Greetings,

I've been progressing fairly quickly with the migration of pkg's
depot.py, face.py, etc. to use CherryPy. The server status summary
page, versions, and even search all appear to be working.

So far, the only warts have been the default dispatch system it uses
and the varying quality of the documentation (some complex parts are
well documented, seemingly simple ones are not) :-(

I do have *a* dispatch structure in place that seems to be working,
but I may end up using Routes instead [1] [2].

Just recently,  I've run into a small snag that is requiring me to
make a *lot* of changes to various parts of the code, and before I do
it, I wanted to get some opinions here.

One nice thing about BaseHTTPServer is that you have the request.wfile
that you can easily write to anytime you need to stream data. I like
the simplicity of having a simple io object to write to.

CherryPy, in contrast, takes a very different approach. To stream
data, you must use Python's generators/yield functionality instead,
which means changing a lot of routines from writing their output and
returning, to "yielding" their output and then returning. That doesn't
sound like a big deal in practice, but from what I understand this
enforces certain limitations.

As such, I was wondering if anyone has any comments or objections to
such a code shift?

This is explained in greater detail here at the CherryPy website [3].

Cheers,
-- 
Shawn Walker

"To err is human -- and to blame it on a computer is even more so." -
Robert Orben

[1] http://routes.groovie.org/
[2] http://tools.cherrypy.org/wiki/RoutesUrlGeneration
[3] http://www.cherrypy.org/wiki/ReturnVsYield
_______________________________________________
pkg-discuss mailing list
[email protected]
http://mail.opensolaris.org/mailman/listinfo/pkg-discuss

Reply via email to