[02:47] <wgrant> them
[21:34] <timrc> cjwatson: re: http://www.chiark.greenend.org.uk/~cjwatson/blog/re-signing-ppas.html -- not groking something.  You state "Then last week I finally got around to timing things on one of our staging systems so that we could estimate how long a full run would take. It was taking a little over two seconds per archive,".  You then did some optimizations and then state "On production, this took about 80
[21:34] <timrc> minutes for 1761 affected archives. This is over two seconds per modified archive".  Am I missing something?
[21:37] <timrc> I would be interest in knowing what the *actual* performance improvement per modified archive was.  Since it seems the avg is skewed by skipping over unmodified archives.
[21:37] <timrc> interested*
[22:22] <cjwatson> timrc: two seconds per modified archive when running over all archives, but as you say the unmodified archives used some time too.  it actually wound up as around .4 seconds for each modified archive.
[22:23] <cjwatson> iyswim
[22:24] <timrc> cjwatson: Ah, nice.
[22:24] <cjwatson> i.e. the final runtime was about 80 minutes, which involved stepping over about 70000 archives of which 1761 were modified.  do the divisions and it makes sense :)
[22:24] <cjwatson> the production system is similar hardware to the staging one, but quite a lot busier.
[22:25] <cjwatson> hence a bit slower.
[22:30] <timrc> Right but relative performance improvement was nice.
[22:30] <timrc> Nice work!
[23:26] <cjwatson> timrc: I've clarified the wording a bit now.