dita message
[Date Prev]
| [Thread Prev]
| [Thread Next]
| [Date Next]
--
[Date Index]
| [Thread Index]
| [List Home]
Subject: Re: [dita] Issue With Requirement for Trailing Space in class= Values
- From: Deborah_Pickett@moldflow.com
- To: Eliot Kimber <ekimber@reallysi.com>
- Date: Tue, 11 Mar 2008 10:44:46 +1100
Eliot Kimber <ekimber@reallysi.com> wrote on
08/03/2008 04:42:27 AM:
> In the process of implementing DITA support in our RSuite product,
we
> discovered that the current version of the MarkLogic XML repository
> strips trailing spaces from attribute values. This of course breaks
all
> hope of generic DITA processing unless we implement a workaround in
> advance of MarkLogic fixing their bug (which they might refuse to
fix).
This might not be much comfort, but that behaviour
is absolutely in contravention to the XML spec:
http://www.w3.org/TR/REC-xml/#AVNormalize
(Interpreting the spec: leading and trailing whitespace
must be left as-is if the attribute is declared as CDATA, which DITA @class
is.)
> - Does anyone know of any processors that would break if class= had
a
> trailing non-blank character.
DITA-OT's generalization code (generalize-on-conref,
as well as the standalone generalize and respecialize transforms) would
probably break. It breaks the class string into whitespace-separated
bits and assumes that each of them is a doctype/element pair.
--
Deborah Pickett
Information Architect, Moldflow Pty Ltd, Melbourne
Deborah_Pickett@moldflow.com
[Date Prev]
| [Thread Prev]
| [Thread Next]
| [Date Next]
--
[Date Index]
| [Thread Index]
| [List Home]