static fz_error
readnewxrefsection(pdf_xref *xref, fz_stream *stm, int i0, int i1, int w0, int w1, int w2)
{
fz_error error;
int i, n;
if (i0 < 0 || i0 + i1 > xref->len)
return fz_throw("xref stream has too many entries");
for (i = i0; i < i0 + i1; i++)
{
int a = 0;
int b = 0;
int c = 0;
if (fz_peekbyte(stm) == EOF)
{
error = fz_readerror(stm);
if (error)
return fz_rethrow(error, "truncated xref stream");
return fz_throw("truncated xref stream");
}
for (n = 0; n < w0; n++)
a = (a << 8) + fz_readbyte(stm);
for (n = 0; n < w1; n++)
b = (b << 8) + fz_readbyte(stm);
for (n = 0; n < w2; n++)
c = (c << 8) + fz_readbyte(stm);
error = fz_readerror(stm);
if (error)
return fz_rethrow(error, "truncated xref stream");
if (!xref->table[i].type)
{
int t = w0 ? a : 1;
xref->table[i].type = t == 0 ? 'f' : t == 1 ? 'n' : t == 2 ? 'o' : 0;
xref->table[i].ofs = w1 ? b : 0;
xref->table[i].gen = w2 ? c : 0;
}
}
return fz_okay;
}
static fz_error
readxref(fz_obj **trailerp, pdf_xref *xref, int ofs, char *buf, int cap)
{
fz_error error;
int c;
error = fz_seek(xref->file, ofs, 0);
if (error)
return fz_rethrow(error, "cannot seek to xref");
while (iswhite(fz_peekbyte(xref->file)))
fz_readbyte(xref->file);
c = fz_peekbyte(xref->file);
error = fz_readerror(xref->file);
if (error)
return fz_rethrow(error, "cannot read trailer");
if (c == 'x')
{
error = readoldxref(trailerp, xref, buf, cap);
if (error)
return fz_rethrow(error, "cannot read xref (ofs=%d)", ofs);
}
else if (c >= '0' && c <= '9')
{
error = readnewxref(trailerp, xref, buf, cap);
if (error)
return fz_rethrow(error, "cannot read xref (ofs=%d)", ofs);
}
else
{
return fz_throw("cannot recognize xref format");
}
return fz_okay;
}
static fz_error
readtrailer(pdf_xref *xref, char *buf, int cap)
{
fz_error error;
int c;
error = fz_seek(xref->file, xref->startxref, 0);
if (error)
return fz_rethrow(error, "cannot seek to startxref");
while (iswhite(fz_peekbyte(xref->file)))
fz_readbyte(xref->file);
c = fz_peekbyte(xref->file);
error = fz_readerror(xref->file);
if (error)
return fz_rethrow(error, "cannot read trailer");
if (c == 'x')
{
error = readoldtrailer(xref, buf, cap);
if (error)
return fz_rethrow(error, "cannot read trailer");
}
else if (c >= '0' && c <= '9')
{
error = readnewtrailer(xref, buf, cap);
if (error)
return fz_rethrow(error, "cannot read trailer");
}
else
{
return fz_throw("cannot recognize xref format: '%c'", c);
}
return fz_okay;
}
fz_error
fz_seek(fz_stream *stm, int offset, int whence)
{
fz_error error;
fz_buffer *buf = stm->buffer;
int t, c;
if (stm->dead)
return fz_throw("assert: seek in dead stream");
if (whence == 1)
{
int cur = fz_tell(stm);
if (cur < 0)
return fz_throw("cannot tell current position");
offset = cur + offset;
whence = 0;
}
buf->eof = 0;
switch (stm->kind)
{
case FZ_SFILE:
t = lseek(stm->file, offset, whence);
if (t < 0)
{
stm->dead = 1;
return fz_throw("syserr: lseek: %s", strerror(errno));
}
buf->rp = buf->bp;
buf->wp = buf->bp;
return fz_okay;
case FZ_SFILTER:
if (whence == 0)
{
if (offset < fz_tell(stm))
{
stm->dead = 1;
return fz_throw("assert: seek backwards in filter");
}
while (fz_tell(stm) < offset)
{
c = fz_readbyte(stm);
if (c == EOF)
{
error = fz_readerror(stm);
if (error)
return fz_rethrow(error, "cannot seek forward in filter");
break;
}
}
return fz_okay;
}
stm->dead = 1;
return fz_throw("assert: relative seek in filter");
case FZ_SBUFFER:
if (whence == 0)
buf->rp = CLAMP(buf->bp + offset, buf->bp, buf->ep);
else
buf->rp = CLAMP(buf->ep + offset, buf->bp, buf->ep);
return fz_okay;
default:
return fz_throw("unknown stream type");
}
}
fz_error
pdf_parseindobj(fz_obj **op, pdf_xref *xref,
fz_stream *file, char *buf, int cap,
int *onum, int *ogen, int *ostmofs)
{
fz_error error = fz_okay;
fz_obj *obj = nil;
int num = 0, gen = 0, stmofs;
pdf_token_e tok;
int len;
int a, b;
error = pdf_lex(&tok, file, buf, cap, &len);
if (error)
return fz_rethrow(error, "cannot parse indirect object (%d %d R)", num, gen);
if (tok != PDF_TINT)
return fz_throw("cannot parse indirect object (%d %d R)", num, gen);
num = atoi(buf);
error = pdf_lex(&tok, file, buf, cap, &len);
if (error)
return fz_rethrow(error, "cannot parse indirect object (%d %d R)", num, gen);
if (tok != PDF_TINT)
return fz_throw("cannot parse indirect object (%d %d R)", num, gen);
gen = atoi(buf);
error = pdf_lex(&tok, file, buf, cap, &len);
if (error)
return fz_rethrow(error, "cannot parse indirect object (%d %d R)", num, gen);
if (tok != PDF_TOBJ)
return fz_throw("cannot parse indirect object (%d %d R)", num, gen);
error = pdf_lex(&tok, file, buf, cap, &len);
if (error)
return fz_rethrow(error, "cannot parse indirect object (%d %d R)", num, gen);
switch (tok)
{
case PDF_TOARRAY:
error = pdf_parsearray(&obj, xref, file, buf, cap);
if (error)
return fz_rethrow(error, "cannot parse indirect object (%d %d R)", num, gen);
break;
case PDF_TODICT:
error = pdf_parsedict(&obj, xref, file, buf, cap);
if (error)
return fz_rethrow(error, "cannot parse indirect object (%d %d R)", num, gen);
break;
case PDF_TNAME:
obj = fz_newname(buf);
break;
case PDF_TREAL:
obj = fz_newreal(atof(buf));
break;
case PDF_TSTRING:
obj = fz_newstring(buf, len);
break;
case PDF_TTRUE:
obj = fz_newbool(1);
break;
case PDF_TFALSE:
obj = fz_newbool(0);
break;
case PDF_TNULL:
obj = fz_newnull();
break;
case PDF_TINT:
a = atoi(buf);
error = pdf_lex(&tok, file, buf, cap, &len);
if (error)
return fz_rethrow(error, "cannot parse indirect object (%d %d R)", num, gen);
if (tok == PDF_TSTREAM || tok == PDF_TENDOBJ)
{
obj = fz_newint(a);
goto skip;
}
if (tok == PDF_TINT)
{
b = atoi(buf);
error = pdf_lex(&tok, file, buf, cap, &len);
if (error)
return fz_rethrow(error, "cannot parse indirect object (%d %d R)", num, gen);
if (tok == PDF_TR)
{
obj = fz_newindirect(a, b, xref);
break;
}
}
return fz_throw("cannot parse indirect object (%d %d R)", num, gen);
case PDF_TENDOBJ:
obj = fz_newnull();
goto skip;
default:
return fz_throw("cannot parse indirect object (%d %d R)", num, gen);
}
error = pdf_lex(&tok, file, buf, cap, &len);
if (error)
{
fz_dropobj(obj);
return fz_rethrow(error, "cannot parse indirect object (%d %d R)", num, gen);
}
skip:
if (tok == PDF_TSTREAM)
{
int c = fz_readbyte(file);
if (c == '\r')
{
c = fz_peekbyte(file);
if (c != '\n')
fz_warn("line feed missing after stream begin marker (%d %d R)", num, gen);
else
c = fz_readbyte(file);
}
error = fz_readerror(file);
if (error)
{
fz_dropobj(obj);
return fz_rethrow(error, "cannot parse indirect object (%d %d R)", num, gen);
}
stmofs = fz_tell(file);
}
else if (tok == PDF_TENDOBJ)
{
stmofs = 0;
}
else
{
fz_warn("expected endobj or stream keyword (%d %d R)", num, gen);
stmofs = 0;
}
if (onum) *onum = num;
if (ogen) *ogen = gen;
if (ostmofs) *ostmofs = stmofs;
*op = obj;
return fz_okay;
}
static fz_error
readoldtrailer(pdf_xref *xref, char *buf, int cap)
{
fz_error error;
int ofs, len;
char *s;
int n;
int t;
pdf_token_e tok;
int c;
pdf_logxref("load old xref format trailer\n");
error = fz_readline(xref->file, buf, cap);
if (error)
return fz_rethrow(error, "cannot read xref marker");
if (strncmp(buf, "xref", 4) != 0)
return fz_throw("cannot find xref marker");
while (1)
{
c = fz_peekbyte(xref->file);
if (!(c >= '0' && c <= '9'))
break;
error = fz_readline(xref->file, buf, cap);
if (error)
return fz_rethrow(error, "cannot read xref count");
s = buf;
ofs = atoi(strsep(&s, " "));
if (!s)
return fz_throw("invalid range marker in xref");
len = atoi(strsep(&s, " "));
/* broken pdfs where the section is not on a separate line */
if (s && *s != '\0')
{
error = fz_seek(xref->file, -(2 + strlen(s)), 1);
if (error)
return fz_rethrow(error, "cannot seek in file");
}
t = fz_tell(xref->file);
if (t < 0)
return fz_throw("cannot tell in file");
error = fz_seek(xref->file, t + 20 * len, 0);
if (error)
return fz_rethrow(error, "cannot seek in file");
}
error = fz_readerror(xref->file);
if (error)
return fz_rethrow(error, "cannot read from file");
error = pdf_lex(&tok, xref->file, buf, cap, &n);
if (error)
return fz_rethrow(error, "cannot parse trailer");
if (tok != PDF_TTRAILER)
return fz_throw("expected trailer marker");
error = pdf_lex(&tok, xref->file, buf, cap, &n);
if (error)
return fz_rethrow(error, "cannot parse trailer");
if (tok != PDF_TODICT)
return fz_throw("expected trailer dictionary");
error = pdf_parsedict(&xref->trailer, xref, xref->file, buf, cap);
if (error)
return fz_rethrow(error, "cannot parse trailer");
return fz_okay;
}
fz_error
pdf_lex(pdf_token_e *tok, fz_stream *f, char *buf, int n, int *sl)
{
fz_error error;
int c;
while (1)
{
c = fz_peekbyte(f);
if (c == EOF)
{
*tok = PDF_TEOF;
goto cleanupokay;
}
else if (iswhite(c))
lexwhite(f);
else if (c == '%')
lexcomment(f);
else if (c == '/')
{
fz_readbyte(f);
lexname(f, buf, n);
*sl = strlen(buf);
*tok = PDF_TNAME;
goto cleanupokay;
}
else if (c == '(')
{
fz_readbyte(f);
*sl = lexstring(f, buf, n);
*tok = PDF_TSTRING;
goto cleanupokay;
}
else if (c == '<')
{
fz_readbyte(f);
c = fz_peekbyte(f);
if (c == '<')
{
fz_readbyte(f);
*tok = PDF_TODICT;
goto cleanupokay;
}
else
{
*sl = lexhexstring(f, buf, n);
*tok = PDF_TSTRING;
goto cleanupokay;
}
}
else if (c == '>')
{
fz_readbyte(f);
c = fz_readbyte(f);
if (c == '>')
{
*tok = PDF_TCDICT;
goto cleanupokay;
}
*tok = PDF_TERROR;
goto cleanuperror;
}
else if (c == '[')
{
fz_readbyte(f);
*tok = PDF_TOARRAY;
goto cleanupokay;
}
else if (c == ']')
{
fz_readbyte(f);
*tok = PDF_TCARRAY;
goto cleanupokay;
}
else if (c == '{')
{
fz_readbyte(f);
*tok = PDF_TOBRACE;
goto cleanupokay;
}
else if (c == '}')
{
fz_readbyte(f);
*tok = PDF_TCBRACE;
goto cleanupokay;
}
else if (isnumber(c))
{
lexnumber(f, buf, n);
*sl = strlen(buf);
if (strchr(buf, '.'))
{
*tok = PDF_TREAL;
goto cleanupokay;
}
*tok = PDF_TINT;
goto cleanupokay;
}
else if (isregular(c))
{
lexname(f, buf, n);
*sl = strlen(buf);
*tok = pdf_tokenfromkeyword(buf);
goto cleanupokay;
}
else
{
*tok = PDF_TERROR;
goto cleanuperror;
}
}
cleanupokay:
error = fz_readerror(f);
if (error)
{
*tok = PDF_TERROR;
return fz_rethrow(error, "cannot read token");
}
return fz_okay;
cleanuperror:
error = fz_readerror(f);
if (error)
{
*tok = PDF_TERROR;
return fz_rethrow(error, "cannot read token");
}
*tok = PDF_TERROR;
return fz_throw("lexical error");
}
