version 1.31, 2014/07/15 19:03:07 |
version 1.44, 2015/02/14 12:47:09 |
|
|
/* $Id$ */ |
/* $Id$ */ |
/* |
/* |
* Copyright (c) 2014 Kristaps Dzonsons <kristaps@bsd.lv> |
* Copyright (c) 2014 Kristaps Dzonsons <kristaps@bsd.lv> |
|
* Copyright (c) 2014, 2015 Ingo Schwarze <schwarze@openbsd.org> |
* |
* |
* Permission to use, copy, modify, and distribute this software for any |
* Permission to use, copy, modify, and distribute this software for any |
* purpose with or without fee is hereby granted, provided that the above |
* purpose with or without fee is hereby granted, provided that the above |
|
|
#include <string.h> |
#include <string.h> |
#include <unistd.h> |
#include <unistd.h> |
|
|
|
#include "dict.h" |
|
|
/* |
/* |
* In what section can we find Perl module manuals? |
* In what section can we find Perl module manuals? |
* Sometimes (Mac OS X) it's 3pm, sometimes (OpenBSD, etc.) 3p. |
* Sometimes (Mac OS X) it's 3pm, sometimes (OpenBSD, etc.) 3p. |
|
|
SECT_SYNOPSIS, /* SYNOPSIS section */ |
SECT_SYNOPSIS, /* SYNOPSIS section */ |
}; |
}; |
|
|
|
enum outstate { |
|
OUST_NL = 0, /* just started a new output line */ |
|
OUST_TXT, /* text line output in progress */ |
|
OUST_MAC /* macro line output in progress */ |
|
}; |
|
|
struct state { |
struct state { |
const char *fname; /* file being parsed */ |
const char *fname; /* file being parsed */ |
int parsing; /* after =cut of before command */ |
int parsing; /* after =cut of before command */ |
|
|
enum list lstack[LIST_STACKSZ]; /* open lists */ |
enum list lstack[LIST_STACKSZ]; /* open lists */ |
size_t lpos; /* where in list stack */ |
size_t lpos; /* where in list stack */ |
int haspar; /* in paragraph: do we need Pp? */ |
int haspar; /* in paragraph: do we need Pp? */ |
int hasnl; /* in text: just started a new line */ |
enum outstate oust; /* state of the mdoc output stream */ |
|
int wantws; /* let mdoc(7) output whitespace here */ |
char *outbuf; /* text buffered for output */ |
char *outbuf; /* text buffered for output */ |
size_t outbufsz; /* allocated size of outbuf */ |
size_t outbufsz; /* allocated size of outbuf */ |
size_t outbuflen; /* current length of outbuf */ |
size_t outbuflen; /* current length of outbuf */ |
Line 125 static const char fmts[FMT__MAX] = { |
|
Line 135 static const char fmts[FMT__MAX] = { |
|
'Z' /* FMT_NULL */ |
'Z' /* FMT_NULL */ |
}; |
}; |
|
|
static int last; |
static unsigned char last; |
|
|
|
|
static void |
static void |
Line 161 outbuf_addstr(struct state *st, const char *str) |
|
Line 171 outbuf_addstr(struct state *st, const char *str) |
|
if (st->outbuflen + slen >= st->outbufsz) |
if (st->outbuflen + slen >= st->outbufsz) |
outbuf_grow(st, slen); |
outbuf_grow(st, slen); |
memcpy(st->outbuf + st->outbuflen, str, slen+1); |
memcpy(st->outbuf + st->outbuflen, str, slen+1); |
|
st->outbuflen += slen; |
last = str[slen - 1]; |
last = str[slen - 1]; |
} |
} |
|
|
Line 171 outbuf_flush(struct state *st) |
|
Line 182 outbuf_flush(struct state *st) |
|
if (0 == st->outbuflen) |
if (0 == st->outbuflen) |
return; |
return; |
|
|
|
if (OUST_TXT == st->oust && st->wantws) |
|
putchar(' '); |
|
|
fputs(st->outbuf, stdout); |
fputs(st->outbuf, stdout); |
*st->outbuf = '\0'; |
*st->outbuf = '\0'; |
st->outbuflen = 0; |
st->outbuflen = 0; |
st->hasnl = 0; |
|
|
if (OUST_NL == st->oust) |
|
st->oust = OUST_TXT; |
} |
} |
|
|
static void |
static void |
outbuf_newln(struct state *st) |
mdoc_newln(struct state *st) |
{ |
{ |
|
|
if ('\n' == last) |
if (OUST_NL == st->oust) |
return; |
return; |
outbuf_flush(st); |
|
putchar('\n'); |
putchar('\n'); |
last = '\n'; |
last = '\n'; |
st->hasnl = 1; |
st->oust = OUST_NL; |
|
st->wantws = 1; |
} |
} |
|
|
/* |
/* |
* 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. |
* Sets start to be one after the '>'. |
* Sets start to be one after the '>'. |
|
* |
|
* This function does not care about output modes, |
|
* it merely appends text to the output buffer, |
|
* which can then be used in any mode. |
*/ |
*/ |
static void |
static void |
formatescape(struct state *st, const char *buf, size_t *start, size_t end) |
formatescape(struct state *st, const char *buf, size_t *start, size_t end) |
Line 226 formatescape(struct state *st, const char *buf, size_t |
|
Line 247 formatescape(struct state *st, const char *buf, size_t |
|
outbuf_addstr(st, "\\(la"); |
outbuf_addstr(st, "\\(la"); |
else if (0 == strcmp(esc, "gt")) |
else if (0 == strcmp(esc, "gt")) |
outbuf_addstr(st, "\\(ra"); |
outbuf_addstr(st, "\\(ra"); |
else if (0 == strcmp(esc, "vb")) |
else if (0 == strcmp(esc, "verbar")) |
outbuf_addstr(st, "\\(ba"); |
outbuf_addstr(st, "\\(ba"); |
else if (0 == strcmp(esc, "sol")) |
else if (0 == strcmp(esc, "sol")) |
outbuf_addstr(st, "\\(sl"); |
outbuf_addstr(st, "\\(sl"); |
Line 237 formatescape(struct state *st, const char *buf, size_t |
|
Line 258 formatescape(struct state *st, const char *buf, size_t |
|
* I set "start" to be the end of the sequence (last right-carrot) so |
* I set "start" to be the end of the sequence (last right-carrot) so |
* that the caller can safely just continue processing. |
* that the caller can safely just continue processing. |
* If this is just an empty tag, I'll return 0. |
* If this is just an empty tag, I'll return 0. |
|
* |
|
* Always operates in OUST_MAC mode. |
|
* Mode handling is done by the caller. |
*/ |
*/ |
static int |
static int |
trylink(const char *buf, size_t *start, size_t end, size_t dsz) |
trylink(const char *buf, size_t *start, size_t end, size_t dsz) |
Line 371 trylink(const char *buf, size_t *start, size_t end, si |
|
Line 395 trylink(const char *buf, size_t *start, size_t end, si |
|
* Our flag might be followed by an argument, so make sure that we're |
* Our flag might be followed by an argument, so make sure that we're |
* accounting for that, too. |
* accounting for that, too. |
* If we don't have a flag at all, however, then assume we're an "Ar". |
* If we don't have a flag at all, however, then assume we're an "Ar". |
|
* |
|
* Always operates in OUST_MAC mode. |
|
* Mode handlinf is done by the caller. |
*/ |
*/ |
static void |
static void |
dosynopsisfl(const char *buf, size_t *start, size_t end) |
dosynopsisfl(const char *buf, size_t *start, size_t end) |
|
|
* like X<...> and can contain nested format codes. |
* like X<...> and can contain nested format codes. |
* This consumes the whole format code, and any nested format codes, til |
* This consumes the whole format code, and any nested format codes, til |
* the end of matched production. |
* the end of matched production. |
* If "reentrant", then we're being called after a macro has already |
|
* been printed to the current line. |
|
* If "nomacro", then we don't print any macros, just contained data |
* If "nomacro", then we don't print any macros, just contained data |
* (e.g., following "Sh" or "Nm"). |
* (e.g., following "Sh" or "Nm"). |
* "pos" is only significant in SYNOPSIS, and should be 0 when invoked |
* "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"), |
* as the first format code on a line (for decoration as an "Nm"), |
* non-zero otherwise. |
* 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 |
* Output mode handling is most complicated here. |
* reentrant). |
* We may enter in any mode. |
|
* We usually exit in OUST_MAC mode, except when |
|
* entering without OUST_MAC and the code is invalid. |
*/ |
*/ |
static int |
static int |
formatcode(struct state *st, const char *buf, size_t *start, |
formatcode(struct state *st, const char *buf, size_t *start, |
size_t end, int reentrant, int nomacro, int pos) |
size_t end, int nomacro, int pos) |
{ |
{ |
enum fmt fmt; |
|
size_t i, j, dsz; |
size_t i, j, dsz; |
int white; |
enum fmt fmt; |
|
int wantws; |
|
unsigned char uc; |
|
|
assert(*start + 1 < end); |
assert(*start + 1 < end); |
assert('<' == buf[*start + 1]); |
assert('<' == buf[*start + 1]); |
Line 529 formatcode(struct state *st, const char *buf, size_t * |
|
Line 557 formatcode(struct state *st, const char *buf, size_t * |
|
* suppressed in, e.g., "Nm" and "Sh" macros). |
* suppressed in, e.g., "Nm" and "Sh" macros). |
*/ |
*/ |
if (FMT__MAX != fmt && !nomacro) { |
if (FMT__MAX != fmt && !nomacro) { |
white = ' ' == last || '\n' == last || |
|
' ' == buf[*start]; |
|
|
|
/* |
/* |
|
* We may already have wantws if there was whitespace |
|
* before the code ("text B<text"), but initial |
|
* whitespace inside our scope ("textB< text") |
|
* allows to break at this point as well. |
|
*/ |
|
|
|
wantws = ' ' == buf[*start] || |
|
(OUST_MAC == st->oust ? st->wantws : ! st->outbuflen); |
|
|
|
/* |
* If we are on a text line and there is no |
* If we are on a text line and there is no |
* whitespace before our content, we have to make |
* whitespace before our content, we have to make |
* the previous word a prefix to the macro line. |
* the previous word a prefix to the macro line. |
|
* In the following, mdoc_newln() must not be used |
|
* lest we clobber out output state. |
*/ |
*/ |
|
|
if ( ! white && ! reentrant) { |
if (OUST_MAC != st->oust && ! wantws) { |
if ( ! st->hasnl) |
if (OUST_NL != st->oust) |
putchar('\n'); |
putchar('\n'); |
printf(".Pf "); |
printf(".Pf "); |
|
st->wantws = 0; |
} |
} |
|
|
outbuf_flush(st); |
outbuf_flush(st); |
|
|
/* Whitespace is easier to suppress on macro lines. */ |
/* Whitespace is easier to suppress on macro lines. */ |
|
|
if ( ! white && reentrant) |
if (OUST_MAC == st->oust && ! wantws) |
printf(" Ns"); |
printf(" Ns "); |
|
|
/* Unless we are on a macro line, start one. */ |
/* Unless we are on a macro line, start one. */ |
|
|
if (white && ! reentrant) { |
if (OUST_MAC != st->oust && wantws) { |
if (last != '\n') |
if (OUST_NL != st->oust) |
putchar('\n'); |
putchar('\n'); |
putchar('.'); |
putchar('.'); |
} else |
} else |
putchar(' '); |
putchar(' '); |
|
|
/* Print the macro corresponding to this format code. */ |
/* |
|
* Print the macro corresponding to this format code, |
|
* and update the output state afterwards. |
|
*/ |
|
|
switch (fmt) { |
switch (fmt) { |
case (FMT_ITALIC): |
case (FMT_ITALIC): |
Line 575 formatcode(struct state *st, const char *buf, size_t * |
|
Line 617 formatcode(struct state *st, const char *buf, size_t * |
|
else |
else |
printf("Ar "); |
printf("Ar "); |
break; |
break; |
} |
} |
if (0 == strncmp(buf + *start, "NULL", 4) && |
i = 0; |
('=' == buf[*start + 4] || |
uc = buf[*start]; |
'>' == buf[*start + 4])) |
while (isalnum(uc) || '_' == uc || ' ' == uc) |
|
uc = buf[*start + ++i]; |
|
if ('=' != uc && '>' != uc) |
|
i = 0; |
|
if (4 == i && ! strncmp(buf + *start, "NULL", 4)) { |
printf("Dv "); |
printf("Dv "); |
else |
break; |
|
} |
|
switch (i ? dict_get(buf + *start, i) : MDOC_MAX) { |
|
case MDOC_Fa: |
|
printf("Fa "); |
|
break; |
|
case MDOC_Vt: |
|
printf("Vt "); |
|
break; |
|
default: |
printf("Sy "); |
printf("Sy "); |
|
break; |
|
} |
break; |
break; |
case (FMT_CODE): |
case (FMT_CODE): |
printf("Qo Li "); |
printf("Qo Li "); |
Line 600 formatcode(struct state *st, const char *buf, size_t * |
|
Line 657 formatcode(struct state *st, const char *buf, size_t * |
|
default: |
default: |
abort(); |
abort(); |
} |
} |
|
st->oust = OUST_MAC; |
|
st->wantws = 1; |
} else |
} else |
outbuf_flush(st); |
outbuf_flush(st); |
|
|
Line 630 formatcode(struct state *st, const char *buf, size_t * |
|
Line 689 formatcode(struct state *st, const char *buf, size_t * |
|
break; |
break; |
} |
} |
} |
} |
if (*start + 1 < end && '<' == buf[*start + 1]) { |
if (*start + 1 < end && '<' == buf[*start + 1] && |
formatcode(st, buf, start, end, 1, nomacro, 1); |
'A' <= buf[*start] && 'Z' >= buf[*start]) { |
|
if ( ! formatcode(st, buf, start, end, nomacro, 1)) |
|
st->wantws = 1; |
continue; |
continue; |
} |
} |
|
|
/* |
/* Suppress newlines and multiple spaces. */ |
* Make sure that any macro-like words (or |
|
* really any word starting with a capital |
|
* letter) is assumed to be a macro that must be |
|
* escaped. |
|
* This matches "Xx " and "XxEOLN". |
|
*/ |
|
if ((' ' == last || '\n' == last) && |
|
end - *start > 1 && |
|
isupper((int)buf[*start]) && |
|
islower((int)buf[*start + 1]) && |
|
(end - *start == 2 || |
|
' ' == buf[*start + 2])) |
|
printf("\\&"); |
|
|
|
/* Suppress newline. */ |
last = buf[(*start)++]; |
if ('\n' == buf[*start]) |
if (' ' == last || '\n' == last) { |
putchar(last = ' '); |
putchar(' '); |
else |
while (*start < end && ' ' == buf[*start]) |
putchar(last = buf[*start]); |
(*start)++; |
|
continue; |
|
} |
|
|
|
if (OUST_MAC == st->oust && FMT__MAX != fmt) { |
|
if ( ! st->wantws) { |
|
printf(" Ns "); |
|
st->wantws = 1; |
|
} |
|
|
|
/* |
|
* Escape macro-like words. |
|
* This matches "Xx " and "XxEOLN". |
|
*/ |
|
|
|
if (end - *start > 0 && |
|
isupper((unsigned char)last) && |
|
islower((unsigned char)buf[*start]) && |
|
(end - *start == 1 || |
|
' ' == buf[*start + 1] || |
|
'>' == buf[*start + 1])) |
|
printf("\\&"); |
|
} |
|
|
|
putchar(last); |
|
|
/* Protect against character escapes. */ |
/* Protect against character escapes. */ |
|
|
if ('\\' == last) |
if ('\\' == last) |
putchar('e'); |
putchar('e'); |
|
|
(*start)++; |
|
|
|
if (' ' == last) |
|
while (*start < end && ' ' == buf[*start]) |
|
(*start)++; |
|
} |
} |
|
|
if (FMT__MAX == fmt) |
|
return(0); |
|
|
|
if ( ! nomacro && FMT_CODE == fmt) |
if ( ! nomacro && FMT_CODE == fmt) |
printf(" Qc "); |
printf(" Qc "); |
|
|
/* |
st->wantws = ' ' == last; |
* We're now subsequent the format code. |
return(FMT__MAX != fmt); |
* If there isn't a space (or newline) here, and we haven't just |
|
* printed a space, then suppress space. |
|
*/ |
|
if ( ! nomacro && ' ' != last) |
|
if (' ' != buf[*start] && '\n' != buf[*start]) |
|
printf(" Ns "); |
|
|
|
return(1); |
|
} |
} |
|
|
/* |
/* |
* Calls formatcode() til the end of a paragraph. |
* Calls formatcode() til the end of a paragraph. |
|
* Goes to OUST_MAC mode and stays there when returning, |
|
* such that the caller can add arguments to the macro line |
|
* before closing it out. |
*/ |
*/ |
static void |
static void |
formatcodeln(struct state *st, const char *buf, |
formatcodeln(struct state *st, const char *linemac, |
size_t *start, size_t end, int nomacro) |
const char *buf, size_t *start, size_t end, int nomacro) |
{ |
{ |
|
int gotmacro, wantws; |
|
|
last = ' '; |
assert(OUST_NL == st->oust); |
|
assert(st->wantws); |
|
printf(".%s ", linemac); |
|
st->oust = OUST_MAC; |
|
|
|
gotmacro = 0; |
while (*start < end) { |
while (*start < end) { |
if (*start + 1 < end && '<' == buf[*start + 1]) { |
wantws = ' ' == buf[*start] || '\n' == buf[*start]; |
formatcode(st, buf, start, end, 1, nomacro, 1); |
if (wantws) { |
|
last = ' '; |
|
do { |
|
(*start)++; |
|
} while (*start < end && ' ' == buf[*start]); |
|
} |
|
|
|
if (*start + 1 < end && '<' == buf[*start + 1] && |
|
'A' <= buf[*start] && 'Z' >= buf[*start]) { |
|
st->wantws |= wantws; |
|
gotmacro = formatcode(st, buf, |
|
start, end, nomacro, 1); |
continue; |
continue; |
} |
} |
|
|
|
if (gotmacro) { |
|
if (*start < end || st->outbuflen) { |
|
if (st->wantws || |
|
(wantws && !st->outbuflen)) |
|
printf(" No "); |
|
else |
|
printf(" Ns "); |
|
} |
|
gotmacro = 0; |
|
} |
|
outbuf_flush(st); |
|
st->wantws = wantws; |
|
|
|
if (*start >= end) |
|
break; |
|
|
|
if (st->wantws) { |
|
putchar(' '); |
|
st->wantws = 0; |
|
} |
|
|
/* |
/* |
* Since we're already on a macro line, we want to make |
* Since we're already on a macro line, we want to make |
* sure that we don't inadvertently invoke a macro. |
* sure that we don't inadvertently invoke a macro. |
Line 707 formatcodeln(struct state *st, const char *buf, |
|
Line 805 formatcodeln(struct state *st, const char *buf, |
|
* something that needn't be escaped. |
* something that needn't be escaped. |
*/ |
*/ |
if (' ' == last && end - *start > 1 && |
if (' ' == last && end - *start > 1 && |
isupper((int)buf[*start]) && |
isupper((unsigned char)buf[*start]) && |
islower((int)buf[*start + 1]) && |
islower((unsigned char)buf[*start + 1]) && |
(end - *start == 2 || |
(end - *start == 2 || ' ' == buf[*start + 2])) |
' ' == buf[*start + 2])) |
|
printf("\\&"); |
printf("\\&"); |
|
|
if ('\n' == buf[*start]) |
putchar(last = buf[*start]); |
putchar(last = ' '); |
|
else |
|
putchar(last = buf[*start]); |
|
|
|
/* Protect against character escapes. */ |
/* Protect against character escapes. */ |
|
|
if ('\\' == last) |
if ('\\' == last) |
putchar('e'); |
putchar('e'); |
|
|
Line 752 listguess(const char *buf, size_t start, size_t end) |
|
Line 847 listguess(const char *buf, size_t start, size_t end) |
|
* A command paragraph, as noted in the perlpod manual, just indicates |
* A command paragraph, as noted in the perlpod manual, just indicates |
* that we should do something, optionally with some text to print as |
* that we should do something, optionally with some text to print as |
* well. |
* well. |
|
* From the perspective of external callers, |
|
* always stays in OUST_NL/wantws mode, |
|
* but its children do use OUST_MAC. |
*/ |
*/ |
static void |
static void |
command(struct state *st, const char *buf, size_t start, size_t end) |
command(struct state *st, const char *buf, size_t start, size_t end) |
Line 795 command(struct state *st, const char *buf, size_t star |
|
Line 893 command(struct state *st, const char *buf, size_t star |
|
* The behaviour of head= follows from a quick glance at |
* The behaviour of head= follows from a quick glance at |
* how pod2man handles it. |
* how pod2man handles it. |
*/ |
*/ |
printf(".Sh "); |
|
st->sect = SECT_NONE; |
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)) |
Line 804 command(struct state *st, const char *buf, size_t star |
|
Line 901 command(struct state *st, const char *buf, size_t star |
|
if (0 == memcmp(&buf[start], "SYNOPSIS", 8)) |
if (0 == memcmp(&buf[start], "SYNOPSIS", 8)) |
st->sect = SECT_SYNOPSIS; |
st->sect = SECT_SYNOPSIS; |
} |
} |
formatcodeln(st, buf, &start, end, 1); |
formatcodeln(st, "Sh", buf, &start, end, 1); |
putchar(last = '\n'); |
mdoc_newln(st); |
st->haspar = 1; |
st->haspar = 1; |
break; |
break; |
case (CMD_HEAD2): |
case (CMD_HEAD2): |
printf(".Ss "); |
formatcodeln(st, "Ss", buf, &start, end, 1); |
formatcodeln(st, buf, &start, end, 1); |
mdoc_newln(st); |
putchar(last = '\n'); |
|
st->haspar = 1; |
st->haspar = 1; |
break; |
break; |
case (CMD_HEAD3): |
case (CMD_HEAD3): |
puts(".Pp"); |
puts(".Pp"); |
printf(".Em "); |
formatcodeln(st, "Em", buf, &start, end, 0); |
formatcodeln(st, buf, &start, end, 0); |
mdoc_newln(st); |
putchar(last = '\n'); |
|
puts(".Pp"); |
puts(".Pp"); |
st->haspar = 1; |
st->haspar = 1; |
break; |
break; |
case (CMD_HEAD4): |
case (CMD_HEAD4): |
puts(".Pp"); |
puts(".Pp"); |
printf(".No "); |
formatcodeln(st, "No", buf, &start, end, 0); |
formatcodeln(st, buf, &start, end, 0); |
mdoc_newln(st); |
putchar(last = '\n'); |
|
puts(".Pp"); |
puts(".Pp"); |
st->haspar = 1; |
st->haspar = 1; |
break; |
break; |
Line 878 command(struct state *st, const char *buf, size_t star |
|
Line 972 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 "); |
formatcodeln(st, "It", buf, &start, end, 0); |
formatcodeln(st, buf, &start, end, 0); |
mdoc_newln(st); |
putchar(last = '\n'); |
|
break; |
break; |
case (LIST_ENUM): |
case (LIST_ENUM): |
/* FALLTHROUGH */ |
/* FALLTHROUGH */ |
Line 931 command(struct state *st, const char *buf, size_t star |
|
Line 1024 command(struct state *st, const char *buf, size_t star |
|
} |
} |
|
|
/* |
/* |
|
* Put the type provided as an argument into the dictionary. |
|
*/ |
|
static void |
|
register_type(const char *ptype) |
|
{ |
|
const char *pname, *pend; |
|
|
|
pname = ptype; |
|
while (isalnum((unsigned char)*pname) || '_' == *pname) |
|
pname++; |
|
if ((pname - ptype == 6 && ! strncmp(ptype, "struct", 6)) || |
|
(pname - ptype == 4 && ! strncmp(ptype, "enum", 4))) { |
|
while (' ' == *pname) |
|
pname++; |
|
pend = pname; |
|
while (isalnum((unsigned char)*pend) || '_' == *pend) |
|
pend++; |
|
if (pend > pname) |
|
dict_put(pname, pend - pname, MDOC_Vt); |
|
} else |
|
pend = pname; |
|
if (pend > ptype) |
|
dict_put(ptype, pend - ptype, MDOC_Vt); |
|
} |
|
|
|
/* |
* Just pump out the line in a verbatim block. |
* Just pump out the line in a verbatim block. |
|
* From the perspective of external callers, |
|
* always stays in OUST_NL/wantws mode. |
*/ |
*/ |
static void |
static void |
verbatim(struct state *st, const char *buf, size_t start, size_t end) |
verbatim(struct state *st, char *buf, size_t start, size_t end) |
{ |
{ |
size_t i; |
size_t i, ift, ifo, ifa, ifc, inl; |
|
char *cp, *cp2; |
|
int nopen; |
|
|
if ( ! st->parsing || st->paused) |
if ( ! st->parsing || st->paused || start == end) |
return; |
return; |
again: |
again: |
/* |
/* |
|
|
*/ |
*/ |
if (SECT_SYNOPSIS == st->sect) { |
if (SECT_SYNOPSIS == st->sect) { |
i = start; |
i = start; |
for (i = start; i < end && ' ' == buf[i]; i++) |
while (i < end && buf[i] == ' ') |
/* Spin. */ ; |
i++; |
if (i == end) |
if (i == end) |
return; |
return; |
|
|
/* We're an include block! */ |
/* We're an include block! */ |
if (end - i > 10 && |
if (end - i > 10 && |
0 == memcmp(&buf[i], "#include <", 10)) { |
0 == memcmp(&buf[i], "#include <", 10)) { |
|
|
start++; |
start++; |
if (start < end && '\n' == buf[start]) |
if (start < end && '\n' == buf[start]) |
start++; |
start++; |
|
goto again; |
|
} |
|
|
|
/* Other preprocessor directives. */ |
|
if ('#' == buf[i]) { |
|
fputs(".Fd ", stdout); |
|
start = i; |
|
while(start < end && '\n' != buf[start]) |
|
putchar(buf[start++]); |
|
putchar('\n'); |
|
if (start < end && '\n' == buf[start]) |
|
start++; |
|
goto again; |
|
} |
|
|
|
/* Parse function declaration. */ |
|
ifo = ifa = ifc = 0; |
|
inl = end; |
|
nopen = 0; |
|
for (ift = i; i < end; i++) { |
|
if (ifc) { |
|
if (buf[i] != '\n') |
|
continue; |
|
inl = i; |
|
break; |
|
} |
|
switch (buf[i]) { |
|
case ' ': |
|
if ( ! ifa) |
|
ifo = i; |
|
break; |
|
case '(': |
|
if (ifo) { |
|
nopen++; |
|
if ( ! ifa) |
|
ifa = i; |
|
} else |
|
i = end; |
|
break; |
|
case ')': |
|
switch (nopen) { |
|
case 0: |
|
i = end; |
|
break; |
|
case 1: |
|
ifc = i; |
|
break; |
|
default: |
|
nopen--; |
|
break; |
|
} |
|
break; |
|
default: |
|
break; |
|
} |
|
} |
|
|
|
/* Encode function declaration. */ |
|
if (ifc) { |
|
for (i = ifa; i < ifc; i++) |
|
if (buf[i] == '\n') |
|
buf[i] = ' '; |
|
buf[ifo++] = '\0'; |
|
register_type(buf + ift); |
|
printf(".Ft %s", buf + ift); |
|
if (buf[ifo] == '*') { |
|
fputs(" *", stdout); |
|
ifo++; |
|
} |
|
putchar('\n'); |
|
buf[ifa++] = '\0'; |
|
printf(".Fo %s\n", buf + ifo); |
|
dict_put(buf + ifo, 0, MDOC_Fo); |
|
buf[ifc++] = '\0'; |
|
for (;;) { |
|
cp = strchr(buf + ifa, ','); |
|
if (cp != NULL) { |
|
cp2 = cp; |
|
*cp++ = '\0'; |
|
} else |
|
cp2 = strchr(buf + ifa, '\0'); |
|
while (isalnum((unsigned char)cp2[-1]) || |
|
'_' == cp2[-1]) |
|
cp2--; |
|
if ('\0' != *cp2) |
|
dict_put(cp2, 0, MDOC_Fa); |
|
register_type(buf + ifa); |
|
printf(".Fa \"%s\"\n", buf + ifa); |
|
if (cp == NULL) |
|
break; |
|
while (*cp == ' ') |
|
cp++; |
|
ifa = cp - buf; |
|
} |
|
puts(".Fc"); |
|
if (buf[ifc] == ';') |
|
ifc++; |
|
if (ifc < inl) { |
|
buf[inl] = '\0'; |
|
puts(buf + ifc); |
|
} |
|
start = inl + 1; |
if (start < end) |
if (start < end) |
goto again; |
goto again; |
return; |
return; |
} |
} |
} |
} |
|
|
if (start == end) |
|
return; |
|
puts(".Bd -literal"); |
puts(".Bd -literal"); |
for (last = ' '; start < end; start++) { |
for (last = ' '; start < end; start++) { |
/* |
/* |
Line 1020 hasmatch(const char *buf, size_t start, size_t end) |
|
Line 1244 hasmatch(const char *buf, size_t start, size_t end) |
|
* If we're an ending bracket, see if we have a stack already. |
* If we're an ending bracket, see if we have a stack already. |
*/ |
*/ |
static int |
static int |
dosynopsisop(const char *buf, size_t *start, size_t end, size_t *opstack) |
dosynopsisop(struct state *st, const char *buf, |
|
size_t *start, size_t end, size_t *opstack) |
{ |
{ |
|
|
assert('[' == buf[*start] || ']' == buf[*start]); |
assert('[' == buf[*start] || ']' == buf[*start]); |
|
|
if ('[' == buf[*start] && hasmatch(buf, *start + 1, end)) { |
if ('[' == buf[*start] && hasmatch(buf, *start + 1, end)) { |
if ('\n' != last) |
mdoc_newln(st); |
putchar('\n'); |
|
puts(".Oo"); |
puts(".Oo"); |
(*opstack)++; |
(*opstack)++; |
} else if ('[' == buf[*start]) |
} else if ('[' == buf[*start]) |
return(0); |
return(0); |
|
|
if (']' == buf[*start] && *opstack > 0) { |
if (']' == buf[*start] && *opstack > 0) { |
if ('\n' != last) |
mdoc_newln(st); |
putchar('\n'); |
|
puts(".Oc"); |
puts(".Oc"); |
(*opstack)--; |
(*opstack)--; |
} else if (']' == buf[*start]) |
} else if (']' == buf[*start]) |
Line 1050 dosynopsisop(const char *buf, size_t *start, size_t en |
|
Line 1273 dosynopsisop(const char *buf, size_t *start, size_t en |
|
|
|
/* |
/* |
* Format multiple "Nm" manpage names in the NAME section. |
* Format multiple "Nm" manpage names in the NAME section. |
|
* From the perspective of external callers, |
|
* always stays in OUST_NL/wantws mode, |
|
* but its children do use OUST_MAC. |
*/ |
*/ |
static void |
static void |
donamenm(struct state *st, const char *buf, size_t *start, size_t end) |
donamenm(struct state *st, const char *buf, size_t *start, size_t end) |
{ |
{ |
size_t word; |
size_t word; |
|
|
|
assert(OUST_NL == st->oust); |
|
assert(st->wantws); |
|
|
while (*start < end && ' ' == buf[*start]) |
while (*start < end && ' ' == buf[*start]) |
(*start)++; |
(*start)++; |
|
|
Line 1065 donamenm(struct state *st, const char *buf, size_t *st |
|
Line 1294 donamenm(struct state *st, const char *buf, size_t *st |
|
} |
} |
|
|
while (*start < end) { |
while (*start < end) { |
fputs(".Nm ", stdout); |
|
for (word = *start; word < end; word++) |
for (word = *start; word < end; word++) |
if (',' == buf[word]) |
if (',' == buf[word]) |
break; |
break; |
formatcodeln(st, buf, start, word, 1); |
formatcodeln(st, "Nm", buf, start, word, 1); |
if (*start == end) { |
if (*start == end) { |
putchar(last = '\n'); |
mdoc_newln(st); |
continue; |
break; |
} |
} |
assert(',' == buf[*start]); |
assert(',' == buf[*start]); |
puts(" ,"); |
printf(" ,"); |
|
mdoc_newln(st); |
(*start)++; |
(*start)++; |
while (*start < end && ' ' == buf[*start]) |
while (*start < end && ' ' == buf[*start]) |
(*start)++; |
(*start)++; |
Line 1089 donamenm(struct state *st, const char *buf, size_t *st |
|
Line 1318 donamenm(struct state *st, const char *buf, size_t *st |
|
* Lots of other snakes in the grass: escaping a newline followed by a |
* Lots of other snakes in the grass: escaping a newline followed by a |
* period (accidental mdoc(7) control), double-newlines after macro |
* period (accidental mdoc(7) control), double-newlines after macro |
* passages, etc. |
* passages, etc. |
|
* |
|
* Uses formatcode() to go to OUST_MAC mode |
|
* and outbuf_flush() to go to OUST_TXT mode. |
|
* In text mode, wantws requests white space before the text |
|
* currently contained in the outbuf, not before upcoming text. |
|
* Must make sure to go back to OUST_NL/wantws mode before returning. |
*/ |
*/ |
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) |
{ |
{ |
size_t i, j, opstack; |
size_t i, j, opstack, wend; |
int seq; |
enum mdoc_type mtype; |
|
int eos, noeos, seq; |
|
|
if ( ! st->parsing || st->paused) |
if ( ! st->parsing || st->paused) |
return; |
return; |
Line 1119 ordinary(struct state *st, const char *buf, size_t sta |
|
Line 1355 ordinary(struct state *st, const char *buf, size_t sta |
|
start = j + 1; |
start = j + 1; |
while (start < end && ' ' == buf[start]) |
while (start < end && ' ' == buf[start]) |
start++; |
start++; |
fputs(".Nd ", stdout); |
formatcodeln(st, "Nd", buf, &start, end, 1); |
formatcodeln(st, buf, &start, end, 1); |
mdoc_newln(st); |
putchar(last = '\n'); |
|
return; |
return; |
} |
} |
} |
} |
Line 1130 ordinary(struct state *st, const char *buf, size_t sta |
|
Line 1365 ordinary(struct state *st, const char *buf, size_t sta |
|
puts(".Pp"); |
puts(".Pp"); |
|
|
st->haspar = 0; |
st->haspar = 0; |
st->hasnl = 1; |
|
last = '\n'; |
last = '\n'; |
opstack = 0; |
opstack = 0; |
|
|
Line 1140 ordinary(struct state *st, const char *buf, size_t sta |
|
Line 1374 ordinary(struct state *st, const char *buf, size_t sta |
|
* Escape initial control characters. |
* Escape initial control characters. |
*/ |
*/ |
while (start < end) { |
while (start < end) { |
if (start < end - 1 && '<' == buf[start + 1]) |
if (start < end - 1 && '<' == buf[start + 1] && |
|
'A' <= buf[start] && 'Z' >= buf[start]) |
break; |
break; |
else if ('\n' == buf[start]) |
else if ('\n' == buf[start]) |
break; |
break; |
Line 1153 ordinary(struct state *st, const char *buf, size_t sta |
|
Line 1388 ordinary(struct state *st, const char *buf, size_t sta |
|
* brackets indicate that we're opening and |
* brackets indicate that we're opening and |
* closing an optional context. |
* closing an optional context. |
*/ |
*/ |
|
|
if (SECT_SYNOPSIS == st->sect && |
if (SECT_SYNOPSIS == st->sect && |
('[' == buf[start] || |
('[' == buf[start] || |
']' == buf[start]) && |
']' == buf[start]) && |
dosynopsisop(buf, &start, end, &opstack)) |
dosynopsisop(st, buf, |
|
&start, end, &opstack)) |
continue; |
continue; |
|
|
|
/* Merely buffer non-whitespace. */ |
|
|
last = buf[start++]; |
last = buf[start++]; |
if (' ' == last) { |
if ( ! isspace(last)) |
outbuf_flush(st); |
|
putchar(' '); |
|
} else |
|
outbuf_addchar(st); |
outbuf_addchar(st); |
|
if (start < end && |
|
! isspace((unsigned char)buf[start])) |
|
continue; |
|
|
|
/* |
|
* Found the end of a word. |
|
* Rewind trailing delimiters. |
|
*/ |
|
|
|
eos = noeos = 0; |
|
for (wend = st->outbuflen; wend; wend--) |
|
if ('.' == st->outbuf[wend - 1] || |
|
'!' == st->outbuf[wend - 1] || |
|
'?' == st->outbuf[wend - 1]) |
|
eos = 1; |
|
else if ('|' == st->outbuf[wend - 1] || |
|
',' == st->outbuf[wend - 1] || |
|
';' == st->outbuf[wend - 1] || |
|
':' == st->outbuf[wend - 1]) |
|
noeos = 1; |
|
else if ('\'' != st->outbuf[wend - 1] && |
|
'"' != st->outbuf[wend - 1] && |
|
')' != st->outbuf[wend - 1] && |
|
']' != st->outbuf[wend - 1]) |
|
break; |
|
eos &= ! noeos; |
|
|
|
/* |
|
* Detect function names. |
|
*/ |
|
|
|
mtype = MDOC_Fa; |
|
if (wend && ')' == st->outbuf[wend] && |
|
'(' == st->outbuf[wend - 1]) { |
|
mtype = dict_get(st->outbuf, --wend); |
|
if (MDOC_Fo == mtype || MDOC_MAX == mtype) { |
|
st->outbuflen = wend; |
|
st->outbuf[wend] = '\0'; |
|
mdoc_newln(st); |
|
if (MDOC_Fo == mtype) |
|
fputs(".Fn ", stdout); |
|
else |
|
fputs(".Xr ", stdout); |
|
st->oust = OUST_MAC; |
|
} |
|
} |
|
|
|
/* |
|
* On whitespace, flush the output buffer |
|
* and allow breaking to a macro line. |
|
*/ |
|
|
|
outbuf_flush(st); |
|
|
|
/* |
|
* End macro lines, and |
|
* end text lines at the end of sentences. |
|
*/ |
|
|
|
if (OUST_MAC == st->oust || (eos && wend > 1 && |
|
islower((unsigned char)st->outbuf[wend - 1]))) { |
|
if (MDOC_MAX == mtype) |
|
fputs(" 3", stdout); |
|
if (MDOC_Fa != mtype) |
|
for (wend += 2; |
|
'\0' != st->outbuf[wend]; |
|
wend++) |
|
printf(" %c", |
|
st->outbuf[wend]); |
|
mdoc_newln(st); |
|
} |
|
|
|
/* Advance to the next word. */ |
|
|
|
while ('\n' != buf[start] && |
|
isspace((unsigned char)buf[start])) |
|
start++; |
|
st->wantws = 1; |
} |
} |
|
|
if (start < end - 1 && '<' == buf[start + 1]) { |
if (start < end - 1 && '<' == buf[start + 1] && |
if (formatcode(st, buf, &start, end, 0, 0, seq)) { |
'A' <= buf[start] && 'Z' >= buf[start]) { |
|
formatcode(st, buf, &start, end, 0, seq); |
|
if (OUST_MAC == st->oust) { |
/* |
/* |
* Let mdoc(7) handle trailing punctuation. |
* Let mdoc(7) handle trailing punctuation. |
* XXX Some punctuation characters |
* XXX Some punctuation characters |
Line 1182 ordinary(struct state *st, const char *buf, size_t sta |
|
Line 1499 ordinary(struct state *st, const char *buf, size_t sta |
|
putchar(' '); |
putchar(' '); |
putchar(buf[start++]); |
putchar(buf[start++]); |
} |
} |
/* End the macro line. */ |
|
putchar(last = '\n'); |
if (st->wantws || |
st->hasnl = 1; |
' ' == buf[start] || |
|
'\n' == buf[start]) |
|
mdoc_newln(st); |
|
|
/* |
/* |
* Consume all whitespace |
* Consume all whitespace |
* so we don't accidentally start |
* so we don't accidentally start |
* an implicit literal line. |
* an implicit literal line. |
*/ |
*/ |
|
|
while (start < end && ' ' == buf[start]) |
while (start < end && ' ' == buf[start]) |
start++; |
start++; |
|
|
|
/* |
|
* Some text is following. |
|
* Implement requested spacing. |
|
*/ |
|
|
|
if ( ! st->wantws && start < end && |
|
('<' != buf[start + 1] || |
|
'A' > buf[start] || |
|
'Z' < buf[start])) { |
|
printf(" Ns "); |
|
st->wantws = 1; |
|
} |
} |
} |
} else if (start < end && '\n' == buf[start]) { |
} else if (start < end && '\n' == buf[start]) { |
outbuf_newln(st); |
outbuf_flush(st); |
|
mdoc_newln(st); |
if (++start >= end) |
if (++start >= end) |
continue; |
continue; |
/* |
/* |
Line 1212 ordinary(struct state *st, const char *buf, size_t sta |
|
Line 1547 ordinary(struct state *st, const char *buf, size_t sta |
|
break; |
break; |
} |
} |
} |
} |
outbuf_newln(st); |
outbuf_flush(st); |
|
mdoc_newln(st); |
} |
} |
|
|
/* |
/* |
Line 1221 ordinary(struct state *st, const char *buf, size_t sta |
|
Line 1557 ordinary(struct state *st, const char *buf, size_t sta |
|
* (default: starts with "="). |
* (default: starts with "="). |
*/ |
*/ |
static void |
static void |
dopar(struct state *st, const char *buf, size_t start, size_t end) |
dopar(struct state *st, char *buf, size_t start, size_t end) |
{ |
{ |
|
|
|
assert(OUST_NL == st->oust); |
|
assert(st->wantws); |
|
|
if (end == start) |
if (end == start) |
return; |
return; |
if (' ' == buf[start] || '\t' == buf[start]) |
if (' ' == buf[start] || '\t' == buf[start]) |
Line 1240 dopar(struct state *st, const char *buf, size_t start, |
|
Line 1579 dopar(struct state *st, const char *buf, size_t start, |
|
*/ |
*/ |
static void |
static void |
dofile(const struct args *args, const char *fname, |
dofile(const struct args *args, const char *fname, |
const struct tm *tm, const char *buf, size_t sz) |
const struct tm *tm, char *buf, size_t sz) |
{ |
{ |
char datebuf[64]; |
char datebuf[64]; |
struct state st; |
struct state st; |
Line 1305 dofile(const struct args *args, const char *fname, |
|
Line 1644 dofile(const struct args *args, const char *fname, |
|
|
|
free(title); |
free(title); |
|
|
|
dict_init(); |
memset(&st, 0, sizeof(struct state)); |
memset(&st, 0, sizeof(struct state)); |
|
st.oust = OUST_NL; |
|
st.wantws = 1; |
|
|
assert(sz > 0); |
assert(sz > 0); |
|
|
/* Main loop over file contents. */ |
/* Main loop over file contents. */ |
Line 1329 dofile(const struct args *args, const char *fname, |
|
Line 1672 dofile(const struct args *args, const char *fname, |
|
dopar(&st, buf, cur, end); |
dopar(&st, buf, cur, end); |
cur = sup; |
cur = sup; |
} |
} |
|
dict_destroy(); |
} |
} |
|
|
/* |
/* |