@2010-12-09 05:25, Neil Kandalgaonkar: > On 12/8/10 3:40 PM, Andrew Garrett wrote: >> On Wed, Dec 8, 2010 at 3:11 AM, Trevor Parscal<[email protected]>wrote: >>> This particular change (the "don't delete line breaks" part of r73196) >>> should be reverted, Tim's good changes should be pushed upstream, and we >>> should be using a standard JSMin distribution whenever possible. >> I think the point is that it makes bug reports that much easier to >> understand when they're reported by somebody who isn't in debug mode. This >> has to be traded off against the performance advantages of removing blank >> lines. Is there any data on how significant the performance improvements >> are? > We already had this debate, in epic fashion, around September 30 - Oct > 1st 2010. > > Rough estimates were traded, that suggested that these hacks to JSMin would > > - increase the total Javascript weight by 2KB; > [...] 2KB with or without compression? AFAIR all calls were supposed to be gzipped and if gzip is not the lamest compressing method of all it will compress "\n" x 30 in 2B and so having 30 whitespace or just one shouldn't really be that much of a difference.
Besides 2KB? Powered by MW icon is heavier. Just lowering colour depth to 16 would save just as much. Changing two power icons into background + text + two small icons would probably save as much too. Regards, Nux. _______________________________________________ Wikitech-l mailing list [email protected] https://lists.wikimedia.org/mailman/listinfo/wikitech-l
