[Halld-offline] bug fixes to the evio-2.0 io library

[Richard Jones]

Elliott,

Back when I was first trying to read events from the evio files created during our recent test run, I ran into a number of problems with the evio library.  After I fixed the initial seg faults that were happening because of a mismatched new/delete pair, there was a nasty memory leak that prevented me from getting past a few million events.  The resident set size would grow past 4GB and then the application would crash when it hit my limits defined in /etc/security/limits.conf.

I was able to trace the bug to parts of the evio library source and install a fix.  You can look at my fixes and decide whether to use them or fix it in some other way.  My revised source tree is available at

http://zeus.phys.uconn.edu/halld/gridwork/packages/evio/2.0b/evio-2.0b.tgz

My memory is alreadying fading about the details.  I fixed it late one night when I just wanted it to work so I could look at my data, and forgot that I needed to pass the changes back upstream.  I hope this is not already too old to be useful.

-Richard J.

-------------- next part --------------
A non-text attachment was scrubbed...
Name: smime.p7s
Type: application/pkcs7-signature
Size: 4092 bytes
Desc: S/MIME Cryptographic Signature
URL: <https://mailman.jlab.org/pipermail/halld-offline/attachments/20110524/60ea2639/attachment.p7s>