On Nov 13, 2007 12:47 PM, Nagy Gabor <ngaba@bibl.u-szeged.hu> wrote:
Any notable optimization worth the hack.
I just need to comment here that I'd prefer we _not_ do this in pacman. "Any optimization is worth it" is a REALLY bad mantra when it comes to code, in my experience. Code should be clean and elegant _first_ and as optimized as possible _second_. So if optimizations break code clarity and add unnecessary stuff for a 0.1 second improvement, it's not worth it.
Sure large improvements are worth a loss in code clarity, but, as Dan actually proved with the timings, there's very little room for improvement here.
First of all, I don't want to start a flame (again): I wouldn't have replied anything if you just had said: "I don't like this. Forget it.". You are the leader, you can do it. No question. (I forgot it ;-). But you should see, that your reasoning is simply impossible to keep, because "clean code" and "elegance" are _subjetive_ things: -_imho_ sync.c is not elegant at all -_imho_ your alpm_list patch (first prev == tail) is not elegant (I would have preferred alpm_list_add_first) -_imho_ a function which we could _radically_ speed-up with a 'first' parameter and an 'if(first) return ...' line, is not elegant ... We are different, that's it. Well, maybe I misled you with this: "Any notable optimization worth the hack.", sry. Notable and hack are subjective things too ;-) Bye, ngaba ---------------------------------------------------- SZTE Egyetemi Könyvtár - http://www.bibl.u-szeged.hu This mail sent through IMP: http://horde.org/imp/