On Mon, Feb 11, Michael Matz wrote:
you want to test how the machine behaves with O(N^2) algorithms :)
This of course could be fixed in libzypp by building a name->poolitem index itself and use it in the ByName iterator. But in the old libzypp a considerable amount of space and time was used for just building several dozen indices, and ideally we don't repeat this again. For one or two ByName filters (e.g. when done in reaction to a user action) such index is not necessary.
In most cases the applications use the byName iterator and filter by kind (or vice versa). This was ok as we mainatined a name index. But ideally we'll provide (and the the applications should use) a combined (kind,name) iterator. Unlike byName, we can process by(kind,name) without any string operations. (libsatsolver does not provide individual name and kind values). And if it's still to slow, we can index it. by(kind,name) is more frequently used than the individual ones. (e.g. to build a selectable). -- cu, Michael Andres +------------------------------------------------------------------+ Key fingerprint = 2DFA 5D73 18B1 E7EF A862 27AC 3FB8 9E3A 27C6 B0E4 +------------------------------------------------------------------+ Michael Andres YaST Development ma@novell.com SUSE LINUX Products GmbH, GF: Markus Rex, HRB 16746 (AG Nuernberg) Maxfeldstrasse 5, D-90409 Nuernberg, Germany, ++49 (0)911 - 740 53-0 +------------------------------------------------------------------+ -- To unsubscribe, e-mail: zypp-devel+unsubscribe@opensuse.org For additional commands, e-mail: zypp-devel+help@opensuse.org