Re: Assertion id's

Yes -- these names were also very long to go hand in hand with
the long filenames.  Sure would be nice to have some documentation
that covers these conditions -- otherwise, test writer's are just guessing
at what else needs to be done.  Out of 130 tests, we've got problems with
36 of them -- we're going to play with them until about 3:00 and then
we'll commit them for you to play with tonight.

--Mary


----- Original Message -----
From: "Arnold, Curt" <Curt.Arnold@hyprotech.com>
To: <www-dom-ts@w3.org>
Sent: Wednesday, October 24, 2001 12:49 PM
Subject: Assertion id's


> In a previous message, I suggested that the id attributes for
<assertDOMException> should be something like id="throw_INDEX_SIZE_ERR", the
motivation behind this was that exception assertions are use
> xUnit's fail() method and the id is passed as a description.  Since unlike
assertEquals, there is no additional information in a failure report, so
having an id that clearly indicates the nature of
> the failure is very helpful when reviewing the test failures.
>
> I believe that I had also suggested that assertEquals, assertNull, et al
have shorter names, typically the name of the variable being checked.  All
that is necessary of the assertion id is that it be
> unique in the document so that we can identify which assertion failed and
if it can help in the superficial interpretation of a failure, the better.
>
> The test just posted had this assertion:
>
> <assertNull actual="localName" id="throw_Null"/>
>
> "throw_Null" is not helpful, actually it could be misleading.  Since it is
an assertNull, the test output will already indicate that something that was
supposed to be null wasn't.  And nothing is
> being thrown.
>
> In this case, I would suggest an id="localName" which would result in
something approximately:
>
> Traceback: localName: expected null was ....
>
> Again, I wouldn't hold up a commit to change these, but I did want to
point this out.
>
>

Received on Wednesday, 24 October 2001 13:06:14 UTC