[tahoe-lafs-trac-stream] [tahoe-lafs] #1418: "cannot convert float NaN to integer" in next_power_of_k, during upload via helper

[tahoe-lafs]

#1418: "cannot convert float NaN to integer" in next_power_of_k, during upload via
helper
-------------------------+-------------------------------------------------
     Reporter:  rycee    |      Owner:  rycee
         Type:  defect   |     Status:  new
     Priority:  major    |  Milestone:  undecided
    Component:  code     |    Version:  1.8.2
   Resolution:           |   Keywords:  helper upload heisenbug
Launchpad Bug:           |  floatingpoint
-------------------------+-------------------------------------------------
Changes (by zooko):

 * owner:  zooko => rycee
 * status:  assigned => new


Comment:

 rycee: oh boy! You've managed to reproducibly isolate a way to make the
 bug happen or not happen! Now here's a question: can you run {{{git
 bisect}}} and get it to automatically run the unit tests in between each
 version change so that it finds exactly what patch makes the difference?
 Or do the equivalent manually -- try different versions and narrow down
 exactly what causes the problem?[[Image()]]

 Honestly, I'm still skeptical that anything we do in our 100% Python code
 could cause this bug, but perhaps something about the sequence of
 instructions that we're sending through the CPU accidentally avoids or
 accidentally triggers it...

 Also I am still willing to help you use some diagnostic tools like
 valgrind or the fp env from comment:29...

-- 
Ticket URL: <https://tahoe-lafs.org/trac/tahoe-lafs/ticket/1418#comment:34>
tahoe-lafs <https://tahoe-lafs.org>
secure decentralized storage