Closed jakubczaplicki closed 12 years ago
Hi Jakub, thanks for the patch. Do you have data collected with a recent bootchart2 that exhibit the problem you can share or are you just using the one on google code issue? If it is the latter since it looks like that's exactly the same data i think it's better to skip it if ltime == time instead of accounting the same values twice.
Hi Riccardo, difficult to say which version of bootchart was used to collect the data, as I used the one built in within Android code (Gingerbread). I used the pybootchartui to visualise data only. I suspect it's the old version of the bootchart (I haven’t looked into this specific Android code).
Ok, so if you cannot upload the tarball somewhere could you please verify that the chunks with the same time have the same data inside? The files are proc_diskstats.log and proc_ps.log.
I guess I could upload the file but I've got it on another machine to which I don't have access at the moment. But I can confirm that the chunks with the same time do NOT have the same data inside.
So i've reworked the patch a bit and while at it fixed another crash triggered by another tarball available in the same google code issue. Thanks!
Sometimes the log files may contain two samples timed with at same number. Running the script would fail because of interval division in parsing.py and samples.py. I've added ZeroDivisionError workaround by setting interval=1 if interval=0.
Also seen on http://code.google.com/p/pybootchartgui/issues/detail?id=13