Added try/catch into unescapeuri()

[GregorMatheis]

In the current version unescapeuri() gives an error if the %... character is an invalid base 16. However some url parameters can have %... as part of their value.
For those cases the function now gives a warning but returns the string without changes and doesn't throw an error.

[c42f]

Hi, thanks for the PR!

Can you include one or several test cases to show how this is supposed to work?

It's very useful to tie all URI parsing and escaping back to one of the RFC standards, or to a library with similar functionality so that we know we're doing something sensible. Do you have a reference for the desired behavior? https://tools.ietf.org/html/rfc3986 is the primary RFC for URI syntax. It can be a little hard to interpret though, so referring to how other software libraries implement these things can also be useful.

I don't think these low level parsers should throw warnings - that should be up to higher level code in the user's application. From the point of view of URIs.jl I think we should try to be definitive about whether URIs are malformed or not.

• If malformed, we should throw an error by default, and perhaps have a "permissive" flag to allow more permissive parsing of malformed URIs.
• If not malformed, then this is a bug in our code which we should just fix!

[GregorMatheis]

Hi @c42f

first, I'm super sorry for the late response, somehow I missed the notification that you answered.
My idea originated in a real life issue with data coming from Google Analytics that includes the full URL. For some reasons there are '%' in the URI that aren't valid base 16 character (and additionally last character '%'). Not sure if that is an issue that more people are facing.

Anyways I'm happy to not throw a warning and have something like a flag. I'll edit that in the code so you can have a look.