Segmentation fault with not well formed urls

[ivan-is]

Code sample for reproducing this error:

In [1]: from reppy.robots import Robots

In [2]: Robots.robots_url('http://:::cnn.com/47_50_95.html')

terminate called after throwing an instance of 'std::invalid_argument'
  what():  stoi
Aborted (core dumped)

Environment:

pip freeze | grep reppy
reppy==0.4.3

cat /etc/redhat-release 
Fedora release 24 (Twenty Four)

[Brandon Forehand]

This is actually a bug in url-cpp, which is why it is causing a segmentation fault.

[Dan Lecocq]

I guess my question is what the interpretation of this URL should be.

[Brandon Forehand]

I don't think we need to interpret it magically, but not causing a core dump would be good.

[Dan Lecocq]

That's fair. That's also something that could change in our reppy cythonization -- to catch that exception and raise something sensible.

I think url-cpp raising a new std::invalid_argument is sane, but bubbling up the stoi is lame on my part.

[Brandon Forehand]

In this case, it's not actually bubbling up the exception, terminate is being called which means an exception is being thrown during the exception stack unwinding.

[Brandon Forehand]

seomoz/url-cpp#37 addresses this issue, but it will need to be merged upstream to rep-cpp and then to reppy before this issue can be resolved.