W3C home > Mailing lists > Public > www-svg@w3.org > September 2010

Re: Required understanding of animate-elem-219-t test case

From: Dr. Olaf Hoffmann <Dr.O.Hoffmann@gmx.de>
Date: Fri, 24 Sep 2010 17:29:29 +0200
To: www-svg@w3.org, pavank.e87@gmail.com
Message-Id: <201009241729.30666.Dr.O.Hoffmann@gmx.de>
Hello,

the reference PNG is wrong. 
The bottom right triangle does not disappear at all - there is no such 
information within the file or in the description, but it is known, that for
example several versions of Opera have a bug, that results in destroying the
element some specific attributeType combinations are applied too.
At the time, the test was created, the Opera versions did not have
this bug, therefore it was not considered, that such a bug can happen
related to this test, therefore such a behaviour is not covered by the
simple test rule red=wrong.
Because Squiggle, typically used to create the PNGs, has bugs too
related to this test,  presumably someone created a PNG manually
relying on the wrong behaviour of Opera later ;o) 
I think, this was already discussed in the SVG-WG some time
ago on a f2f-conference, but obviously the PNG was not fixed
afterwards ;o)

Looking into the source code, I think, the test was not updated due
to a late incompatible change of the recommendation (id/xml:id issue), 
that happened after the test was created, therefore it surely needs an 
update anyway, not just the PNG.
Well, typically the PNGs are not very relevant for several animation tests
anyway, therefore it is always the problem for the test author to indicate
test relevant bugs in a simple way. However I'm not sure, if the specific bug
of Opera is something, that can be considered, before it appears.


Olaf
Received on Friday, 24 September 2010 15:31:48 GMT

This archive was generated by hypermail 2.3.1 : Friday, 8 March 2013 15:54:46 GMT