Adam said:
First of all thanks for all the work you've all done on this and
Thank you Adam! Well, you have successfully gotten us on to the trail of
using acid3 as the guide for what to prioritize supporting. But now that
we have discovered this whitespace thing is part of the world according to
acid3, I wonder what their quality threshhold is about. Does it go
further than we really need to go?
Here's a question for Karl, Adam and everyone. Do you think it would be
worthwhile if I prepared a little concordance to all 100 tests with what
functions it expects to be able to call?
Here's a sample. Test 7 is an example of a test with all kinds of crud
I've never heard of before. Test 17 is an example of a test that walks
in the footsteps of work we have already addressed and that is used
constantly by real websites. I propose that I do the following sample
entries for all 100. Then we would have a clue about how to triage. I
don't suppose it would be that much work as I am just going to be scanning
for unique javascript property and method names, and saying "these
keywords pass the sanity test because we know they are actually called a
lot, and these other keywords are obscure." Then we can use it to rank or
bucket the 100 tests according to which ones have obscure "features" like
the whitespace "feature" and which ones are about fundamental DOM
topology and definitely worth our while.
-------------------------------------------------------
Test 7: Basic ranges test
KC: On a likert scale of 1 to 7, 1 being tests that hew closely to basic
node operations and and 7 being tests that introduce questionable
constructs which I sure as hell have never used before, my score for Test
7 is .... 7
The unimplemented things we would need to implement to pass Test 7 are:
createRange
commentAncestorContainer
startContainer
startOffset
endContainer
endOffset
cloneContents
cloneRange
collapse
compareBoundaryPoints
deleteContents
insertNode (??! Is there such a thing? We have insertBefore. Since when
is there an insertNode?)
setEnd
If 7 is not already one of our passed tests, maybe we will end up skipping
it because all of this range stuff has never come up in any find&fix
research that I have worked on, at least.
----------------------------------------------------------
Test 17: hasAttribute
KC: On a likert scale of 1 to 7, 1 being tests that hew closely to basic
node operations and and 7 being tests that introduce questionable
constructs I sure as hell have never used before, my score for
Test 17 is..... 1
The unimplemented things we would need to implement to pass Test 7 are:
None! The only things this test expects to be able to use are
hasAttribute and getElementsByTagName. Easy, if not already a pass.
-------------------------------------------------------------
_______________________________________________
Edbrowse-dev mailing list
[email protected]
http://lists.the-brannons.com/mailman/listinfo/edbrowse-dev
