[Sane-analysis] [Fwd: Good news Re: Follow up Re: Random numbers]

[O. A. Rondon]

SANE paw++ users beware:

At yesterday's analysis meeting we noticed that Luwani's plots of his
study of the effect of the tracker on the pairs background seemed to
have a factor of 100 multiplying the bin contents of his histograms.

It's a result of a paw++ artifact (it may apply to plain paw, too. I
won't spend a minute checking that) which just chains the same ntuple
multiple times when the file name (including the path) exceeds ~ 72
characters. The safe filename length limit is buried at the bottom of
the Chain command help window, and it's even wrong: the "too long
filename" warning messages show a safe value of 111 characters, but it's
actually 72 or so.

Hopefully other collaborators were already aware of this issue, or were
lucky to have been using short names.

Cheers,

Oscar



-------------- next part --------------
An embedded message was scrubbed...
From: "O. A. Rondon" <or at virginia.edu>
Subject: Good news Re: Follow up Re: Random numbers
Date: Thu, 20 Mar 2014 18:41:07 -0400
Size: 4592
Url: https://mailman.jlab.org/pipermail/sane-analysis/attachments/20140320/53a90772/attachment.eml