RE: CSS discussion

Ok,Let's try this option.Perhaps doing the full
serialization and other pending TODOs would take us 
too much time in compararison with regEx approach.

A question about the development plan (Perhaps I am missing
something):

>After preprocessing, I'll use XPath queries in Java to pull out relevant CSS
>text. From that information, I'll run a series of regular expression tests.
>From those results, I'll construct a TestResult object and add that test
>result to TestResults so that it gets reserialized into XML for the final
>Results XML document.

So, CSS tests are going to be done without using XSLT templates?
(I thought that the initial idea was using Java from XSLT)

Regards,

Abel.



-----Mensaje original-----
De: Laura Holmes [mailto:holmes@google.com]
Enviado el: mar 07/08/2007 17:51
Para: Sean Owen
CC: Jo Rabin; Abel Rionda; public-mobileok-checker@w3.org
Asunto: Re: CSS discussion
 
Following my e-mail with a proposed moki structure, and a discussion with
Sean, we decided that for this specific test it would be best to implement
it mostly in Java without further inflating the moki.

Just so everyone is in the loop, here's my development plan:

After preprocessing, I'll use XPath queries in Java to pull out relevant CSS
text. From that information, I'll run a series of regular expression tests.
>From those results, I'll construct a TestResult object and add that test
result to TestResults so that it gets reserialized into XML for the final
Results XML document.

I'll be working on this for the next couple days and I'll let you guys know
when it's completed.

Let me know if you have any comments or concerns.

Cheers,
Laura

On 8/7/07, Sean Owen <srowen@google.com> wrote:
>
>
> I think we agree -- you are mostly concerned about validation, which I
> think is important. My first two solutions to that are:
>
> 1) Run the CSS validator on it and take its results -- this is
> probably "good enough" for 1.0
> 2) Where not good enough, use SAC to chew through the CSS and look for
> what we like.
>
> And separately I've said how I think the actual test implementations
> are trivial even without SAC.
>
> This to me makes the serialization piece of this a nice idea, but
> vestigial and optional at this moment. If this serialization were
> already complete, I would not be arguing to rip it out. But as Abel
> says, there is a daunting amount of detail left to consider, and I
> don't find it palatable to leave this in half-baked. As we still need
> to complete core tests, and are not ahead of schedule, I personally
> can't see supporting more work on this right now.
>
> I don't want to throw it away -- on the contrary, actually, I'd like
> to see this expanded into a standalone project that could be re-used
> outside this project. It is a good idea.
>
> My concrete proposal is to:
> 1) Use CSS Validator + SAC to validate
> 2) Let Laura finish and commit her CSS tests as described in this thread
> 3) Comment out the CSS serialization code
>
> On 8/7/07, Jo Rabin <jrabin@mtld.mobi> wrote:
> > I agree that we should think carefully before doing it. Fools step in
> where
> > angels fear to tread. But then again, we have both feet over the
> threshold
> > already, don't we?
>
>