Vitaly Harisov wrote:
> 
> See attached test-case.
> 
> Steps to reproduce:
> 
> 1) Ensure that XXE option "Ignore <?xml-stylesheet?>" is turned off.
> 2) Open first-child-test-case.xml in XXE.
> 3) para2 is first child of issue element, it marked with green border,
>    para1 is not first child.
> 4) remove abstract element, para1 is first child now, but not marked
>    with green border. This is bug.

First of all, thanks for your feedback and your very precise bug report.

* Yes, this is a bug from your point of view because it does not work as 
expected.

* No, this is not a bug (i.e. a logical error in some code) from our 
developer's point of view because we have never taken the time to code 
the desired behavior.

Excerpt from "XMLmind XML Editor - Support of Cascading Style Sheets 
(W3C CSS)", "Chapter 2. Restrictions", 
http://www.xmlmind.com/xmleditor/_distrib/docs/csssupport/ch02.html

---
The style of an element may be specified using a selector containing the 
:first-child pseudo-class but the style of such element is not 
recomputed when sibling elements are inserted or deleted.

The workaround is to manually ``refresh'' the view of the parent of such 
element using Select|Refresh.
---

We can agree to call this ''a limitation which makes the :first-child 
pseudo-class unusable in XXE''.



> 
> ------------------------------------------------------------------------
> 
> *
> {
>       display: block;
>       margin: 1em;
> }
> 
> issue:before
> {
>       content: "Issue";
> }
> 
> issue para:first-child
> {
>       border: 1px solid green;
> }
> 
> 
> ------------------------------------------------------------------------
> 
> <?xml version="1.0" encoding="windows-1251"?>
> <!DOCTYPE first-child-test-case SYSTEM "first-child-test-case.dtd">
> <?xml-stylesheet href="first-child-test-case.css" type="text/css"?>
> <first-child-test-case>
>   <issue>
>     <abstract>abstract</abstract>
> 
>     <para>para1</para>
>   </issue>
> 
>   <issue>
>     <para>para2</para>
>   </issue>
> </first-child-test-case>
> 
> 
> ------------------------------------------------------------------------



Reply via email to