MDOC_EOS;
body->parent->flags |= MDOC_EOS;
}
/*
* If there is an open sub-block requiring explicit close-out,
* postpone closing out the current block
* until the rew_sub() call closing out the sub-block.
*/
for (n = m->last; n && n != body && n != blk->parent; n = n->parent) {
if (MDOC_BLOCK == n->type &&
MDOC_EXPLICIT & mdoc_macros[n->tok].flags &&
! (MDOC_VALID & n->flags)) {
make_pending(n, tok, m, line, ppos);
if ( ! mdoc_endbody_alloc(m, line, ppos,
tok, body, ENDBODY_NOSPACE))
return(0);
return(1);
}
}
/*
* If we can't rewind to our body, then our scope has already
* been closed by another macro (like `Oc' closing `Op'). This
* is ugly behaviour nodding its head to OpenBSD's overwhelming
* crufty use of `Op' breakage.
*/
if (n != body)
mandoc_vmsg(MANDOCERR_SCOPENEST, m->parse, line, ppos,
"%s broken", mdoc_macronames[tok]);
if (n && ! rew_sub(MDOC_BODY, m, tok, line, ppos))
return(0);
/* Standard appending of delimiters. */
if (nl && ! append_delims(m, line, pos, buf))
return(0);
/* Rewind scope, if applicable. */
if (n && ! rew_sub(MDOC_BLOCK, m, tok, line, ppos))
return(0);
return(1);
}
void
in_line_eoln(MACRO_PROT_ARGS)
{
int la;
char *p;
struct roff_node *n;
roff_elem_alloc(man, line, ppos, tok);
n = man->last;
for (;;) {
if (buf[*pos] != '\0' && (tok == MAN_br ||
tok == MAN_fi || tok == MAN_nf)) {
mandoc_vmsg(MANDOCERR_ARG_SKIP,
man->parse, line, *pos, "%s %s",
man_macronames[tok], buf + *pos);
break;
}
if (buf[*pos] != '\0' && man->last != n &&
(tok == MAN_PD || tok == MAN_ft || tok == MAN_sp)) {
mandoc_vmsg(MANDOCERR_ARG_EXCESS,
man->parse, line, *pos, "%s ... %s",
man_macronames[tok], buf + *pos);
break;
}
la = *pos;
if ( ! man_args(man, line, pos, buf, &p))
break;
if (man_macros[tok].flags & MAN_JOIN &&
man->last->type == ROFFT_TEXT)
roff_word_append(man, p);
else
roff_word_alloc(man, line, la, p);
}
/*
* Append MAN_EOS in case the last snipped argument
* ends with a dot, e.g. `.IR syslog (3).'
*/
if (n != man->last &&
mandoc_eos(man->last->string, strlen(man->last->string)))
man->last->flags |= MAN_EOS;
/*
* If no arguments are specified and this is MAN_SCOPED (i.e.,
* next-line scoped), then set our mode to indicate that we're
* waiting for terms to load into our context.
*/
if (n == man->last && man_macros[tok].flags & MAN_SCOPED) {
assert( ! (man_macros[tok].flags & MAN_NSCOPED));
man->flags |= MAN_ELINE;
return;
}
assert(man->last->type != ROFFT_ROOT);
man->next = ROFF_NEXT_SIBLING;
/* Rewind our element scope. */
for ( ; man->last; man->last = man->last->parent) {
man_state(man, man->last);
if (man->last == n)
break;
}
}
static int
man_ptext(struct roff_man *man, int line, char *buf, int offs)
{
int i;
/* Literal free-form text whitespace is preserved. */
if (man->flags & MAN_LITERAL) {
roff_word_alloc(man, line, offs, buf + offs);
man_descope(man, line, offs);
return 1;
}
for (i = offs; buf[i] == ' '; i++)
/* Skip leading whitespace. */ ;
/*
* Blank lines are ignored in next line scope and right
* after headings but add a single vertical space elsewhere.
*/
if (buf[i] == '\0') {
if (man->flags & (MAN_ELINE | MAN_BLINE))
mandoc_msg(MANDOCERR_BLK_BLANK, man->parse,
line, 0, NULL);
else if (man->last->tok != MAN_SH &&
man->last->tok != MAN_SS) {
roff_elem_alloc(man, line, offs, ROFF_sp);
man->next = ROFF_NEXT_SIBLING;
}
return 1;
}
/*
* Warn if the last un-escaped character is whitespace. Then
* strip away the remaining spaces (tabs stay!).
*/
i = (int)strlen(buf);
assert(i);
if (' ' == buf[i - 1] || '\t' == buf[i - 1]) {
if (i > 1 && '\\' != buf[i - 2])
mandoc_msg(MANDOCERR_SPACE_EOL, man->parse,
line, i - 1, NULL);
for (--i; i && ' ' == buf[i]; i--)
/* Spin back to non-space. */ ;
/* Jump ahead of escaped whitespace. */
i += '\\' == buf[i] ? 2 : 1;
buf[i] = '\0';
}
roff_word_alloc(man, line, offs, buf + offs);
/*
* 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))
man->last->flags |= NODE_EOS;
man_descope(man, line, offs);
return 1;
}
/* ARGSUSED */
int
in_line_eoln(MACRO_PROT_ARGS)
{
int la;
char *p;
struct man_node *n;
if ( ! man_elem_alloc(man, line, ppos, tok))
return(0);
n = man->last;
for (;;) {
la = *pos;
if ( ! man_args(man, line, pos, buf, &p))
break;
if ( ! man_word_alloc(man, line, la, p))
return(0);
}
/*
* Append MAN_EOS in case the last snipped argument
* ends with a dot, e.g. `.IR syslog (3).'
*/
if (n != man->last &&
mandoc_eos(man->last->string, strlen(man->last->string)))
man->last->flags |= MAN_EOS;
/*
* If no arguments are specified and this is MAN_SCOPED (i.e.,
* next-line scoped), then set our mode to indicate that we're
* waiting for terms to load into our context.
*/
if (n == man->last && MAN_SCOPED & man_macros[tok].flags) {
assert( ! (MAN_NSCOPED & man_macros[tok].flags));
man->flags |= MAN_ELINE;
return(1);
}
/* Set ignorable context, if applicable. */
if (MAN_NSCOPED & man_macros[tok].flags) {
assert( ! (MAN_SCOPED & man_macros[tok].flags));
man->flags |= MAN_ILINE;
}
assert(MAN_ROOT != man->last->type);
man->next = MAN_NEXT_SIBLING;
/*
* Rewind our element scope. Note that when TH is pruned, we'll
* be back at the root, so make sure that we don't clobber as
* its sibling.
*/
for ( ; man->last; man->last = man->last->parent) {
if (man->last == n)
break;
if (man->last->type == MAN_ROOT)
break;
if ( ! man_valid_post(man))
return(0);
}
assert(man->last);
/*
* Same here regarding whether we're back at the root.
*/
if (man->last->type != MAN_ROOT && ! man_valid_post(man))
return(0);
return(1);
}
