[Babeldoc-devel] Scanner and Feeders

[Dejan K. <dej...@nb...>]

Hi everybody!
I have just retuned from the conference were I had Babeldoc presentation
last week. There were about 50-100 people in the room during presentation
and it seems to me that I haven't been too boring ;-)

Anyway I am now fresh and ready for working on Babeldoc! We are planning to
have some modifications in the scanner module. In a few words, plan is to
separate processes for scaning and feeding documents. I have already written
some kind of queue and feeder implementations for documents, but after
reviewing Bruce's code for asynchronous feeder I have realized that there is
no need for another feeder implementation in scanner module. The biggest
difference between them is that mine implementation serializes documents so
in case of some crach, documents will not be lost. This is important in case
of e-mail messages because these messages have to be deleted from server
(pop3 protocol does not provide functionality for moving messages to
different folders). There are other situations, too. I think that this
feature is important so it should be implemented in core module, not just
scanner. So instead of subclasses Bruce's class I suggest adding serialize
functionality to basic AsynchronousFeeder class. Opinions?

Dejan