W3C home > Mailing lists > Public > public-webfonts-wg@w3.org > November 2010

Re: User Agent Test Plan Page

From: Tal Leming <tal@typesupply.com>
Date: Tue, 23 Nov 2010 14:59:17 -0500
Cc: WOFF Working Group FONT <public-webfonts-wg@w3.org>
Message-Id: <58B2C57F-B846-43DD-8FC4-0024A362CF59@typesupply.com>
To: Chris Lilley <chris@w3.org>

On Nov 23, 2010, at 1:53 PM, Chris Lilley wrote:

> TL> - valid woff files
> Yes. There needs to be a set of test cases with valid files of various sorts, so we can tell whether a UA handles woff 
> a) at all
> b) well, for a good selection of the cases where it should accept the font

I have eight so far:


> TL> - document test cases (I'm not sure what the right term is. I'm
> TL> thinking of things like the CORS, CSS3, etc. tests.)
> For the valid woff files UA tests, they will need to use @font-face anyway, no? 
> Maybe I am not seeing what characteristics this group has.

What I mean is that there are some tests that deal with where the file comes from and there are tests that deal with the file structure:

Where It Comes From:
	- http://dev.w3.org/webfonts/WOFF/spec/#conform-same-origin
	- http://dev.w3.org/webfonts/WOFF/spec/#conform-doc-origin
	- http://dev.w3.org/webfonts/WOFF/spec/#conform-css3font-available
	- http://dev.w3.org/webfonts/WOFF/spec/#conform-cors
Fire Structure:
	- http://dev.w3.org/webfonts/WOFF/spec/#conform-extraneous-reject
	- http://dev.w3.org/webfonts/WOFF/spec/#conform-overlap-reject
	- ...

I don't know the proper term for categorizing the same-orign, doc-orign, CSS3 and CORS tests. I hope this makes more sense.

> TL> - file integrity test cases
> Yes. These are mostly the 'must reject' cases, yes?


> You currently have a block of tests that don't correspond directly to testable assertions (but correspond to combinations of them, or to file format conformance statements, or to the general 'must reject invalid' and 'must accept valid' meta-statements.

Yes. This is what I was trying to get at last week on the call when I brought up the checksums. We have some testable assertions that are tagged with "file format" that could possibly be tagged with "user agent" as well. These are the ones that I have identified so far:


> Should these be collected in a group, or divided up?

I'm still trying to figure this out. Maybe I'll move them into appropriate sections and mark them as not being derived from testable assertions.

> Do some of them correspond to explicit conformance statements that should be added to the spec?

Probably. I'll add notes about this to the page.

Received on Tuesday, 23 November 2010 19:59:53 UTC

This archive was generated by hypermail 2.3.1 : Tuesday, 6 January 2015 20:34:14 UTC