Description Christian Vogt 2007-11-13 16:43:42 EST

Created attachment 82809 [details]
patch for AbstractEditHelper

I have attached a patch which reduces the peak memory consumption in large delete scenarios. This is done by reordering the command creation such that the parent delete command is created after the child delete commands.

As well, I have skipped the check to canExecute (see patch) because under normal circumstances canExecute will return true, and this check ends up causing slow downs in large scenarios. By skipping this check, the delete scenario will increase in speed, but delaying this check in a failing scenario will cause the fail to propagate slower. Its a trade off, but the happy path scenario is improved.

One more performance fix is related to Bug 209709. Prior to returning the edit command produced by the edit helper, the command should be reduced to minimize the amount of nested commands.