On Sat, Sep 30, 2006 at 11:46:19PM +0200, Torsten Jerzembeck wrote: > >If I remove this line in the config file, however, I can actually > >reproduce the bug. It writes the file all right, but after that simply > >hangs with strace only showing futex()es and stat()s of /etc/localtime. > > "articlesperpage" is an additional configuration command provided by a > plugin, "paged-output.py". THe behaviour of rawdog doesn't change for me > either if I remove this plugin, so I don't think this has something to > do with the actual bug.
Hi, I can reproduce the bug, it's not related to download, but to writing out the pages. The problem lies with feedparser, it chokes on one of the articles in the output file, namely with Zugschlus' "Alturo-Ersatz" posting. If someone wants to verify this, check out this patch: http://greek0.net/div/rawdog-diff I've boiled down a simple script which demonstrates the problem in feedparser, it's attached. It fails with the feedparser version shipped with rawdog, the version from python-feedparser, and the upstream version 4.1. I'm hesitant to just reassign/clone this bug to feedparser though, as the problem happens when rawdog calls _resolveRelativeURIs(), which looks like an internal function which should not just be called from the outside of feedparser (by convention, the _ in the beginning indicates this). As the problem happens only with one posting on one blog I'm inclined to downgrade this bug to important, any objections? Cheers, Christian Aichinger
fp-testcase.tar.gz
Description: Binary data
signature.asc
Description: Digital signature
