W3C home > Mailing lists > Public > www-svg@w3.org > September 2008

some comments about SVG 1.2 Tiny: Beta3 of Test Suite

From: Dr. Olaf Hoffmann <Dr.O.Hoffmann@gmx.de>
Date: Fri, 19 Sep 2008 14:41:21 +0200
To: www-svg@w3.org
Message-Id: <200809191441.21813.Dr.O.Hoffmann@gmx.de>

Hello SVG WG,

meanwhile I had a look on a few samples in the beta3 version
of the test suite, here are some comments to improve it:


--------------------
- font-family="Arial" - I think, this can be removed in the following
  tests, because there is already the
  font-family="SVGFreeSansASCII,sans-serif" 
  for the SVGT1.2 tests:

animate-elem-02 
animate-elem-09
animate-elem-10
animate-elem-11
animate-elem-12
animate-elem-13
animate-elem-14
animate-elem-15
animate-elem-17
animate-elem-19
animate-elem-32

--------------------
animate-elem-09
- testcase shows only discrete animation, text talks about discrete, linear, 
paced and spline animation.


--------------------
animate-elem-210

- wrong: 'The main indication for a failed test is the appearance of red.'
- correct: 'Different coloured rings in the circles are the main indication of 
a failed test.'

--------------------
animate-elem-41

- still wrong timing indication for discrete animation (fill-rule)
  either the timing of the related animateTransform has to be changed 
  or better the animation changed to a from-to animation
 
--------------------
animate-elem-53

- calcMode paced for points not defined acording to the new draft, what is
correct/understandable, but then tests about this can either be skipped or 
modified in such a way, that it tests linear behaviour as suggested now.

--------------------
animate-elem-65
animate-elem-66

- references SMIL2.0, not SMIL2.1 as the current SVGT1.2 draft


--------------------
animate-elem-82

- describes still (wrong) behaviour of SVG 1.1 for paced animation
- paced translate animation still insensitive for implemented timing and
therefore useless
(see long discussion from last year about this test including a provided 
alternative).

--------------------
animate-elem-83

- text still mentiones path d commands A/a, neither used in the test nor 
available in SVGT1.2

--------------------
animate-elem-86

- PNG wrong, neither does it fit to the description nor the test
(the described behaviour itself fits to the current draft, clarifying the 
ambiguities in SMIL2/3 about frozen discrete animations and ensures 
backward compatibility with SVG 1.1 and the SMIL  animation 
recommendation).

--------------------

coords-constr-201

- description is very confusing.
- the reference image does not show the file after performing a user agent
zoom, it shows the initial appearance
- After a zoom the three circles will not align vertically in general, but
this is noted as pass criteria. Typically they do not align, because the first
circle/line is not constrained, the other two are.

------------------------

coords-constr-201

- the animation is applied to the parent g element. A g element has no
viewBox.
Maybe for the animation element the following is missing?
xlink:href="#svg-root" 

-------------------------
extend-namespace-02

- has still the draft sign


--------------------------
fonts-glyph-201
fonts-glyph-202
fonts-overview-201

- contains still path-data with more than 4 digits in the fractional part of a
number
- what is the correct display for such paths, because the current draft notes, 
that numbers must not have more than 4 digits?
The tests seem to assume that they are displayed anyway, but what have 
the user-agents to do?  to round, ceil, floor?
Obviously it is easy to modify these tests, to avoid this question here, but
it would be in general very useful, if the behaviour is defined and the same
in any viewer, for example: 'viewers may round numbers with more than
4 digits in the fractional part of a number'. 

--------------------------
interact-event-201
interact-event-202
struct-discard-207
text-edit-201


- no indication, that text/ecmascript is required 
  or how the test passes, if a viewer supports only another script language

----------------------------
linking-a-201
paint-stroke-01 - 04
struct-cond-01
struct-defs-01
struct-group-01
text-ws-01 / 02

text notes that font-family Arial is inside, but it isn't,
maybe the complete paragraph is redundant?

------------------------------
linking-frag-201

- PNG not related to test?

------------------------------
linking-uri-03

- wrong description, not the first but the second PNG seems to be shown


---------------------------
metadata-example-01

- description mentions DTD - this is not available in the sample and not for 
SVGT1.2

--------------------------
paint-stroke-205

- not updated

---------------

paint-vfill-204

- I think, the PNG is wrong, viewport of the referenced raster-image should be 
yellow as mentioned in the text

---------------

paint-vfill-206

- note, that the outer XHTML element has no defined background color (not even
specified with CSS), therefore the color of the viewport can be quite
different.
For example I have a default gray background, this means neither the PNG nor
the right rectangle have the correct color. At least for the right rectangle
this can be done using a system color, 'Window' maybe.

---------------
struct-cond-02

- parts of the description, an ol list, seem to miss, compare with the related 
1.1 test...
Received on Friday, 19 September 2008 12:43:51 GMT

This archive was generated by hypermail 2.3.1 : Friday, 8 March 2013 15:54:40 GMT