Re: [xml-dev] Pete's blend for XMLlite

["Pete Cordell" <petexmldev@codalogic.com>]

Original Message From: "David Carlisle"

Thanks for the feedback David.  See comments below.

> On 08/12/2010 12:19, Pete Cordell wrote:
> evelopers for each suggestion.
>>
>> It's at:
>>
>> http://codalogic.com/xmllite/xmllite.html
>>
>> I intend to update it from time-to-time as I see things on the list, and
>> also as my position changes.
>>
> Could you fix the encoding, it's served as utf-8 but it apparently is 
> encoded as latin1 (I have to change the encoding by hand in the browser 
> view menu)
>
> In 2.1 (called 3.1 in the toc) you say remove the _internal_ DTD but the 
> rest of that section seems to imply that you mean removinbg the external 
> dtd as well, since you say no attribute defaults or entities.
> Either is a reasonable position to take for a suggestion like this, but I 
> wasn't clear which you intended.

What I was trying to do was avoid suggesting that DTDs should be erased from 
existence!  What I want to say is that XML files shouldn't rely on a DTD in 
order to be parsed correctly.  And you're right, this includes internal DTDs 
and externally referenced DTDs.  External DTDs can obviously still exist as 
standalone schema definitions.

I've attempted to update the page accordingly.

> > 2.2 - Discard CDATA Sections
>
> Discarding CDATA sections would have some benefits, but not the one 
> stated, as the sequence ]]> would still have to be trapped and escaped if 
> xmllite was to be compatible with xml 1.0.

That's a good point.  I had forgotten about that!  I've put something in 
about this.

> > 2.3 - Limit Character Encodings to UTF-8 and UTF-16
> see comment above:-)

What's the smiley for AARGGH!!!!

> > 3.1.1 - Impact on existing XML parsers
> this should also mention the (much bigger) impact on XPath that is 
> proposed here.

I'm not that up on XPath, particularly on the implementation of XPath 
libraries.  If someone else wants to comment I could incorporate it, or link 
to it.

> > 4.1.3 - Impact on applications
>
> > No impact.
> The impact here is that applications reading files intended for xml1.0 
> using & n b s p ; will be subject to silent data corruption.Changes that 
> make previously illegal syntax legal, or previously legal syntax illegal 
> are _much_ safer than changes that change the meaning of existing syntax.

There is a versioning issue to be worked out here.  I'm not sure what the 
answer is yet.  I think all of the points below are affected by this.

> > 4.3 - Allow Nested Comments
>
> you don't say what parsing  rules you want for comments.
>
> <!--  aaaa <!-- -->
>
> is currently illegal. Your suggested change 4.2 would make it legal and 
> have comment text " aaaa <! "  but the addition of 4.3 would presumably 
> change the rules so that --> closed only the inner comment, and so some
> later part of the document is also commented. So effectively 4.3 would 
> imply changing the existing rule that you can't have a comment 
> containing -- to a new rule saying you can't have a comment containing 
> <!--.
>
> > 4.4 - Preserve White Space in Attributes
> again changing the meaning of existing syntax implies silent data 
> corruption.

Thanks,

Pete Cordell
Codalogic Ltd
Interface XML to C++ the easy way using C++ XML
data binding to convert XSD schemas to C++ classes.
Visit http://codalogic.com/lmx/ or http://www.xml2cpp.com
for more info