version 1.5, 2014/03/23 13:00:24 |
version 1.9, 2014/03/24 01:07:30 |
Line 108 static const char fmts[FMT__MAX] = { |
|
Line 108 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 143 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 151 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 recognise L<foo::bar> as a Perl manpage, printing it in section 3p; |
|
* or a general UNIX foo(5) manpage. |
|
* If I recognise one, I set "start" to be the end of the sequence so |
|
* that the caller can safely just continue processing. |
|
* Otherwise, I don't touch "start". |
*/ |
*/ |
static int |
static int |
skipspace(const char *buf, size_t *start, size_t end) |
trylink(const char *buf, size_t *start, size_t end, size_t dsz) |
{ |
{ |
size_t sv = *start; |
size_t sv, nstart, nend, i, j; |
|
int hasdouble; |
|
|
while (*start < end && ' ' == buf[*start]) |
/* |
(*start)++; |
* Scan to the start of the terminus. |
|
* This function is more or less replicated in the formatcode() |
|
* for null or index formatting codes. |
|
*/ |
|
hasdouble = 0; |
|
for (sv = nstart = *start; nstart < end; nstart++) { |
|
/* Do we have a double-colon? */ |
|
if (':' == buf[nstart] && |
|
nstart > sv && |
|
':' == buf[nstart - 1]) |
|
hasdouble = 1; |
|
if ('>' != buf[nstart]) |
|
continue; |
|
else if (dsz == 1) |
|
break; |
|
assert(nstart > 0); |
|
if (' ' != buf[nstart - 1]) |
|
continue; |
|
i = nstart; |
|
for (j = 0; i < end && j < dsz; j++) |
|
if ('>' != buf[i++]) |
|
break; |
|
if (dsz == j) |
|
break; |
|
} |
|
|
|
/* We don't care about stubs. */ |
|
if (nstart == end || nstart == *start) |
|
return(0); |
|
|
return(*start > sv); |
/* Set nend to the end of content. */ |
|
nend = nstart; |
|
if (dsz > 1) |
|
nend--; |
|
|
|
/* |
|
* Provide for some common invocations of the link primitive. |
|
* First, allow us to link to other Perl manuals. |
|
*/ |
|
if (hasdouble) |
|
printf("Xr %.*s 3p", |
|
(int)(nend - sv), &buf[sv]); |
|
else if (nend - sv > 3 && isalnum(buf[sv]) && |
|
')' == buf[nend - 1] && |
|
isdigit((int)buf[nend - 2]) && |
|
'(' == buf[nend - 3]) |
|
printf("Xr %.*s %c", |
|
(int)(nend - 3 - sv), |
|
&buf[sv], buf[nend - 2]); |
|
else |
|
return(0); |
|
|
|
*start = nstart; |
|
return(1); |
} |
} |
|
|
/* |
/* |
Line 172 skipspace(const char *buf, size_t *start, size_t end) |
|
Line 235 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. |
* 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(const char *buf, size_t *start, |
size_t end, int reentrant, int last, int nomacro) |
size_t end, int reentrant, int nomacro) |
{ |
{ |
enum fmt fmt; |
enum fmt fmt; |
size_t i, j, dsz; |
size_t i, j, dsz; |
Line 186 formatcode(const char *buf, size_t *start, |
|
Line 251 formatcode(const char *buf, size_t *start, |
|
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++) |
|
if (buf[*start] == fmts[fmt]) |
|
break; |
|
|
|
if (FMT__MAX == fmt) { |
|
putchar(last = buf[(*start)++]); |
|
if ('\\' == last) |
|
putchar('e'); |
|
return(0); |
|
} |
|
|
/* |
/* |
* Determine whether we're overriding our delimiter. |
* Determine whether we're overriding our delimiter. |
* According to POD, if we have more than one '<' followed by a |
* According to POD, if we have more than one '<' followed by a |
Line 201 formatcode(const char *buf, size_t *start, |
|
Line 281 formatcode(const char *buf, size_t *start, |
|
if (dsz > 1 && (i >= end || ' ' != buf[i])) |
if (dsz > 1 && (i >= end || ' ' != buf[i])) |
dsz = 1; |
dsz = 1; |
|
|
for (fmt = 0; fmt < FMT__MAX; fmt++) |
|
if (buf[*start] == fmts[fmt]) |
|
break; |
|
|
|
/* Invalid macros are just regular text. */ |
|
|
|
if (FMT__MAX == fmt) { |
|
putchar(buf[*start]); |
|
(*start)++; |
|
return(0); |
|
} |
|
|
|
/* Remember, if dsz>1, to jump the trailing space. */ |
/* Remember, if dsz>1, to jump the trailing space. */ |
*start += dsz + 1 + (dsz > 1 ? 1 : 0); |
*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. |
* Just consume til the end delimiter, accounting for |
* Be wary of encountering custom delimiters (dsz>1), |
* whether it's a custom one. |
* which require special handling. |
|
*/ |
*/ |
for ( ; *start < end; (*start)++) { |
for ( ; *start < end; (*start)++) { |
if ('>' != buf[*start]) |
if ('>' != buf[*start]) |
Line 249 formatcode(const char *buf, size_t *start, |
|
Line 317 formatcode(const char *buf, size_t *start, |
|
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 257 formatcode(const char *buf, size_t *start, |
|
Line 329 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 "); |
Line 281 formatcode(const char *buf, size_t *start, |
|
Line 355 formatcode(const char *buf, size_t *start, |
|
printf("Qo Li "); |
printf("Qo Li "); |
break; |
break; |
case (FMT_LINK): |
case (FMT_LINK): |
printf("Lk "); |
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 296 formatcode(const char *buf, size_t *start, |
|
Line 370 formatcode(const char *buf, size_t *start, |
|
} |
} |
|
|
/* |
/* |
* Read until we reach the end market (e.g., '>') or until we |
* Process until we reach the end marker (e.g., '>') or until we |
* find a 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. |
Line 323 formatcode(const char *buf, size_t *start, |
|
Line 397 formatcode(const char *buf, size_t *start, |
|
} |
} |
} |
} |
if (*start + 1 < end && '<' == buf[*start + 1]) { |
if (*start + 1 < end && '<' == buf[*start + 1]) { |
formatcode(buf, start, end, 1, last, nomacro); |
formatcode(buf, start, end, 1, nomacro); |
continue; |
continue; |
} |
} |
|
|
Line 343 formatcode(const char *buf, size_t *start, |
|
Line 417 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); |
|
|
|
/* FIXME: with the "Qc", this doens't work good. */ |
|
|
|
/* |
/* |
* 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] && '\n' != buf[*start]) |
|
printf(" Ns "); |
|
|
if (',' != buf[*start] && '.' != buf[*start] && |
|
'!' != buf[*start] && '?' != buf[*start] && |
|
')' != 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 389 formatcode(const char *buf, size_t *start, |
|
Line 454 formatcode(const char *buf, size_t *start, |
|
static void |
static void |
formatcodeln(const char *buf, size_t *start, size_t end, int nomacro) |
formatcodeln(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(buf, start, end, 1, nomacro); |
continue; |
continue; |
} |
} |
/* |
/* |
Line 411 formatcodeln(const char *buf, size_t *start, size_t en |
|
Line 475 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 470 command(struct state *st, const char *buf, size_t star |
|
Line 539 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 534 command(struct state *st, const char *buf, size_t star |
|
Line 605 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 615 command(struct state *st, const char *buf, size_t star |
|
Line 694 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"); |
} |
} |
|
|
Line 635 verbatim(struct state *st, const char *buf, size_t sta |
|
Line 727 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; |
size_t i, j; |
|
|
if ( ! st->parsing || st->paused) |
if ( ! st->parsing || st->paused) |
Line 689 ordinary(struct state *st, const char *buf, size_t sta |
|
Line 780 ordinary(struct state *st, const char *buf, size_t sta |
|
else if ('\n' == last && '\'' == buf[start]) |
else if ('\n' == last && '\'' == buf[start]) |
printf("\\&"); |
printf("\\&"); |
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 698 ordinary(struct state *st, const char *buf, size_t sta |
|
Line 791 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 (formatcode(buf, &start, end, 0, last, 0)) |
if (formatcode(buf, &start, end, 0, 0)) { |
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 735 ordinary(struct state *st, const char *buf, size_t sta |
|
Line 833 ordinary(struct state *st, const char *buf, size_t sta |
|
else if ('\n' == last && '\'' == buf[start]) |
else if ('\n' == last && '\'' == buf[start]) |
printf("\\&"); |
printf("\\&"); |
putchar(last = buf[start++]); |
putchar(last = buf[start++]); |
|
if ('\\' == last) |
|
putchar('e'); |
} |
} |
} |
} |
|
|