3 ms·
> 20 files changed, 307 insertions(+), 746 deletions(-) Two orders of magnitude faster and half the code size? That's nice.
by zokier 11y ago
> 20 files changed, 307 insertions(+), 746 deletions(-)
Two orders of magnitude faster and half the code size? That's nice.
- goalieca 11y agoi always find less is more with programming.
- __s 11y agoUntil you simplify your sort function into insert sort (which, admittedly, is really fast before container size hits triple digits)
- brianwawok 11y agoeven Quicksort got faster if you check for runs i.e. already sorted parts. More code but faster.
- lfx 11y ago> This effort started two years ago [..]. And it is well thought changes.
- snowwindwaves 11y agoThey probably worked on other things over the last 2 years as well. I wonder how many hours were actually spent on modifying 1000 lines of code. Probably a staggeringly high number, certainly all LOC are not equal.
- luckydude 11y agoWhat you said. I love seeing changesets like that.
- ionforce 11y agoI hate large change sets. They are impossible to code review. I would rather have death by one million patches than that.
- mackal 11y agoIt's 17 commits, each that do one thing.
- Too 11y agoSome times the problem isn't one piece of code that can be solved by adding an if-statement somewhere, some times it's the whole architecture of the module and the only way to solve that is to rip the whole thing out and replace it with something new. Reviewing such a "change" (it's more of a new feature actually) does require you to understand it completely and not just within the context of an insertion between two other lines of code. Being too scared to do such a replacement can easily lead to stagnation and fire-extinguishing that never solves the root cause of a problem.