Re: Soliciting feedback on draft-abarth-url

On 18.04.2011 18:08, Adam Barth wrote:
> ...
>> I think it would be extremely useful to have a comparison how those tests
>> with absolute URIs would be parsed by the regexp in the URI spec:
>>
>> <http://greenbytes.de/tech/webdav/rfc3986.html#rfc.section.B>
>>
>> Are there differences? And if there are, do UAs agree on what to do with
>> those inputs?
>
> Is there an easy way to execute that regular expression?  That
> document doesn't provide  a reference defining the semantics of the
> regular expression.  It alludes to POSIX regular expressions.  Is
> there some way to write a POSIX program that executes it?  In any
> case, I'd be happy to consider it if someone can explain to me how to
> run it over the test suite mentioned above.

How about just using JS regexps, and see where we can get from there?

>>> Test cases are especially helpful because they allow us to compare the
>>> behavior of different user agents and will ensure that the net result
>>> of this process is interoperable behavior.
>>
>> <brokenrecord>There are more components than UAs processing
>> URIs</brokenrecord>. So, for example, we should test with libraries
>> (jave.net.URI, the .NET equivalent etc...) as well.
>
> Two points:
>
> 1) We're not talking about URIs.

Then let's be clear what we're talking about (note java.net.URI indeed 
parses IRIs).

> 2) I'd be happy to consider other implementations if someone would be
> willing to send me the results of running the test suite over those
> implementations.
>
> Ideally, such a contributor would collate the results by test, as in
> https://github.com/abarth/url-spec/blob/master/tests/gurl-results/by-browser.txt,
> which makes it easy to compare different behavior and understand how
> implementations differ today.

I'll put that on my TODO list :-)

Best regards, Julian

Received on Monday, 18 April 2011 16:19:30 UTC