Re: Draft [URL] reference update to informative text

On 10/10/2014 05:59 AM, Anne van Kesteren wrote:
> On Fri, Oct 10, 2014 at 11:22 AM, Sam Ruby <rubys@intertwingly.net> wrote:
>> I'll note that these inputs are not consistent.
>
> If there are changes to the specification that are an improvement and
> you would be willing to make them, I don't see why those would not be
> accepted. I just don't personally think that's what we need to move
> this to the next level.

Cool.

Now let me describe my next steps.

1) I'm going to attempt to build a faithful implementation of the URL 
Standard.  By that, I mean that it will do the steps as written, with 
little regard for things like performance.  This effort may identify 
problems with the spec.  As a concrete example, I don't see how c could 
be the EOF code point at in scheme data state step 3.1[1].  I'll collect 
these up and report them on the whatwg mailing list, and once the 
discussion dies down, will open bug reports on whatever is remaining.

2) I'm then going to convert urltestdata.txt[2] to a series of unit 
tests, and run them against my code.  For each failure, I will determine 
whether that is as a result of a bug in my code or in the test data 
itself.  In both cases, I'll directly make the fix.  Once this step is 
complete, I'll post my code and submit a pull request for evaluation on 
urltestdata.txt.

3) I'll then work to identify changes to the code which would bring the 
test results in greater alignment with existing RFC's and/or browser 
implementations.  In terms of the color coded results[3], I'll work to 
eliminate the Hot Pink reduce the Pale Red.  As I do so, I'll submit a 
series of small pull requests against the URL Standard.

- Sam Ruby

[1] https://url.spec.whatwg.org/#scheme-data-state

[2] 
https://raw.githubusercontent.com/w3c/web-platform-tests/master/url/urltestdata.txt

[3] http://intertwingly.net/stories/2014/10/05/urltest-results/

Received on Friday, 10 October 2014 11:35:38 UTC