On 15-May-18 7:24 AM, Remy Horton wrote:
On 14/05/2018 12:29, Burakov, Anatoly wrote:
[..]
This failure is not caused by this patchset, and you should get similar
failures on master if you get these while testing my patchset. I am not
able to reproduce this issue, but i'll double-check the bounded reserve
code with a fine-toothed comb anyway.
I reliably get the failure with V3 applied to RC2, but so far haven't
been able to replicate with any other combination (clean RC2, RC2+V5,
master+V4, etc). Odd..
I've retested on plain v3+rc2 (i was previously testing on master), and
i can reproduce the failure as well. The bug that was causing this issue
was fixed on rebase in v4. So, rc2+v4/v5 no longer has this issue.
--
Thanks,
Anatoly
