Re: [PATCH] Atom 1.0 feeds for subversion

Malcolm, Thanks for your review!
I have made necessary changes and i'll attach the patch shortly in next
message. Please see my comments below which will be reflected in my
patch.

> * The usage message could be a lot clearer. I couldn't determine what
> the effect of some of the switches would be without using them.

Ok. I have added the enough document to such messages.

> * The output Atom file declares the XHTML namespace, but doesn't use it.

Yeah, you are correct. Henceforth i don't declare the XHTML namespace.

> * There's a fair bit of unnecessary trailing whitespace in the file.

I have corrected them.

> In my testing, the script generated a file called '.atom', not
> REPOS_NAME.atom.

I'm unable to repeat this issue. Can you please provide the test case ?

> The generated URL for each item is ITEM_URL?rev=N. That's a bit odd
> (it assumes a particular URL scheme), but if we decide to keep it,
> we should document it here.

Yes, i've now documented.

> Atom feeds require both a feed URL and item URL - or perhaps you can drop
> the feed URL if every item has a URL, I'm not exactly sure. Whichever it
> is, there's no point in allowing the user to be able to generate invalid
> Atom feeds (with URL's such as 'None', as currently happens).

The feed URL and item URL are required in the Atom feed. The domain part
of the feed URL and item URL may be different, so, we can't drop the
feed URL. If the user don't pass the feed URL or item URL, it is set as
"".

> * The generated feeds don't pass the feed validator[1], since they lack an
> atom:link entry with rel="self".

Yeah, you are correct. But i guess, it is just a warning message.
Moreover, as per Atom syndication format specified in [1], this entry is
optional. Anyhow, i tried to add this entry with 'rel="self"' attribute,
but still the warning continues to appear. I'm unsure howto overcome
this warning message.

> Finally, I needed to apply the attached patch to get this to work
> at all. I'm assuming you're using something outside of the standard
> Python libraries for XML pretty-printing, so I just disabled that bit.
> I think you should also be using a different method to get at the
> DOMImplementation object (see the patch).

Thanks for the patch. I've incorporated it in my patch.

[1] http://www.atomenabled.org/developers/syndication/#link , Atom
syndication format - Introduction

-- 
Regards,
Bhuvaneswaran