[Orca-dev] changes in number of data points per interval

[jontankersleyatsabre at yahoo.com]

Has anybody thought through dealing with one percol file with more 
than one set of data points?  I've noticed at times that orca will 
complain (especially if we've done something with disks or network 
interfaces) about the wrong number of elements on a data line.  The 
code bails out of doing anything with that line.

If there are changes made to the system (new interface, disks, etc.) 
and the system is rebooted, there will be multiple heading lines in 
the file (or on an unstable system there will be multiple).  This will 
cause the data to NOT be plotted.

If it is a new heading line, can we 'reset' things to have it use the 
new line (until the next heading line) to count args, etc?

It has really cropped up with my trying to play with getting diskio 
stats out of orca.  I have the hooks in, but it is increases the 
chance for new heading lines and missing data from the plots.

I've been trying to think it through, but have been having too many 
interruptions lately.