On Oct 25, 2011, at 16:32 , Patrick Ohly wrote: > On Tue, 2011-10-25 at 15:46 +0200, Lukas Zeller wrote: >> But if the superdatastore was changed from probing with ReadItem to >> directly probing/deleting with DeleteItem, the error reporting would >> look similar, as it would show failed deletes for those subdatastores >> not containing the object in question. > > No, because I never consider attempts to delete a non-existent item as > an error and thus would always suppress the error message.
But still return 404 to the engine, I assume. > After all, the desired outcome was achieved. > > This is different for reading where not getting the item data may or may > not be relevant. Yes, seen from a standpoint where the result of the operation counts, and not the process, a failing read is relevant while a failing delete is not - the desired result for the latter (item gone) is ok, even if the item has disappeared sometime before. Makes all perfect sense to me now :-) Lukas Zeller, plan44.ch l...@plan44.ch - www.plan44.ch _______________________________________________ os-libsynthesis mailing list os-libsynthesis@synthesis.ch http://lists.synthesis.ch/mailman/listinfo/os-libsynthesis