Triggering bug! panic with StrVals passed to sections

[ibabushkin]

Given a template such as this one:

{{# some_var }}
some text {{{ some_var }}}
{{/ some_var }}

and using

let mut data = HashMap::new();
data.insert("some_var", "some value");

as the data passed to the compiled template, I am receiving the following panic:

thread 'main' panicked at 'bug: unexpected value StrVal(some value). Please report this issue on GitHub if you find an input that triggers this case.'

A code search in the repo yields the following lines in src/template.rs as the culprit:

match self.find(path, stack) {
    None => {}
    Some(value) => {
        match *value {
            Bool(true) => {
                try!(self.render(wr, stack, children));
            }
            Bool(false) => {}
            VecVal(ref vs) => {
                for v in vs.iter() {
                    stack.push(v);
                    try!(self.render(wr, stack, children));
                    stack.pop();
                }
            }
            Map(_) => {
                stack.push(value);
                try!(self.render(wr, stack, children));
                stack.pop();
            }
            Fun(ref fcell) => {
                let f = &mut *fcell.borrow_mut();
                let tokens = try!(self.render_fun(src, otag, ctag, f));
                try!(self.render(wr, stack, &tokens));
            }
            OptVal(ref val) => {
                if let Some(ref val) = *val {
                    stack.push(val);
                    try!(self.render(wr, stack, children));
                    stack.pop();
                }
            }
            _ => bug!("unexpected value {:?}", value),
        }
    }
};

Now, the spec doesn't quite specify a clear behaviour on whether string values are to be interpreted as "thruthy", "falsey" or whether this decision is up to the host language's interpretation of this question. Now, in either case, either some filtering at template-compile-time should be done, or strings should be explicitly handled.

I can provide a fix for this (I suggest handling strings as true), but I'd like to know how the authors/maintainers see this issue.

[therealbstern]

Empty strings should be handled as "false" as they are a falsy value (see janl/mustache.js#186 for a rationale).

[therealbstern]

FYI, the code referenced above seems to be at line 278.

[ibabushkin]

Does this commit ibabushkin@1bd7ebf satisfy your requirements? If so, I'd open a pull-request.

[jolhoeft]

Yes please, we'd like a pull request.