619630 - Support no comma or whitespace between points in polyline

Status: RESOLVED FIXED

(Core :: SVG, defect)

Description

[Jonathan Watt [:jwatt]]

If the "next" number in a polyline/polygon 'points' attribute is negative, then the grammar says that a space or comma is not required. (So points="270-225 300-245..." is valid.)

Comment 2

[Robert Longson [:longsonr]]

Attachment: patch

Comment 3

[Jonathan Watt [:jwatt]]

Attachment: testcase

Opera and Webkit also both pass this.

Comment 4

[Robert Longson [:longsonr]]

I'll leave you to decide whether to raise bugs on either Opera and Webkit or the SVG Specification.

coordinate-pairs:
    coordinate-pair
    | coordinate-pair comma-wsp coordinate-pairs
coordinate-pair:
    coordinate comma-wsp coordinate
    | coordinate negative-coordinate

Comment 5

[Robert Longson [:longsonr]]

so I don't see how comment 3 is valid.

Comment 6

[Jonathan Watt [:jwatt]]

You're right, it's not. In that case, can you change your test to check for that by changing it to:

<svg xmlns="http://www.w3.org/2000/svg" version="1.1">
  <title>Testing for x-y |points| attribute</title>

  <!-- From https://bugzilla.mozilla.org/show_bug.cgi?id=619630 -->

  <rect fill="lime" height="100%" width="100%"/>
  <rect width="100" height="100" fill="red"/>

  <!-- this should render -->
  <polygon points="-10-110 -10-10 -110-10 -110-110" fill="lime"
    transform="translate(110,110)"/>

  <!-- this should not -->
  <polygon fill="red" transform="translate(110,110)"
    points="-110-10-110-110-10-110-10-10"/>
</svg>

Note the change I made to your existing path in addition to the new path. From looking at your patch I think my modifications to your existing path will cause the test to fail (because it erroneously returns an error if the Y coordinate of the last coordinate pair is negative and against the X coordinate).

Comment 7

[Robert Longson [:longsonr]]

Attachment: that I can do.

Comment 8

[Jonathan Watt [:jwatt]]

You didn't make my modifications to the existing polygon...and note what I said about your current patch failing on the modified polygon.

Comment 9

[Robert Longson [:longsonr]]

Attachment: reftest as jwatt wants

Anything for a quiet life on the reftest front but your code comment isn't right as my code works just fine :-)

Comment 10

[Robert Longson [:longsonr]]

(In reply to comment #6)
> Note the change I made to your existing path in addition to the new path. From
> looking at your patch I think my modifications to your existing path will cause
> the test to fail (because it erroneously returns an error if the Y coordinate
> of the last coordinate pair is negative and against the X coordinate).

It doesn't.

Comment 11

[Robert Longson [:longsonr]]

Attachment: fix reftest.list

Comment 12

[Jonathan Watt [:jwatt]]

Comment on attachment 498534 [details] [diff] [review]
fix reftest.list

r=jwatt

Yeah, sorry, I'm gonna blame jetlag on misreading the code.

Thanks for changing both polygons in the test! I think having all the coordinate pairs test x-y is better even though your patch was correct.

Comment 13

[Jonathan Watt [:jwatt]]

http://hg.mozilla.org/mozilla-central/rev/b2dc4b542229