On Tue, Nov 27, 2001 at 02:05:30AM +0100, Emiliano wrote:
> > Proper one requires some rethinking how to place data in CDATA
> > because XML standard assumes that data there should be unchanged by XML
> > parser so Repligard follows this asumption. May be escaping would help but
> > I'm not sure -- it may cause other problems.
>
> I don't think the XML standard assumes data would be unchanged. XML tools
> work from the XML InfoSet, not the literal data in the file. You could have
> everything in CDATA sections, or everything qouted, and tools that properly
> use the XML API (be it SAX or the DOM) can't and shouldn't be able to tell
> the difference.
I mean that we can't simply replace ]]> by its &symbol;-based variant
because this sequence might already be in the data and restoration of ]]>
will lead to data corruption (as opposed to unescaping). This means that
we have to implement full XML-aware escaping mechanism for all the data
going as 'string/text' types into XML file (in terms of Repligard data types).
This is serious change and I can't say how it is safe to do now,
especially given that output encoding is dictated now by a SQL server and
thus additional conversion to UTF-8 before the transformation would be
desirable to avoid unpredicted side effects of the transformation.
--
/ Alexander Bokovoy
$ cat /proc/identity >~/.signature
`Senior software developer and analyst for SaM-Solutions Ltd.`
---
Nov 21 20:58:58 alconost kernel: VFS: Busy inodes after unmount.
Self-destruct in 5 seconds. Have a nice day...
---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
