#611: test_repairer test_repair_from_corruption_of_1 sometimes errors
---------------------------+------------------------------------------------
Reporter: warner | Owner: zooko
Type: defect | Status: assigned
Priority: major | Milestone: 1.3.1
Component: code-encoding | Version: 1.2.0
Keywords: | Launchpad_bug:
---------------------------+------------------------------------------------
Changes (by zooko):
* cc: tahoe-...@… (added)
Comment:
Okay, I found the bug. If the uploader requests a read from the
DownUpConnector after the downloader has finished writing and before the
downloader closes, then that request will never be satisfied.
[20090212230447-92b7f-11ef58cee99e471c69580d675b5bffedc038df29] fixes this
and tests it. Please review! This is the last code change that is
scheduled to go into tahoe-1.3.0 and I would hate to have inserted a bug
into tahoe-1.3.0 at the last moment, so please review this patch while I
work on CREDITS and relnotes.txt. :-)
By the way, I also ran another 4876 iterations of
{{{test_repair_from_deletion_of_1}}} (this time without debugprintouts)
with no deviation in behavior, and I am wondering why there is any
variation at all in the behavior of this test on Brian's machine. Could
it be that system test is using loopback and that the loopback device
drops or reorders packets very rarely on Brian's machine but never or at
least even more rarely on mine?
--
Ticket URL: <http://allmydata.org/trac/tahoe/ticket/611#comment:5>
tahoe-lafs <http://allmydata.org>
secure decentralized file storage grid
_______________________________________________
tahoe-dev mailing list
[email protected]
http://allmydata.org/cgi-bin/mailman/listinfo/tahoe-dev
