I had an issue with an Apache output filter written in Python, using mod_python, when together with Apache. After several days (a whole week to be more precise) trying to figure out what was causing it I seem to be able to find the problem, apparently. Here’s my post to the mod_python mailing list detailing the issue.

Ok, I was finally able to figure out part of the problem and a solution for it. The issue is that if the response fits in ‘io_buffer_size’ which defaults to 8k, then the output filters get all called in the same iteration.

In particular, there’s one filter called ‘HTTP_HEADER’ which is responsible for emitting the headers. Obviously, to make things work right, it should be the last one to be called, which it is apparently. However, when the response is larger than io_buffer_size, the filters get called multiple times.

 Unfortunately, HTTP_HEADER is one such filter, so the first time around the mod_python filter gets called, but hasn’t seen EOS, so it doesn’t change the headers. Then the HTTP_HEADER gets called and sends the headers. Then after that you can’t change the headers anymore.

My initial hack was to set ‘req.assbackwards’, which prevented the HTTP_HEADER from doing it’s thing, and then when the mod_python filter sees EOS then I set ‘req.assbackwards’ to 1 and call req.add_output_filter(‘HTTP_HEADER’). That works, but depends on filter implementation internals, so I felt I needed a better fix.

Then I stumble upon ‘mod_filter’, which has lots of interesting configuration knobs. The interesting one is ‘FilterProtocol’, which lets you specify if your filter will change the content length and if it supports byte ranges, and also that you want your filter to be the first one in the chain. So I toggled those settings, and it magically works.

Here’s my configuration file: https://svn.enfoldsystems.com/browse/public/enfold.lxml/trunk/docs/sample-apache.conf?rev=1321&view=auto

It would be great to document this somewhere, so that the next person doesn’t have to suffer that much. Maybe add some notes to mod_python docs? I will certainly write a blog entry.

Hopefully that will save someone’s time when trying to write output filters for Apache with mod_python. I just hope more people would post their solutions when they have to work around or fix complex issues. It would make everyone’s life much better.

As usual, those binaries are statically compiled so there’s no extra dependencies to install. They have been built against libxml2 2.6.28 and libxslt 1.1.19.

Have fun, and please report any bugs/crashes to the lxml-dev at codespeak dot net mailing list.

Along with the Plone Announcement, he also announced the future
availability of Enfold Server 2.9 and 3.0, sporting Plone 2.1.3 and
Plone 2.5 respectively.

Of special note, as previously mentioned, are the inclusion of XML
tools libxml2 and lxml.

Alan Runyan is the head of Enfold Systems, a Texas-based Plone
Consulting firm that provides packaged versions of Plone with tight
Windows integration product firm specializing in low-cost, high-impact commercial software.

Special thanks to Rocky Burt and Hanno for helping out with QA of the
new Plone installers.

When I’ve first seen it, I thought it was a cool idea, but it didn’t
seem like it could achieve a practical value for me. After all, at
that time I was a VB programmer, and my concept of useful was dragging
buttons to draw forms and use Access for databases wink.

But time has passed, and I started to hear more and more about this
beast. Particularly since I’ve met Paul, which I think from the
persons that I talk to regularly, is the one who believes more than
anyone in XMLTopia, a world where all applications will be able to
interchange information using XML.

I fear though, that I may have joined the wave way too late. Now
that I’m starting to effectively use XML, I see more and more starting
to use RDF in practice. Yes, RDF can be serialized to and from
XML. However it seems like this serialization can occur in several
different ways, and still keep the same semantic value.

I’ve been working hard to try to make some sense out of the semantic
web, specially since it seems to be gaining speed amongst some
developers that I know. However, I must confess that it melts my brain
to try to follow their conversations. I’ve spent a couple hours
chatting with Joe ‘Arnia’ Geldart and it took me a full 3 days to
digest just about 5% of the information he gave me :)

Anyway, as I’ve said already, the situation is improving on my
end. I’ve spent most of the last two weeks on building a import/export
framework for Archetypes-based products. There’s a good summary about
it on PLIP44, which btw, I should update to reflect some of the
changes I’ve done to the xml dialect and schema.

To finish, it has been a pleasant time working on this. Much more
pleasant than I would expect, and I really hope more and more people
start drinking the XML kool-aid so we can meet all in XMLTopia one day!

Then I did remember about a chat with Paul, where he mentioned that he
was writing XML documents using XSD as the ‘source’ for the document
format definition, and send him an email asking about it. He
said to be very disappointed with me, because I hadn’t rewritten
XMLSpy while the teacher was starting. :) I told him it would not be a
problem as long as I had Kiwi and Glade with me, but that wasn’t the
case (90% of the desktops in university still run Windows, on a Novell
network, and the remaining 10% are dual boot, with Windows and Red Hat
8).

So, Paul recommended me looking at <oXygen/> and RelaxNG (spelled
relax-en-gee), which seems to be what he’s been using lately. He also
told me about Morphon, which seems to have been turned on a free
license. The downside (for me) is that Morphon seems to be
java-based. On a quick look looks like oXygen is best, having support
for RelaxNG, and being able to be run as an Eclipse plugin. It also
looks like its able to convert from XSD to RelaxNG and forth. Pretty
impressive list of features they have there.

Lastly, all this discussion reminded me that I need to try Eclipse
someday soon, as everyone seems to be talking good about it.

• Take some bottles of wine with me. I promised that to Phillip.
• Ride the Bluesmobile!
• Take lots of photos for my fiancee at home.
• Chat with Ben Saller (and all the folks there) about the future of
  Archetypes
• Lobby for breaking 2.1 compatibility in Plone and Archetypes HEAD
• Talk about backporting parts of the CA to Zope 2, ASAP!
• Listen carefully to Paul’s ideas about XML/RDF and Archetypes
• Meet personally the cool folks that I only know from IRC (that’s
  implicit)

That’s a pretty rough plan, but I wanted to write it down before
forgetting something. I’m sure there’s something missing anyway, and I
hope to remember what is it. :^)