OIIO_NAMESPACE_BEGIN
TypeDesc::TypeDesc (string_view typestring)
: basetype(UNKNOWN), aggregate(SCALAR), vecsemantics(NOXFORM),
reserved(0), arraylen(0)
{
fromstring (typestring);
}
//strtof exists in C99, but let's not use that
bool fromstring(cstring s, float& out)
{
out = 0;
double tmp;
if (!fromstring(s, tmp)) return false;
if (tmp < -FLT_MAX || tmp > FLT_MAX) return false;
out = tmp;
return true;
}
void HTML::entity_decode(string& out, cstring& in, bool isattr)
{
//this follows the HTML5 spec <https://html.spec.whatwg.org/#character-reference-state>
//12.2.5.72 Character reference state
if (isalnum(in[1]))
{
//12.2.5.73 Named character reference state
const char * ent = find_entity(in.substr(1, ~0));
if (!ent) goto fail;
size_t entlen = strlen(ent);
size_t entlen_cons = 1+entlen;
if (ent[entlen-1]!=';')
{
if (in[entlen_cons]==';') entlen_cons++;
else
{
if (isattr && // If the character reference was consumed as part of an attribute,
/* checked above */ // and the last character matched is not a U+003B SEMICOLON character (;),
(in[entlen_cons]=='=' || // and the next input character is either a U+003D EQUALS SIGN character (=)
isalnum(in[entlen_cons]))) // or an ASCII alphanumeric,
{ // then, for historical reasons,
goto fail; // flush code points consumed as a character reference and switch to the return state
}
}
}
in = in.substr(entlen_cons, ~0);
const char * entval = (ent+entlen+1);
if (*entval == '?') entval++;
out += entval;
return;
}
else if (in[1]=='#')
{
//12.2.5.75 Numeric character reference state
size_t ccode;
if (in[2]=='x' || in[2]=='X')
{
//12.2.5.76 Hexademical character reference start state
size_t n = 3;
while (isxdigit(in[n])) n++;
if (n==3) goto fail;
if (!fromstringhex(in.substr(3, n), ccode)) ccode = 0xFFFD;
if (in[n]==';') n++;
in = in.substr(n, ~0);
}
else
{
//12.2.5.77 Decimal character reference start state
size_t n = 2;
while (isdigit(in[n])) n++;
if (n==2) goto fail;
if (!fromstring(in.substr(2, n), ccode)) ccode = 0xFFFD;
if (in[n]==';') n++;
in = in.substr(n, ~0);
}
//12.2.5.80 Numeric character reference end state
if (ccode == 0x00) ccode = 0xFFFD;
if (ccode > 0x10FFFF) ccode = 0xFFFD;
if (ccode >= 0xD800 && ccode <= 0xDFFF) ccode = 0xFFFD;
if (ccode >= 0x80 && ccode <= 0x9F)
{
#define X 0xFFFD
static const uint16_t windows1252[32]={
//00 to 7F map to themselves
0x20AC, X, 0x201A, 0x0192, 0x201E, 0x2026, 0x2020, 0x2021,
0x02C6, 0x2030, 0x0160, 0x2039, 0x0152, X, 0x017D, X,
X, 0x2018, 0x2019, 0x201C, 0x201D, 0x2022, 0x2013, 0x2014,
0x02DC, 0x2122, 0x0161, 0x203A, 0x0153, X, 0x017E, 0x0178,
//A0 to FF map to themselves
};
#undef X
uint32_t newcp = windows1252[ccode-0x80]; // yes, this is in the spec
if (newcp != 0xFFFD) ccode = newcp;
}
out += string::codepoint(ccode);
return;
}
//else fall through
fail:
out += '&';
in = in.substr(1, ~0);
}
template<typename T> void testundec(const char * S, T V)
{
T a;
assert_eq(fromstring(S, a), true);
assert_eq(a, V);
}
assert_eq(bar[1], 0x34);
assert_eq(bar[2], 0x56);
assert_eq(bar[3], 0x78);
assert_eq(bar[4], 0x90);
assert(!fromstringhex("123456", arrayvieww<byte>(foo))); // not 4 bytes
assert(!fromstringhex("1234567", bar)); // odd length
assert(!fromstringhex("0x123456", bar)); // no 0x allowed
uint8_t u8;
int32_t i32;
uint32_t u32;
uint64_t u64;
float f;
double d;
assert(!fromstring("", u32)); // this isn't an integer
assert(!fromstringhex("", u32));
assert(!fromstring("", f));
assert(!fromstring("2,5", f)); // this is not the decimal separator, has never been and will never be
string s;
s += '7';
s += '\0';
assert(!fromstring(s, u32)); // no nul allowed
assert(!fromstring(s, f));
assert(!fromstringhex(s, u32));
assert(!fromstring("-0", u32)); // if -1 is invalid, -0 should be too
assert(!fromstringhex("-0", u32));
assert( fromstring("-0", i32)); // but if -1 is valid, -0 should be too
uuid_t(const std::string & source) {
fromstring(source.c_str());
}
uuid_t(const char * source) {
fromstring(source);
}
TypeDesc::TypeDesc (const char *typestring)
: basetype(UNKNOWN), aggregate(SCALAR), vecsemantics(NOXFORM),
reserved(0), arraylen(0)
{
fromstring (typestring);
}
size_t
TypeDesc::fromstring (const char *typestring)
{
return fromstring (string_view(typestring));
}
