Re: Initial Feed Validator support... and a few questions
- Sam Ruby <rubys@...> wrote:
> There, I would go further and either discourage our outright disallowYour summary of the reasons for why we allow such a thing are correct.
> elements from "shadowing" each other in this way. In other words, you
> would be free to place media:credit at any level you like, but if
> another media:credit element is encountered in an enclosed scope, a
> warning (or an error) would be produced.
Let me mull this over for a bit, but I tend to agree with what you are
saying about disallowing this. It would probably be a good idea in the
- I've completed my second pass.
The first pass was just to validate element and attribute names, which
would flag things like filesize and samplingRate.
The second pass validates values in isolation. I made plenty of
assumptions like channels must be a non-negative integer, filesize must
be a positive integer, and bitrate can be a floating point number.
Feel free to challenge any of these assumptions at any time.
When I said "in isolation" above, what I haven't implemented yet is any
checks that involve multiple elements (exception: I do enforce no
duplicates on adult, copyright, description, and title within an item).
For example, I don't yet validate that each hash element inside an item
has a different value for the algo attribute.
It may be a while before I get to those kind of checks. If any such
checks are important to you, please let me know, preferably in the form
of test cases.
The complete set of test cases implemented so far can be found here:
- Sam Ruby