Hi all, Is anyone else seeing silly amounts of segfaults with Feedtools? They''re regular enough to make it unusable, but I haven''t pinned down what''s triggering them yet. Here''s the error:> deadlock 0xb75970a8: run:-/usr/lib/ruby/1.8/timeout.rb:42: [BUG] Segmentation fault > ruby 1.8.3 (2005-06-23) [i486-linux]Naturally, that knocks over WEBrick, and causes 500 errors in Lighttpd. That ruby 1.8.3 is a reference to the fact that I''m on Ubuntu Breezy - it''s not so far otherwise been a problem. The only parts of the FeedTools API I''m using are FeedTools::Feed.open() and FeedTools::Feed#title. I''ve got feedtools 0.2.23 installed, and the database is sqlite3. Rails is 1.0. Any ideas? sporkmonger.com seems to be down at the moment, so I can''t look there, but Google didn''t throw up anything relevant... -- Alex
Is it possible that the feed that you''re trying to open is not accessible and that''s why it it''s timing out? Does this happen with all feeds or one particular feed? -------------- next part -------------- An HTML attachment was scrubbed... URL: http://wrath.rubyonrails.org/pipermail/rails/attachments/20060322/73d40f33/attachment.html
eduard wrote:> Is it possible that the feed that you''re trying to open is not > accessible and that''s why it it''s timing out?It''s not timing out - when it fails, it fails hard and fast (less than a second after the method call) with a deadlock. It stinks of threads not behaving themselves, but I don''t know what I can be doing differently to everyone else if I''m the only one seeing this. An unavailable RSS feed shouldn''t be able to knock over the Ruby interpreter (and thus the app)... -- Alex
On 2006-03-22 04:21:16 -0500, Alex Young <alex@blackkettle.org> said:> eduard wrote: >> Is it possible that the feed that you''re trying to open is not >> accessible and that''s why it it''s timing out? > It''s not timing out - when it fails, it fails hard and fast (less than > a second after the method call) with a deadlock. It stinks of threads > not behaving themselves, but I don''t know what I can be doing > differently to everyone else if I''m the only one seeing this. An > unavailable RSS feed shouldn''t be able to knock over the Ruby > interpreter (and thus the app)...Are you using feed caching? I have seen the segfaults before, but not since I started driving the feed cache directly, and only calling Feed.open when there is new stuff. My motivation for doing this is that I wanted more visibility and control over the http request and response (for example, I don''t want to bother parsing a cached version of a feed that returns a 304) , but it''s also made things faster. -Ross http://rossnotes.com
Ross Karchner wrote:> Are you using feed caching?My only reason for using it rather than writing my own parser was to get the caching :-)> I have seen the segfaults before, but not since I started driving the > feed cache directly, and only calling Feed.open when there is new stuff.Have you got an example of working code I could take a look at, by any chance?> My motivation for doing this is that I wanted more visibility and > control over the http request and response (for example, I don''t want to > bother parsing a cached version of a feed that returns a 304) , but it''s > also made things faster.That''s sensible, and also something I''d be wanting to do once I''ve got the basics up and running. -- Alex