freebsd-skq/contrib/mandoc/preconv.c
Baptiste Daroussin 7295610f5d Update mandoc to 1.14.5
MFC after:	2 weeks
2019-04-12 10:13:17 +00:00

180 lines
3.8 KiB
C

/* $Id: preconv.c,v 1.17 2018/12/13 11:55:47 schwarze Exp $ */
/*
* Copyright (c) 2011 Kristaps Dzonsons <kristaps@bsd.lv>
* Copyright (c) 2014 Ingo Schwarze <schwarze@openbsd.org>
*
* Permission to use, copy, modify, and distribute this software for any
* purpose with or without fee is hereby granted, provided that the above
* copyright notice and this permission notice appear in all copies.
*
* THE SOFTWARE IS PROVIDED "AS IS" AND THE AUTHOR DISCLAIMS ALL WARRANTIES
* WITH REGARD TO THIS SOFTWARE INCLUDING ALL IMPLIED WARRANTIES OF
* MERCHANTABILITY AND FITNESS. IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR
* ANY SPECIAL, DIRECT, INDIRECT, OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES
* WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR PROFITS, WHETHER IN AN
* ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION, ARISING OUT OF
* OR IN CONNECTION WITH THE USE OR PERFORMANCE OF THIS SOFTWARE.
*/
#include "config.h"
#include <sys/types.h>
#include <assert.h>
#include <stdio.h>
#include <string.h>
#include "mandoc.h"
#include "roff.h"
#include "mandoc_parse.h"
#include "libmandoc.h"
int
preconv_encode(const struct buf *ib, size_t *ii, struct buf *ob, size_t *oi,
int *filenc)
{
const unsigned char *cu;
int nby;
unsigned int accum;
cu = (const unsigned char *)ib->buf + *ii;
assert(*cu & 0x80);
if ( ! (*filenc & MPARSE_UTF8))
goto latin;
nby = 1;
while (nby < 5 && *cu & (1 << (7 - nby)))
nby++;
switch (nby) {
case 2:
accum = *cu & 0x1f;
if (accum < 0x02) /* Obfuscated ASCII. */
goto latin;
break;
case 3:
accum = *cu & 0x0f;
break;
case 4:
accum = *cu & 0x07;
if (accum > 0x04) /* Beyond Unicode. */
goto latin;
break;
default: /* Bad sequence header. */
goto latin;
}
cu++;
switch (nby) {
case 3:
if ((accum == 0x00 && ! (*cu & 0x20)) || /* Use 2-byte. */
(accum == 0x0d && *cu & 0x20)) /* Surrogates. */
goto latin;
break;
case 4:
if ((accum == 0x00 && ! (*cu & 0x30)) || /* Use 3-byte. */
(accum == 0x04 && *cu & 0x30)) /* Beyond Unicode. */
goto latin;
break;
default:
break;
}
while (--nby) {
if ((*cu & 0xc0) != 0x80) /* Invalid continuation. */
goto latin;
accum <<= 6;
accum += *cu & 0x3f;
cu++;
}
assert(accum > 0x7f);
assert(accum < 0x110000);
assert(accum < 0xd800 || accum > 0xdfff);
*oi += snprintf(ob->buf + *oi, 11, "\\[u%.4X]", accum);
*ii = (const char *)cu - ib->buf;
*filenc &= ~MPARSE_LATIN1;
return 1;
latin:
if ( ! (*filenc & MPARSE_LATIN1))
return 0;
*oi += snprintf(ob->buf + *oi, 11,
"\\[u%.4X]", (unsigned char)ib->buf[(*ii)++]);
*filenc &= ~MPARSE_UTF8;
return 1;
}
int
preconv_cue(const struct buf *b, size_t offset)
{
const char *ln, *eoln, *eoph;
size_t sz, phsz;
ln = b->buf + offset;
sz = b->sz - offset;
/* Look for the end-of-line. */
if (NULL == (eoln = memchr(ln, '\n', sz)))
eoln = ln + sz;
/* Check if we have the correct header/trailer. */
if ((sz = (size_t)(eoln - ln)) < 10 ||
memcmp(ln, ".\\\" -*-", 7) || memcmp(eoln - 3, "-*-", 3))
return MPARSE_UTF8 | MPARSE_LATIN1;
/* Move after the header and adjust for the trailer. */
ln += 7;
sz -= 10;
while (sz > 0) {
while (sz > 0 && ' ' == *ln) {
ln++;
sz--;
}
if (0 == sz)
break;
/* Find the end-of-phrase marker (or eoln). */
if (NULL == (eoph = memchr(ln, ';', sz)))
eoph = eoln - 3;
else
eoph++;
/* Only account for the "coding" phrase. */
if ((phsz = eoph - ln) < 7 ||
strncasecmp(ln, "coding:", 7)) {
sz -= phsz;
ln += phsz;
continue;
}
sz -= 7;
ln += 7;
while (sz > 0 && ' ' == *ln) {
ln++;
sz--;
}
if (0 == sz)
return 0;
/* Check us against known encodings. */
if (phsz > 4 && !strncasecmp(ln, "utf-8", 5))
return MPARSE_UTF8;
if (phsz > 10 && !strncasecmp(ln, "iso-latin-1", 11))
return MPARSE_LATIN1;
return 0;
}
return MPARSE_UTF8 | MPARSE_LATIN1;
}