Core: Cannot purge anymore a complex obs that has lost its complex data.

Hi all, thanks for your input,

I have reworded and changed the issue to ‘Add a forcePurge parameter to ObsService#purgeObs.


For some reason AbstractHandler does not increment the name of “same titled files” with a suffix higher than 99 (two digits). See here. Now the nightmare scenario. Imagine a device that sends its captured images always named as ‘image.jpg’. When the 101th picture is sent, AbstractHandler returns an outputfile pointing to ‘image_99.jpg’… for the second time. This confuses the parent handler in believing that it is safe to write there again, leading to a file override and a ghost file for obs #100 (the obs behind the 100th ‘image.jpg’.) I don’t know how many files our client lost like that (maybe 20) before we urge them to stop capturing images with that device. We worked around this by ensuring that VDUI always sends out unique file names from its client-side. We had created swiftly an internal issue about AbstractHandler’s behaviour but ended up pushing a hot fix on VDUI first, then we moved to other things and the actual bug got a little forgotten. I will raise the original issue on TRUNK as well of course. In fact TRUNK-5077 represents an effort to allow VDUI to clean up what was once produced by the original issue.