LString LString::fromUtf8(const std::string &s) {
if (s.size() == 0) return LString();
const uint8_t *begin = reinterpret_cast<const uint8_t*>(s.c_str());
const uint8_t *end = begin + s.size();
LString ret;
ret.reserve(s.size());
LChar *destBegin = ret.begin();
LChar *destEnd = ret.end();
bool conversionSucceeded = utf8ToUtf32(&begin, end, &destBegin, destEnd);
assert(conversionSucceeded && "Failed conversion utf8 -> utf32");
ret.resize(destBegin - ret.begin());
return ret;
}
bool read_constdb(ZString filename)
{
io::ReadFile in(filename);
if (!in.is_open())
{
PRINTF("can't read %s\n"_fmt, filename);
return false;
}
bool rv = true;
AString line_;
while (in.getline(line_))
{
// is_comment only works for whole-line comments
// that could change once the Z dependency is dropped ...
LString comment = "//"_s;
XString line = line_.xislice_h(std::search(line_.begin(), line_.end(), comment.begin(), comment.end())).rstrip();
if (!line)
continue;
// "%m[A-Za-z0-9_] %i %i"
// TODO promote either qsplit() or asplit()
auto _it = std::find(line.begin(), line.end(), ' ');
auto name = line.xislice_h(_it);
auto _rest = line.xislice_t(_it);
while (_rest.startswith(' '))
_rest = _rest.xslice_t(1);
auto _it2 = std::find(_rest.begin(), _rest.end(), ' ');
auto val_ = _rest.xislice_h(_it2);
auto type_ = _rest.xislice_t(_it2);
while (type_.startswith(' '))
type_ = type_.xslice_t(1);
// yes, the above actually DTRT even for underlength input
int val;
int type = 0;
// Note for future archeaologists: this code is indented correctly
if (std::find_if_not(name.begin(), name.end(),
[](char c)
{
return ('0' <= c && c <= '9')
|| ('A' <= c && c <= 'Z')
|| ('a' <= c && c <= 'z')
|| (c == '_');
}) != name.end()
|| !extract(val_, &val)
|| (!extract(type_, &type) && type_))
{
PRINTF("Bad const line: %s\n"_fmt, line_);
rv = false;
continue;
}
P<str_data_t> n = add_strp(name);
n->type = type ? StringCode::PARAM : StringCode::INT;
n->val = val;
}
return rv;
}
