I dont recall what version this replaced, but this appears to be a known bug as revealed by googling '.netrc fetchmail segfault'. The problem appears to be fetchmail trying to clear the passwords it read in .netrc from memory before freeing it. According to http://www.mhonarc.org/archive/html/fetchmail-friends/2006-01/msg00053.html This bug is also present in fetchmail 6.3.2 That link also points to the patch to fix it. My temporary work around is to rename .netrc, so it doesnt get read.
*** Bug 124486 has been marked as a duplicate of this bug. ***
Thanks for heads-up! 6.3.2-r2 ebuild is in CVS now, with the mentioned patch applied. Please give it a try. As for 6.2.* fetchmail, I'm not sure if we want to support them anymore. Users are encouraged to move to 6.3.*, and (at least from my experience), it's a painless transition. I'll have a look at fixing this in our 6.2.* fetchmail in the evening, I'm heading to work now. If you (or anyone) have time, feel free to post such patch here.
(In reply to comment #2) > Thanks for heads-up! 6.3.2-r2 ebuild is in CVS now, with the mentioned patch > applied. Please give it a try. It's actually -r1. Can we stabilize it on x86 and sparc (or rather the stable keywords from 6.3.2 should just have been kept, IMHO). Having a segfaulting version stable isn't good. :) > As for 6.2.* fetchmail, I'm not sure if we want to support them anymore. Users > are encouraged to move to 6.3.*, and (at least from my experience), it's a > painless transition. Indeed.
Ok. I have tested 6.3.2-r1, and it does not segfault when reading .netrc. But now I have to confess that I screwed up this bug report by claiming that 6.2.4.2-r1 was segfaulting. I took that version number from cached data, that did not reflect my current setup. So, to review, I upgraded *from* 6.2.4.2-r1, which I have just tested and verified working, to 6.3.2, which broke, and now 6.3.2-r1 has fixed the situation. My thanks for your quick response on this, and I apologise if my mistakes caused you any unneed work.
(In reply to comment #4) > So, to review, I upgraded *from* 6.2.4.2-r1, which I have just tested > and verified working, to 6.3.2, which broke, and now 6.3.2-r1 has > fixed the situation. My thanks for your quick response on this, and I > apologise if my mistakes caused you any unneed work. > No problem, Paul, I just came back from work, so I didn't investigate this for 6.2.* fetchmail yet. It's good to know that I don't have to, though. :) I'm closing this bug, thanks again for caring!
