version 1.4, 2014/03/20 15:29:57 |
version 1.21, 2014/04/03 11:55:01 |
|
|
#include <string.h> |
#include <string.h> |
#include <unistd.h> |
#include <unistd.h> |
|
|
|
/* |
|
* In what section can we find Perl module manuals? |
|
* Sometimes (Mac OS X) it's 3pm, sometimes (OpenBSD, etc.) 3p. |
|
* XXX IF YOU CHANGE THIS, CHANGE POD2MDOC.1 AS WELL. |
|
*/ |
|
#define PERL_SECTION "3p" |
|
|
struct args { |
struct args { |
const char *title; /* override "Dt" title */ |
const char *title; /* override "Dt" title */ |
const char *date; /* override "Dd" date */ |
const char *date; /* override "Dd" date */ |
|
|
LIST__MAX |
LIST__MAX |
}; |
}; |
|
|
|
enum sect { |
|
SECT_NONE = 0, |
|
SECT_NAME, /* NAME section */ |
|
SECT_SYNOPSIS, /* SYNOPSIS section */ |
|
}; |
|
|
struct state { |
struct state { |
int parsing; /* after =cut of before command */ |
int parsing; /* after =cut of before command */ |
int paused; /* in =begin and before =end */ |
int paused; /* in =begin and before =end */ |
int haspar; /* in paragraph: do we need Pp? */ |
int haspar; /* in paragraph: do we need Pp? */ |
int isname; /* are we the NAME section? */ |
enum sect sect; /* which section are we in? */ |
const char *fname; /* file being parsed */ |
const char *fname; /* file being parsed */ |
#define LIST_STACKSZ 128 |
#define LIST_STACKSZ 128 |
enum list lstack[LIST_STACKSZ]; /* open lists */ |
enum list lstack[LIST_STACKSZ]; /* open lists */ |
Line 108 static const char fmts[FMT__MAX] = { |
|
Line 121 static const char fmts[FMT__MAX] = { |
|
'Z' /* FMT_NULL */ |
'Z' /* FMT_NULL */ |
}; |
}; |
|
|
|
static int last; |
|
|
/* |
/* |
* Given buf[*start] is at the start of an escape name, read til the end |
* Given buf[*start] is at the start of an escape name, read til the end |
* of the escape ('>') then try to do something with it. |
* of the escape ('>') then try to do something with it. |
Line 141 formatescape(const char *buf, size_t *start, size_t en |
|
Line 156 formatescape(const char *buf, size_t *start, size_t en |
|
* TODO: right now, we only recognise the named escapes. |
* TODO: right now, we only recognise the named escapes. |
* Just let the rest of them go. |
* Just let the rest of them go. |
*/ |
*/ |
if (0 == strcmp(esc, "lt")) |
if (0 == strcmp(esc, "lt")) |
printf("\\(la"); |
printf("\\(la"); |
else if (0 == strcmp(esc, "gt")) |
else if (0 == strcmp(esc, "gt")) |
printf("\\(ra"); |
printf("\\(ra"); |
Line 149 formatescape(const char *buf, size_t *start, size_t en |
|
Line 164 formatescape(const char *buf, size_t *start, size_t en |
|
printf("\\(ba"); |
printf("\\(ba"); |
else if (0 == strcmp(esc, "sol")) |
else if (0 == strcmp(esc, "sol")) |
printf("\\(sl"); |
printf("\\(sl"); |
|
else |
|
return; |
|
|
|
last = 'a'; |
} |
} |
|
|
/* |
/* |
* Skip space characters. |
* Run some heuristics to intuit a link format. |
|
* I set "start" to be the end of the sequence (last right-carrot) so |
|
* that the caller can safely just continue processing. |
|
* If this is just an empty tag, I'll return 0. |
*/ |
*/ |
|
static int |
|
trylink(const char *buf, size_t *start, size_t end, size_t dsz) |
|
{ |
|
size_t linkstart, realend, linkend, |
|
i, j, textsz, stack; |
|
const char *text; |
|
|
|
/* |
|
* Scan to the start of the terminus. |
|
* This function is more or less replicated in the formatcode() |
|
* for null or index formatting codes. |
|
*/ |
|
stack = 0; |
|
for (linkstart = realend = *start; realend < end; realend++) { |
|
if ('>' != buf[realend]) |
|
continue; |
|
else if (dsz == 1) |
|
break; |
|
assert(realend > 0); |
|
if (' ' != buf[realend - 1]) |
|
continue; |
|
for (i = realend, j = 0; i < end && j < dsz; j++) |
|
if ('>' != buf[i++]) |
|
break; |
|
if (dsz == j) |
|
break; |
|
} |
|
|
|
/* Ignore stubs. */ |
|
if (realend == end || realend == *start) |
|
return(0); |
|
|
|
/* Set linkend to the end of content. */ |
|
linkend = dsz > 1 ? realend - 1 : realend; |
|
|
|
/* Re-scan to see if we have a title or section. */ |
|
text = &buf[*start]; |
|
for (textsz = *start; textsz < linkend; textsz++) |
|
if ('|' == buf[textsz] || '/' == buf[textsz]) |
|
break; |
|
|
|
if (textsz < linkend && '|' == buf[textsz]) { |
|
/* With title: set start, then end at section. */ |
|
linkstart = textsz + 1; |
|
textsz = textsz - *start; |
|
for (i = linkstart; i < linkend; i++) |
|
if ('/' == buf[i]) |
|
break; |
|
if (i < linkend) |
|
linkend = i; |
|
} else if (textsz < linkend && '/' == buf[textsz]) { |
|
/* With section: set end at section. */ |
|
linkend = textsz; |
|
textsz = 0; |
|
} else |
|
/* No title, no section. */ |
|
textsz = 0; |
|
|
|
*start = realend; |
|
j = linkend - linkstart; |
|
|
|
/* Do we have only subsection material? */ |
|
if (0 == j && '/' == buf[linkend]) { |
|
linkstart = linkend + 1; |
|
linkend = dsz > 1 ? realend - 1 : realend; |
|
if (0 == (j = linkend - linkstart)) |
|
return(0); |
|
printf("Sx %.*s", (int)j, &buf[linkstart]); |
|
return(1); |
|
} else if (0 == j) |
|
return(0); |
|
|
|
/* See if we qualify as being a link or not. */ |
|
if ((j > 4 && 0 == memcmp("http:", &buf[linkstart], j)) || |
|
(j > 5 && 0 == memcmp("https:", &buf[linkstart], j)) || |
|
(j > 3 && 0 == memcmp("ftp:", &buf[linkstart], j)) || |
|
(j > 4 && 0 == memcmp("sftp:", &buf[linkstart], j)) || |
|
(j > 3 && 0 == memcmp("smb:", &buf[linkstart], j)) || |
|
(j > 3 && 0 == memcmp("afs:", &buf[linkstart], j))) { |
|
/* Gross. */ |
|
printf("Lk %.*s", (int)((dsz > 1 ? realend - 1 : |
|
realend) - linkstart), &buf[linkstart]); |
|
return(1); |
|
} |
|
|
|
/* See if we qualify as a mailto. */ |
|
if (j > 6 && 0 == memcmp("mailto:", &buf[linkstart], j)) { |
|
printf("Mt %.*s", (int)j, &buf[linkstart]); |
|
return(1); |
|
} |
|
|
|
/* See if we're a foo(5), foo(5x), or foo(5xx) manpage. */ |
|
if ((j > 3 && ')' == buf[linkend - 1]) && |
|
('(' == buf[linkend - 3])) { |
|
printf("Xr %.*s %c", (int)(j - 3), |
|
&buf[linkstart], buf[linkend - 2]); |
|
return(1); |
|
} else if ((j > 4 && ')' == buf[linkend - 1]) && |
|
('(' == buf[linkend - 4])) { |
|
printf("Xr %.*s %.*s", (int)(j - 4), |
|
&buf[linkstart], 2, &buf[linkend - 3]); |
|
return(1); |
|
} else if ((j > 5 && ')' == buf[linkend - 1]) && |
|
('(' == buf[linkend - 5])) { |
|
printf("Xr %.*s %.*s", (int)(j - 5), |
|
&buf[linkstart], 3, &buf[linkend - 4]); |
|
return(1); |
|
} |
|
|
|
/* Last try: do we have a double-colon? */ |
|
for (i = linkstart + 1; i < linkend; i++) |
|
if (':' == buf[i] && ':' == buf[i - 1]) |
|
break; |
|
|
|
if (i < linkend) |
|
printf("Xr %.*s " PERL_SECTION, |
|
(int)j, &buf[linkstart]); |
|
else |
|
printf("Xr %.*s 1", (int)j, &buf[linkstart]); |
|
|
|
return(1); |
|
} |
|
|
|
/* |
|
* Doclifting: if we're a bold "-xx" and we're in the SYNOPSIS section, |
|
* then it's likely that we're a flag. |
|
* Our flag might be followed by an argument, so make sure that we're |
|
* accounting for that, too. |
|
* If we don't have a flag at all, however, then assume we're an "Ar". |
|
*/ |
static void |
static void |
skipspace(const char *buf, size_t *start, size_t end) |
dosynopsisfl(const char *buf, size_t *start, size_t end) |
{ |
{ |
|
size_t i; |
|
again: |
|
assert(*start + 1 < end); |
|
assert('-' == buf[*start]); |
|
|
while (*start < end && ' ' == buf[*start]) |
if ( ! isalnum((int)buf[*start + 1]) && |
(*start)++; |
'?' != buf[*start + 1] && |
|
'-' != buf[*start + 1]) { |
|
(*start)--; |
|
fputs("Ar ", stdout); |
|
return; |
|
} |
|
|
|
(*start)++; |
|
for (i = *start; i < end; i++) |
|
if (isalnum((int)buf[i])) |
|
continue; |
|
else if ('?' == buf[i]) |
|
continue; |
|
else if ('-' == buf[i]) |
|
continue; |
|
else if ('_' == buf[i]) |
|
continue; |
|
else |
|
break; |
|
|
|
assert(i < end); |
|
|
|
if ( ! (' ' == buf[i] || '>' == buf[i])) { |
|
printf("Ar "); |
|
return; |
|
} |
|
|
|
printf("Fl "); |
|
if (end - *start > 1 && |
|
isupper((int)buf[*start]) && |
|
islower((int)buf[*start + 1]) && |
|
(end - *start == 2 || |
|
' ' == buf[*start + 2])) |
|
printf("\\&"); |
|
printf("%.*s ", (int)(i - *start), &buf[*start]); |
|
*start = i; |
|
|
|
if (' ' == buf[i]) { |
|
while (i < end && ' ' == buf[i]) |
|
i++; |
|
assert(i < end); |
|
if ('-' == buf[i]) { |
|
*start = i; |
|
goto again; |
|
} |
|
printf("Ar "); |
|
*start = i; |
|
} |
} |
} |
|
|
/* |
/* |
Line 169 skipspace(const char *buf, size_t *start, size_t end) |
|
Line 372 skipspace(const char *buf, size_t *start, size_t end) |
|
* the end of matched production. |
* the end of matched production. |
* If "reentrant", then we're being called after a macro has already |
* If "reentrant", then we're being called after a macro has already |
* been printed to the current line. |
* been printed to the current line. |
* "last" is set to the last read character: this is used to determine |
* If "nomacro", then we don't print any macros, just contained data |
* whether we should buffer with space or not. |
* (e.g., following "Sh" or "Nm"). |
* If "nomacro", then we don't print any macros, just contained data. |
* "pos" is only significant in SYNOPSIS, and should be 0 when invoked |
|
* as the first format code on a line (for decoration as an "Nm"), |
|
* non-zero otherwise. |
|
* Return whether we've printed a macro or not--in other words, whether |
|
* this should trigger a subsequent newline (this should be ignored when |
|
* reentrant). |
*/ |
*/ |
static int |
static int |
formatcode(const char *buf, size_t *start, |
formatcode(struct state *st, const char *buf, size_t *start, |
size_t end, int reentrant, int last, int nomacro) |
size_t end, int reentrant, int nomacro, int pos) |
{ |
{ |
enum fmt fmt; |
enum fmt fmt; |
|
size_t i, j, dsz; |
|
|
assert(*start + 1 < end); |
assert(*start + 1 < end); |
assert('<' == buf[*start + 1]); |
assert('<' == buf[*start + 1]); |
|
|
|
/* |
|
* First, look up the format code. |
|
* If it's not valid, then exit immediately. |
|
*/ |
for (fmt = 0; fmt < FMT__MAX; fmt++) |
for (fmt = 0; fmt < FMT__MAX; fmt++) |
if (buf[*start] == fmts[fmt]) |
if (buf[*start] == fmts[fmt]) |
break; |
break; |
|
|
/* Invalid macros are just regular text. */ |
|
|
|
if (FMT__MAX == fmt) { |
if (FMT__MAX == fmt) { |
putchar(buf[*start]); |
putchar(last = buf[(*start)++]); |
(*start)++; |
if ('\\' == last) |
|
putchar('e'); |
return(0); |
return(0); |
} |
} |
|
|
*start += 2; |
/* |
|
* Determine whether we're overriding our delimiter. |
|
* According to POD, if we have more than one '<' followed by a |
|
* space, then we need a space followed by matching '>' to close |
|
* the expression. |
|
* Otherwise we use the usual '<' and '>' matched pair. |
|
*/ |
|
i = *start + 1; |
|
while (i < end && '<' == buf[i]) |
|
i++; |
|
assert(i > *start + 1); |
|
dsz = i - (*start + 1); |
|
if (dsz > 1 && (i >= end || ' ' != buf[i])) |
|
dsz = 1; |
|
|
|
/* Remember, if dsz>1, to jump the trailing space. */ |
|
*start += dsz + 1 + (dsz > 1 ? 1 : 0); |
|
|
/* |
/* |
* Escapes don't print macro sequences, so just output them like |
* Escapes and ignored codes (NULL and INDEX) don't print macro |
* normal text before processing for macros. |
* sequences, so just output them like normal text before |
|
* processing for real macros. |
*/ |
*/ |
if (FMT_ESCAPE == fmt) { |
if (FMT_ESCAPE == fmt) { |
formatescape(buf, start, end); |
formatescape(buf, start, end); |
return(0); |
return(0); |
} else if (FMT_NULL == fmt || FMT_INDEX == fmt) { |
} else if (FMT_NULL == fmt || FMT_INDEX == fmt) { |
/* For indices and nulls, just consume. */ |
/* |
while (*start < end && '>' != buf[*start]) |
* Just consume til the end delimiter, accounting for |
(*start)++; |
* whether it's a custom one. |
if (*start < end) |
*/ |
(*start)++; |
for ( ; *start < end; (*start)++) { |
|
if ('>' != buf[*start]) |
|
continue; |
|
else if (dsz == 1) |
|
break; |
|
assert(*start > 0); |
|
if (' ' != buf[*start - 1]) |
|
continue; |
|
i = *start; |
|
for (j = 0; i < end && j < dsz; j++) |
|
if ('>' != buf[i++]) |
|
break; |
|
if (dsz != j) |
|
continue; |
|
(*start) += dsz; |
|
break; |
|
} |
return(0); |
return(0); |
} |
} |
|
|
|
/* |
|
* Check whether we're supposed to print macro stuff (this is |
|
* suppressed in, e.g., "Nm" and "Sh" macros). |
|
*/ |
if ( ! nomacro) { |
if ( ! nomacro) { |
/* |
/* |
* Print out the macro describing this format code. |
* Print out the macro describing this format code. |
Line 220 formatcode(const char *buf, size_t *start, |
|
Line 469 formatcode(const char *buf, size_t *start, |
|
* indicator. |
* indicator. |
* Otherwise, offset us with a space. |
* Otherwise, offset us with a space. |
*/ |
*/ |
if ( ! reentrant && last != '\n') |
if ( ! reentrant) { |
putchar('\n'); |
if (last != '\n') |
if ( ! reentrant) |
putchar('\n'); |
putchar('.'); |
putchar('.'); |
else |
} else |
putchar(' '); |
putchar(' '); |
|
|
/* |
/* |
* If we don't have whitespace before us, then suppress |
* If we don't have whitespace before us (and none after |
* macro whitespace with Ns. |
* the opening delimiter), then suppress macro |
|
* whitespace with Pf. |
*/ |
*/ |
if (' ' != last) |
if (' ' != last && '\n' != last && ' ' != buf[*start]) |
printf("Ns "); |
printf("Pf "); |
|
|
switch (fmt) { |
switch (fmt) { |
case (FMT_ITALIC): |
case (FMT_ITALIC): |
printf("Em "); |
printf("Em "); |
break; |
break; |
case (FMT_BOLD): |
case (FMT_BOLD): |
|
if (SECT_SYNOPSIS == st->sect) { |
|
if (1 == dsz && '-' == buf[*start]) |
|
dosynopsisfl(buf, start, end); |
|
else if (0 == pos) |
|
printf("Nm "); |
|
else |
|
printf("Ar "); |
|
break; |
|
} |
printf("Sy "); |
printf("Sy "); |
break; |
break; |
case (FMT_CODE): |
case (FMT_CODE): |
printf("Qo Li "); |
printf("Qo Li "); |
break; |
break; |
case (FMT_LINK): |
case (FMT_LINK): |
printf("Lk "); |
/* Try to link; use "No" if it's empty. */ |
|
if ( ! trylink(buf, start, end, dsz)) |
|
printf("No "); |
break; |
break; |
case (FMT_FILE): |
case (FMT_FILE): |
printf("Pa "); |
printf("Pa "); |
break; |
break; |
case (FMT_NBSP): |
case (FMT_NBSP): |
/* TODO. */ |
|
printf("No "); |
printf("No "); |
break; |
break; |
default: |
default: |
Line 259 formatcode(const char *buf, size_t *start, |
|
Line 520 formatcode(const char *buf, size_t *start, |
|
} |
} |
|
|
/* |
/* |
* Read until we reach the end market ('>') or until we find a |
* Process until we reach the end marker (e.g., '>') or until we |
* nested format code. |
* find a nested format code. |
* Don't emit any newlines: since we're on a macro line, we |
* Don't emit any newlines: since we're on a macro line, we |
* don't want to break the line. |
* don't want to break the line. |
*/ |
*/ |
while (*start < end) { |
while (*start < end) { |
if ('>' == buf[*start]) { |
if ('>' == buf[*start] && 1 == dsz) { |
(*start)++; |
(*start)++; |
break; |
break; |
|
} else if ('>' == buf[*start] && |
|
' ' == buf[*start - 1]) { |
|
/* |
|
* Handle custom delimiters. |
|
* These require a certain number of |
|
* space-preceded carrots before we're really at |
|
* the end. |
|
*/ |
|
i = *start; |
|
for (j = 0; i < end && j < dsz; j++) |
|
if ('>' != buf[i++]) |
|
break; |
|
if (dsz == j) { |
|
*start += dsz; |
|
break; |
|
} |
} |
} |
if (*start + 1 < end && '<' == buf[*start + 1]) { |
if (*start + 1 < end && '<' == buf[*start + 1]) { |
formatcode(buf, start, end, 1, last, nomacro); |
formatcode(st, buf, start, end, 1, nomacro, 1); |
continue; |
continue; |
} |
} |
|
|
Line 290 formatcode(const char *buf, size_t *start, |
|
Line 567 formatcode(const char *buf, size_t *start, |
|
printf("\\&"); |
printf("\\&"); |
|
|
/* Suppress newline. */ |
/* Suppress newline. */ |
if ('\n' == (last = buf[(*start)++])) |
if ('\n' == buf[*start]) |
last = ' '; |
putchar(last = ' '); |
|
else |
|
putchar(last = buf[*start]); |
|
|
putchar(last); |
/* Protect against character escapes. */ |
|
if ('\\' == last) |
|
putchar('e'); |
|
|
|
(*start)++; |
|
|
|
if (' ' == last) |
|
while (*start < end && ' ' == buf[*start]) |
|
(*start)++; |
} |
} |
|
|
if ( ! nomacro && FMT_CODE == fmt) |
if ( ! nomacro && FMT_CODE == fmt) |
printf(" Qc "); |
printf(" Qc "); |
|
|
if (reentrant) |
|
return(1); |
|
|
|
/* |
/* |
* If we're not reentrant, we want to put ending punctuation on |
* We're now subsequent the format code. |
* the macro line so that it's properly handled by being |
* If there isn't a space (or newline) here, and we haven't just |
* smooshed against the terminal word. |
* printed a space, then suppress space. |
*/ |
*/ |
skipspace(buf, start, end); |
if ( ! nomacro && ' ' != last) |
if (',' != buf[*start] && '.' != buf[*start] && |
if (' ' != buf[*start] && '\n' != buf[*start]) |
'!' != buf[*start] && '?' != buf[*start] && |
printf(" Ns "); |
')' != buf[*start]) |
|
return(1); |
|
while (*start < end) { |
|
if (',' != buf[*start] && |
|
'.' != buf[*start] && |
|
'!' != buf[*start] && |
|
'?' != buf[*start] && |
|
')' != buf[*start]) |
|
break; |
|
putchar(' '); |
|
putchar(buf[*start]); |
|
(*start)++; |
|
} |
|
skipspace(buf, start, end); |
|
return(1); |
return(1); |
} |
} |
|
|
Line 331 formatcode(const char *buf, size_t *start, |
|
Line 602 formatcode(const char *buf, size_t *start, |
|
* Calls formatcode() til the end of a paragraph. |
* Calls formatcode() til the end of a paragraph. |
*/ |
*/ |
static void |
static void |
formatcodeln(const char *buf, size_t *start, size_t end, int nomacro) |
formatcodeln(struct state *st, const char *buf, |
|
size_t *start, size_t end, int nomacro) |
{ |
{ |
int last; |
|
|
|
last = ' '; |
last = ' '; |
while (*start < end) { |
while (*start < end) { |
if (*start + 1 < end && '<' == buf[*start + 1]) { |
if (*start + 1 < end && '<' == buf[*start + 1]) { |
formatcode(buf, start, end, 1, last, nomacro); |
formatcode(st, buf, start, end, 1, nomacro, 1); |
continue; |
continue; |
} |
} |
/* |
/* |
Line 355 formatcodeln(const char *buf, size_t *start, size_t en |
|
Line 626 formatcodeln(const char *buf, size_t *start, size_t en |
|
' ' == buf[*start + 2])) |
' ' == buf[*start + 2])) |
printf("\\&"); |
printf("\\&"); |
|
|
if ('\n' != buf[*start]) |
if ('\n' == buf[*start]) |
putchar(last = buf[*start]); |
|
else |
|
putchar(last = ' '); |
putchar(last = ' '); |
|
else |
|
putchar(last = buf[*start]); |
|
|
|
/* Protect against character escapes. */ |
|
if ('\\' == last) |
|
putchar('e'); |
|
|
(*start)++; |
(*start)++; |
} |
} |
} |
} |
Line 414 command(struct state *st, const char *buf, size_t star |
|
Line 690 command(struct state *st, const char *buf, size_t star |
|
return; |
return; |
|
|
start += csz; |
start += csz; |
skipspace(buf, &start, end); |
while (start < end && ' ' == buf[start]) |
|
start++; |
|
|
len = end - start; |
len = end - start; |
|
|
if (st->paused) { |
if (st->paused) { |
Line 431 command(struct state *st, const char *buf, size_t star |
|
Line 709 command(struct state *st, const char *buf, size_t star |
|
* how pod2man handles it. |
* how pod2man handles it. |
*/ |
*/ |
printf(".Sh "); |
printf(".Sh "); |
st->isname = 0; |
st->sect = SECT_NONE; |
if (end - start == 4) |
if (end - start == 4) { |
if (0 == memcmp(&buf[start], "NAME", 4)) |
if (0 == memcmp(&buf[start], "NAME", 4)) |
st->isname = 1; |
st->sect = SECT_NAME; |
formatcodeln(buf, &start, end, 1); |
} else if (end - start == 8) { |
|
if (0 == memcmp(&buf[start], "SYNOPSIS", 8)) |
|
st->sect = SECT_SYNOPSIS; |
|
} |
|
formatcodeln(st, buf, &start, end, 1); |
putchar('\n'); |
putchar('\n'); |
st->haspar = 1; |
st->haspar = 1; |
break; |
break; |
case (CMD_HEAD2): |
case (CMD_HEAD2): |
printf(".Ss "); |
printf(".Ss "); |
formatcodeln(buf, &start, end, 1); |
formatcodeln(st, buf, &start, end, 1); |
putchar('\n'); |
putchar('\n'); |
st->haspar = 1; |
st->haspar = 1; |
break; |
break; |
case (CMD_HEAD3): |
case (CMD_HEAD3): |
puts(".Pp"); |
puts(".Pp"); |
printf(".Em "); |
printf(".Em "); |
formatcodeln(buf, &start, end, 0); |
formatcodeln(st, buf, &start, end, 0); |
putchar('\n'); |
putchar('\n'); |
puts(".Pp"); |
puts(".Pp"); |
st->haspar = 1; |
st->haspar = 1; |
Line 456 command(struct state *st, const char *buf, size_t star |
|
Line 738 command(struct state *st, const char *buf, size_t star |
|
case (CMD_HEAD4): |
case (CMD_HEAD4): |
puts(".Pp"); |
puts(".Pp"); |
printf(".No "); |
printf(".No "); |
formatcodeln(buf, &start, end, 0); |
formatcodeln(st, buf, &start, end, 0); |
putchar('\n'); |
putchar('\n'); |
puts(".Pp"); |
puts(".Pp"); |
st->haspar = 1; |
st->haspar = 1; |
Line 478 command(struct state *st, const char *buf, size_t star |
|
Line 760 command(struct state *st, const char *buf, size_t star |
|
st->lstack[st->lpos - 1] = LIST__MAX; |
st->lstack[st->lpos - 1] = LIST__MAX; |
break; |
break; |
case (CMD_ITEM): |
case (CMD_ITEM): |
|
if (0 == st->lpos) { |
|
/* |
|
* Bad markup. |
|
* Try to compensate. |
|
*/ |
|
st->lstack[st->lpos] = LIST__MAX; |
|
st->lpos++; |
|
} |
assert(st->lpos > 0); |
assert(st->lpos > 0); |
/* |
/* |
* If we're the first =item, guess at what our content |
* If we're the first =item, guess at what our content |
Line 502 command(struct state *st, const char *buf, size_t star |
|
Line 792 command(struct state *st, const char *buf, size_t star |
|
switch (st->lstack[st->lpos - 1]) { |
switch (st->lstack[st->lpos - 1]) { |
case (LIST_TAG): |
case (LIST_TAG): |
printf(".It "); |
printf(".It "); |
formatcodeln(buf, &start, end, 0); |
formatcodeln(st, buf, &start, end, 0); |
putchar('\n'); |
putchar('\n'); |
break; |
break; |
case (LIST_ENUM): |
case (LIST_ENUM): |
Line 559 command(struct state *st, const char *buf, size_t star |
|
Line 849 command(struct state *st, const char *buf, size_t star |
|
static void |
static void |
verbatim(struct state *st, const char *buf, size_t start, size_t end) |
verbatim(struct state *st, const char *buf, size_t start, size_t end) |
{ |
{ |
|
int last; |
|
|
if ( ! st->parsing || st->paused) |
if ( ! st->parsing || st->paused) |
return; |
return; |
|
|
puts(".Bd -literal"); |
puts(".Bd -literal"); |
printf("%.*s\n", (int)(end - start), &buf[start]); |
for (last = ' '; start < end; start++) { |
|
/* |
|
* Handle accidental macros (newline starting with |
|
* control character) and escapes. |
|
*/ |
|
if ('\n' == last) |
|
if ('.' == buf[start] || '\'' == buf[start]) |
|
printf("\\&"); |
|
putchar(last = buf[start]); |
|
if ('\\' == buf[start]) |
|
printf("e"); |
|
} |
|
putchar('\n'); |
puts(".Ed"); |
puts(".Ed"); |
} |
} |
|
|
/* |
/* |
|
* See dosynopsisop(). |
|
*/ |
|
static int |
|
hasmatch(const char *buf, size_t start, size_t end) |
|
{ |
|
size_t stack; |
|
|
|
for (stack = 0; start < end; start++) |
|
if (buf[start] == '[') |
|
stack++; |
|
else if (buf[start] == ']' && 0 == stack) |
|
return(1); |
|
else if (buf[start] == ']') |
|
stack--; |
|
return(0); |
|
} |
|
|
|
/* |
|
* If we're in the SYNOPSIS section and we've encounter braces in an |
|
* ordinary paragraph, then try to see whether we're an [-option]. |
|
* Do this, if we're an opening bracket, by first seeing if we have a |
|
* matching end via hasmatch(). |
|
* If we're an ending bracket, see if we have a stack already. |
|
*/ |
|
static int |
|
dosynopsisop(const char *buf, int *last, |
|
size_t *start, size_t end, size_t *opstack) |
|
{ |
|
|
|
assert('[' == buf[*start] || ']' == buf[*start]); |
|
|
|
if ('[' == buf[*start] && hasmatch(buf, *start + 1, end)) { |
|
if ('\n' != *last) |
|
putchar('\n'); |
|
puts(".Oo"); |
|
(*opstack)++; |
|
} else if ('[' == buf[*start]) |
|
return(0); |
|
|
|
if (']' == buf[*start] && *opstack > 0) { |
|
if ('\n' != *last) |
|
putchar('\n'); |
|
puts(".Oc"); |
|
(*opstack)--; |
|
} else if (']' == buf[*start]) |
|
return(0); |
|
|
|
(*start)++; |
|
*last = '\n'; |
|
while (' ' == buf[*start]) |
|
(*start)++; |
|
return(1); |
|
} |
|
|
|
/* |
|
* Format multiple "Nm" manpage names in the NAME section. |
|
*/ |
|
static void |
|
donamenm(struct state *st, const char *buf, size_t *start, size_t end) |
|
{ |
|
size_t word; |
|
|
|
while (*start < end && ' ' == buf[*start]) |
|
(*start)++; |
|
|
|
if (end == *start) { |
|
puts(".Nm unknown"); |
|
return; |
|
} |
|
|
|
while (*start < end) { |
|
fputs(".Nm ", stdout); |
|
for (word = *start; word < end; word++) |
|
if (',' == buf[word]) |
|
break; |
|
formatcodeln(st, buf, start, word, 1); |
|
if (*start == end) { |
|
putchar('\n'); |
|
continue; |
|
} |
|
assert(',' == buf[*start]); |
|
puts(" ,"); |
|
(*start)++; |
|
while (*start < end && ' ' == buf[*start]) |
|
(*start)++; |
|
} |
|
} |
|
|
|
/* |
* Ordinary paragraph. |
* Ordinary paragraph. |
* Well, this is really the hardest--POD seems to assume that, for |
* Well, this is really the hardest--POD seems to assume that, for |
* example, a leading space implies a newline, and so on. |
* example, a leading space implies a newline, and so on. |
Line 579 verbatim(struct state *st, const char *buf, size_t sta |
|
Line 971 verbatim(struct state *st, const char *buf, size_t sta |
|
static void |
static void |
ordinary(struct state *st, const char *buf, size_t start, size_t end) |
ordinary(struct state *st, const char *buf, size_t start, size_t end) |
{ |
{ |
int last; |
size_t i, j, opstack; |
size_t i, j; |
int seq; |
|
|
if ( ! st->parsing || st->paused) |
if ( ! st->parsing || st->paused) |
return; |
return; |
Line 591 ordinary(struct state *st, const char *buf, size_t sta |
|
Line 983 ordinary(struct state *st, const char *buf, size_t sta |
|
* we're in "name - description" format. |
* we're in "name - description" format. |
* To wit, print out a "Nm" and "Nd" in that format. |
* To wit, print out a "Nm" and "Nd" in that format. |
*/ |
*/ |
if (st->isname) { |
if (SECT_NAME == st->sect) { |
for (i = end - 1; i > start; i--) |
for (i = end - 2; i > start; i--) |
if ('-' == buf[i]) |
if ('-' == buf[i] && ' ' == buf[i + 1]) |
break; |
break; |
if ('-' == buf[i]) { |
if ('-' == buf[i]) { |
j = i; |
j = i; |
Line 601 ordinary(struct state *st, const char *buf, size_t sta |
|
Line 993 ordinary(struct state *st, const char *buf, size_t sta |
|
for ( ; i > start; i--) |
for ( ; i > start; i--) |
if ('-' != buf[i]) |
if ('-' != buf[i]) |
break; |
break; |
/* FIXME: escape macro-like words etc. */ |
donamenm(st, buf, &start, i + 1); |
printf(".Nm %.*s\n", |
start = j + 1; |
(int)((i + 1) - start), &buf[start]); |
while (start < end && ' ' == buf[start]) |
printf(".Nd %.*s\n", |
start++; |
(int)(end - (j + 1)), &buf[j + 1]); |
fputs(".Nd ", stdout); |
|
formatcodeln(st, buf, &start, end, 1); |
|
putchar('\n'); |
return; |
return; |
} |
} |
} |
} |
Line 615 ordinary(struct state *st, const char *buf, size_t sta |
|
Line 1009 ordinary(struct state *st, const char *buf, size_t sta |
|
|
|
st->haspar = 0; |
st->haspar = 0; |
last = '\n'; |
last = '\n'; |
|
opstack = 0; |
|
|
while (start < end) { |
for (seq = 0; start < end; seq++) { |
/* |
/* |
* Loop til we get either to a newline or escape. |
* Loop til we get either to a newline or escape. |
* Escape initial control characters. |
* Escape initial control characters. |
Line 630 ordinary(struct state *st, const char *buf, size_t sta |
|
Line 1025 ordinary(struct state *st, const char *buf, size_t sta |
|
printf("\\&"); |
printf("\\&"); |
else if ('\n' == last && '\'' == buf[start]) |
else if ('\n' == last && '\'' == buf[start]) |
printf("\\&"); |
printf("\\&"); |
|
/* |
|
* If we're in the SYNOPSIS, have square |
|
* brackets indicate that we're opening and |
|
* closing an optional context. |
|
*/ |
|
if (SECT_SYNOPSIS == st->sect && |
|
('[' == buf[start] || |
|
']' == buf[start]) && |
|
dosynopsisop(buf, &last, |
|
&start, end, &opstack)) |
|
continue; |
putchar(last = buf[start++]); |
putchar(last = buf[start++]); |
|
if ('\\' == last) |
|
putchar('e'); |
} |
} |
|
|
if (start < end - 1 && '<' == buf[start + 1]) { |
if (start < end - 1 && '<' == buf[start + 1]) { |
Line 640 ordinary(struct state *st, const char *buf, size_t sta |
|
Line 1048 ordinary(struct state *st, const char *buf, size_t sta |
|
* what, so print a newline now. |
* what, so print a newline now. |
* Then print the (possibly nested) macros and |
* Then print the (possibly nested) macros and |
* following that, a newline. |
* following that, a newline. |
|
* Consume all whitespace so we don't |
|
* accidentally start an implicit literal line. |
|
* If the macro ends with a flush comma or |
|
* period, let mdoc(7) handle it for us. |
*/ |
*/ |
if (formatcode(buf, &start, end, 0, last, 0)) |
if (formatcode(st, buf, &start, end, 0, 0, seq)) { |
|
if ((start == end - 1 || |
|
(start < end - 1 && |
|
(' ' == buf[start + 1] || |
|
'\n' == buf[start + 1]))) && |
|
('.' == buf[start] || |
|
',' == buf[start])) { |
|
putchar(' '); |
|
putchar(buf[start++]); |
|
} |
putchar(last = '\n'); |
putchar(last = '\n'); |
|
while (start < end && ' ' == buf[start]) |
|
start++; |
|
} |
} else if (start < end && '\n' == buf[start]) { |
} else if (start < end && '\n' == buf[start]) { |
/* |
/* |
* Print the newline only if we haven't already |
* Print the newline only if we haven't already |
Line 667 ordinary(struct state *st, const char *buf, size_t sta |
|
Line 1091 ordinary(struct state *st, const char *buf, size_t sta |
|
for ( ; start < end; start++) |
for ( ; start < end; start++) |
if (' ' != buf[start] && '\t' != buf[start]) |
if (' ' != buf[start] && '\t' != buf[start]) |
break; |
break; |
} else if (start < end) { |
} |
/* |
|
* Default: print the character. |
|
* Escape initial control characters. |
|
*/ |
|
if ('\n' == last && '.' == buf[start]) |
|
printf("\\&"); |
|
else if ('\n' == last && '\'' == buf[start]) |
|
printf("\\&"); |
|
putchar(last = buf[start++]); |
|
} |
|
} |
} |
|
|
if (last != '\n') |
if (last != '\n') |
Line 741 dofile(const struct args *args, const char *fname, |
|
Line 1155 dofile(const struct args *args, const char *fname, |
|
if (NULL != (cp = strrchr(title, '.'))) { |
if (NULL != (cp = strrchr(title, '.'))) { |
*cp++ = '\0'; |
*cp++ = '\0'; |
if (0 == strcmp(cp, "pm")) |
if (0 == strcmp(cp, "pm")) |
section = "3p"; |
section = PERL_SECTION; |
} |
} |
} |
} |
|
|