Crash (pointer deref) in StrPair::GetStr()

[weblkle]

Some invalid (and unescaped) characters in the XMLText element are causing a crash when invoking GetText().
Here is a sample xml:
<root><Child Level=\"50\" Time=\"1518637009\">Valid characters, Not valid characters - €SÞ&#x0C-ung done</Child></root>

And some basic code parsing it:

int main(int argc, char* argv[])
{
    printf("Hello tinyxml2\n");

    const char* str_xml = "<root><Child Level=\"50\" Time=\"1518637009\">Valid characters, Not valid characters - €SÞ&#x0C-ung done</Child></root>";
    printf("Parsing:\n%s\n", str_xml);
    tinyxml2::XMLDocument doc;
    tinyxml2::XMLError rc = doc.Parse(str_xml);
    printf("Text parse result: %d\n", rc);
    if (rc != tinyxml2::XML_NO_ERROR)
        return rc;

    tinyxml2::XMLElement* elem = doc.FirstChildElement();
    if (!elem) return -1;

    elem = elem->FirstChildElement();
    const char* text = elem->GetText();

    printf("\nGetText()\n%s\n", text);
}

This potentially could be resolved with #255 (optional entity parsing). I know the characters are not valid and CData should be used, but it should probably not crash.

[leethomason]

As I run it, it works. I suspect the crash is that TinyXML-2 interprets everything as UTF-8, but printf/console isn't.

That said I wouldn't be at all surprised if there is a crasher in that code. I just can't reproduce it with this sample.

[weblkle]

This was an issue with 2.1, which clearly was very outdated when I wrote it up originally. Sorry. And Thanks!