Re: ISSUE-2368: Problems with grammars for numbers

Hi Cameron,

See my comments below.

Le 06/12/2010 23:03, Cameron McCormack a écrit :
> ISSUE-2368 is as follows:
>
>    The grammar for path data uses the following grammar:
>    [[
>    floating-point-constant:
>        fractional-constant exponent?
>        | digit-sequence exponent
>    ]]
>    It seems that if no fractional-constant is used (no '.') an exponent
>    component is required.
>
>    The same problem is present in the grammar for points in polygons and
>    polylines, because the grammar is duplicated.
>
> I don’t think there is a problem, because floating-point-constant is not
> intended to capture numbers that match /^[0-9]+$/.  The only references
> to floating-point-constant are:
>
>    nonnegative-number:
>      integer-constant
>      | floating-point-constant
>    number:
>      sign? integer-constant
>      | sign? floating-point-constant
>    integer-constant:
>      digit-sequence
>
> so we’re safe.  Cyril: can you confirm?
I see my mistake. I had missed the 'integer-constant' type which indeed captures the numbers that match /^[0-9]+$/. I confirm we're safe.

>
> The issue also has:
>
>    Finally, there should be tests in the test suite with:
>    - a number with no digit after and before a "."
>    - a number no digit after exponent
>
> so I’ve raised ISSUE-2393 on the SVG 2 test suite to add such tests.
Thanks.

>
> Finally, I see that ISSUE-2368 is on “SVG 1.1 F2  Last Call”, but the
> original Last Call comment is not linked from the issue.  Was there one?
> Or was it just from Cyril?  If the latter, I guess it shouldn’t be a LC
> comment.
You're probably right. There was no LC comment, just a comment during the Paris F2F meeting.

Cyril
-- 
Cyril Concolato
Maître de Conférences/Associate Professor
Groupe Multimedia/Multimedia Group
Telecom ParisTech
46 rue Barrault
75 013 Paris, France
http://concolato.wp.institut-telecom.fr/

Received on Monday, 7 March 2011 08:53:50 UTC