On Wed, 4 May 2005, Chris Mason wrote: > On Wednesday 04 May 2005 17:47, Geert Bosch wrote: > > From your tests it would seem that the zdelta version is the only one > > to provide a uniform improvement over plain git. As it also seems the > > simplest approach, I wonder why the consensus is that using xdiff > > would be better? > > zdelta seems to be a research project. It does compress better than the xdiff > lib, but the speed improvements against xdiff(1) are probably because the > resulting tree is smaller. I favor the xdiff code because it's so much > smaller, and seems easier for us to maintain. Yep. And compression can be improved without changing de decompressor since the decompressor is only a replay of what the compressor found to be redundent. That redundency searching can probably be improved wrt to the current code. And FRankly considering about 300 lines of code to create a delta and 60 lines to expand it is hard to beat maintenance wise. > For performance, there's still quite a bit of tuning that can be done in terms > of when and how we delta. Indeed. Nicolas - To unsubscribe from this list: send the line "unsubscribe git" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.htmlReceived on Thu May 05 13:11:32 2005
This archive was generated by hypermail 2.1.8 : 2005-05-05 13:11:33 EST