gnu-arch-users
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

[Gnu-arch-users] Re: File-tpye plug-in architecture for Arch?


From: michael josenhans
Subject: [Gnu-arch-users] Re: File-tpye plug-in architecture for Arch?
Date: Tue, 23 Dec 2003 02:54:56 +0100
User-agent: Mozilla/5.0 (X11; U; Linux i686; en-US; rv:1.5) Gecko/20031016

Charles Duffy wrote:
On Mon, 2003-12-22 at 17:16, michael josenhans wrote:

Any good XML patch should take the DTD into account when merging.


Much of the argument here is that knowing that the new document
validates is insufficient to know that the new document is semantically
valid (ie. that when loaded into the app which created the two documents
from which it was created, said app will represent a document that is in
fact The Right Thing).

If there is an DTD. The application must be able to eat any valid DTD.
Especially, if there are open formats like OpenOffice, where not everthing is under control of one implemetation, this in not an disadvantage.

Consider the case where an app loading a merged document presumes that
each style element will have a distinct "name" value, but where each
branch added a different style element with the same name;

This is the same problem like archs file tagging. Just lets the application put a unique tag to the node. Either a running number containing the branch name or use a random number.

If the attched tage to the node is identical, it was a change otherwise a new node.

a DTD cannot
validate that this did not occur,

and so may produce a document which
validates against the DTD but still crashes the app which created the
document.

If the application is buggy, it must be fixed. I can not see why we should respect tools that creast when someone changes anything in a libary that is used by the output filter.

What you need here for XML files is:

xml-md5sum


Better approach: Normalize the trees, then use a regular plain-vanille
md5sum.

fast & easy. cool.

Execept that we likely need a specical normalization reatment for unorderd nodes. These likely need to be come sorted alpahabetiaclly before applying the checksum.

Michael






reply via email to

[Prev in Thread] Current Thread [Next in Thread]