Ignore querystrings

[whostolemyhat]

Fix for #207 - strips querystrings from urls when serving files locally

[whostolemyhat]

Travis is failing on cargo fmt - I ran the formatter on my code but this produced more errors, so not sure where to go from here.

[epage]

Thanks for contributing! I probably won't get a chance to review this for a day or two.

Travis is failing on cargo fmt - I ran the formatter on my code but this produced more errors, so not sure where to go from here.

I just ran into the same problem with liquid. The CI code uses a fixed version of rustfmt which is easy to get out of sync with.

https://github.com/cobalt-org/cobalt.rs/blob/master/ci/install.sh#L7

One option is to include updating the version of rustfmt as part of your change. I took the quick and dirty route in liquid and am using cargo to install rustfmt but that slows down the CI. The problem with that is that I want to be respectful of travis/appveyor providing these services for us.

[epage]

Rather than split/collect, what if we

• rfind + truncate
• rsplitn(2, "?").last().expect("always at least one element")

[whostolemyhat]

Cool, I'll have a look into this

[epage]

So what to do about the CI issue.

Personally, I'd recommend stripping off the last of your commits and rebasing against upstream master. At that point it should be resolved if you use the same version of rust fmt (0.8.3 I believe). This will get the CI passing and will keep history clean. I know "rebase + push -f" is generally considered bad form but no else should really be broken because of it and not sure what else would be recommended to do in this case.

If you don't feel up to that, we could probably commit as is (once the issue above gets cleaned up) or I could clone your branch and clean it up if you wish.

[whostolemyhat]

I've rebased and deleted the whitespace that fmt was complaining about, and the CI seems happy now.

[whostolemyhat]

I've removed the split().collect() and replaced it with rfind - if there's no match then the original req_uri is used.

[epage]

I've removed the split().collect() and replaced it with rfind - if there's no match then the original req_uri is used.

Looks good. Once the clippy issue is resolved, this is good to go in.