On Sun, Aug 29, 2010 at 10:59 AM, Dustin J. Mitchell <dus...@zmanda.com> wrote: > Fubar means "f*'d up beyond all repair", which I hardly think applies > to a few warnings! These are minor bugs, which I'll work on a patch > for. Chris's analysis of the error on line 1312 is correct.
Here's the patch to fix the warnings. This isn't ready for review, because it doesn't have any new tests. Some of these may represent more significant bugs in the report code, but I won't know until I see the trace log. http://github.com/djmitche/amanda/commit/z11827 By the way, this is some particularly hairy perl code. There's no clean way to parse these logs! That said, there's definitely some cleanup that could be done here. On Sun, Aug 29, 2010 at 2:24 PM, Chris Nighswonger <cnighswon...@foundations.edu> wrote: > Regardless of the real cause of this problem, it seems to be very bad > practice to have a sub return a non-numeric value to a test looking > for a numeric value. And at the very least, fixing it would clean up > the logs from distracting errors. Absolutely. The divzero_col sub, in operation, formats its arguments into a string for display in the report. The bug was using that $compression in a numeric context, and in fact trying to format it again by calling $fmt_col_field. Dustin -- Open Source Storage Engineer http://www.zmanda.com
