Finally with those segfaults out of the way I got this fix cooked up. By the way, is there a semi-comprehensive test suite for rrdtool commands? I.e. with these sample rrds and these sample rrdtool graph calls, we should get pictures that look like this and what do they actually look like?
See both modes (--full-size-mode with and without --no-legend) here: http://fenchurch.mc.vanderbilt.edu/ganglia/?m=mem_report <http://fenchurch.mc.vanderbilt.edu/ganglia/?m=mem_report&r=hour&s=descendin g&c=Vogon+Constructor+Fleet&h=&sh=1&hc=4> &r=hour&s=descending&c=Vogon+Constructor+Fleet&h=&sh=1&hc=4 The next step to finish this page is to add an option like --only-legend which produces an image consisting of ONLY the legend part of the graph. Obviously such a command could have a laxer syntax since it doesn't require any actual data, but what do you all think? I would put such an image at the top of the page. Actually, it would only need to be generated once for each metric and then the page could just call the static/cached legend image. This would basically be a convenience option for users to generate the legend for when the same graph from multiple sources many times and they don't want a legend on each one. Regards, Matt
rrdtool-fullsizemode+nolegend-fix.diff
Description: Binary data
_______________________________________________ rrd-developers mailing list rrd-developers@lists.oetiker.ch https://lists.oetiker.ch/cgi-bin/listinfo/rrd-developers
