Parser does not catch variable syntax error

[prsnca]

The original liquid ruby gem fails on parsing if this case is given:
{{ user_name }
and generates this error:
Liquid syntax error: Variable '{{ user_name }' was not properly terminated with regexp: /\}\}/
I would expect this parser to catch these syntax errors as well and return a ParseError.
When trying to parse something with this syntax error, the tokenizer returns 'nested endless hash' or i don't know how to name it :)

I have looked through the config and haven't found anything that will help me :/
@harttle - I'm willing to contribute on this - just give me some guidance ;)

Thanks!

[harttle]

Currently the tokenizer uses a regex-based loop to find all OutputTokens and TagTokens, and characters between them are parsed as HTMLTokens. So there's no exception thrown out for unmatched outputs and tags.

I agree that the parse loop should be implemented manually, iterate character by character for example. We'll have fine-grained control over that process, so we can throw the exceptions you expected.

You can find the source code for the tokenizer here: https://github.com/harttle/liquidjs/blob/master/src/tokenizer.js And test cases here: https://github.com/harttle/liquidjs/blob/master/test/tokenizer.js

[harttle]

Should be resolved along with #89 after re-implement the tokenizer.

[harttle]

Now supported since @7.1.0 by this commit. Closing.