Hi list, I found a workaround for a lot's of random segfault from polipo.
Please take a look at https://github.com/caeies/polipo-caeies/tree/topic/FixInfiniteLoop for my first fix. I'm not happy with it since I'm just fixing the symptom not the root cause. But despite my effort I was not able to find it (So polipo guru review welcome !). In short the commit message ( https://github.com/caeies/polipo-caeies/commit/27c0aa4725dd6febebd970d74f1bd919d3552899) : Client.c: Fix a recursive segfault. * httpClientNoticeRequest is called by httpServeObject with a novalidate set to 1. * But the reasons why we enter in this case (the loop case) is very obscure to me. The real fix is probably not here. * We were crashing the stack. * This looks like a very old problem and a reason of lot's of complaining on the web ... * Testing this case is rather simple on my computer. * Build a huge cache (~4000 url on the same host) * use the lowest memory footprint possible (as per the sample config) * use the proxyOffline to true * use http_proxy=127.0.0.1:8123 wget --input-file=list_url.txt --delete-after * before the fix, 2 run max make polipo enter the loop. * After, it never crashes again (and a log is displayed). [Not commited in] * some other bug is triggered, but that's another story. For me it looks like if the state of the object was nos consistent due to the memory pressure ... But not sure at all of this. Any ideas welcome ! Best regards, Caeies. ------------------------------------------------------------------------------ Live Security Virtual Conference Exclusive live event will cover all the ways today's security and threat landscape has changed and how IT managers can respond. Discussions will include endpoint security, mobile security and the latest in malware threats. http://www.accelacomm.com/jaw/sfrnl04242012/114/50122263/ _______________________________________________ Polipo-users mailing list Polipo-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/polipo-users
