On 28 February 2018 at 20:55, Linus Torvalds <torvalds@xxxxxxxxxxxxxxxxxxxx> wrote: > Bugs and mis-designed happen. They can be annoying and irritating and > embarrassing. But they are also just reality of any development. > As I am trying to use Sparse as the basis of a compiler I guess I'd rather not have bugs introduced ... it is hard to sell a compiler to users if that compiler occasionally just generates bad code. I think the problem is that the Sparse IR optimization has not yet reached a point of stability / correctness - and in that situation, adding more unstable changes rapidly means that it will never hit that stable point. Once it has reached that point I guess things would be different. Personally I can't see how in the absence of solid regression test framework this can be overcome. Again I am coming at this from the point of view that I cannot have a compiler that generates wrong code every so often. Regards Dibyendu -- To unsubscribe from this list: send the line "unsubscribe linux-sparse" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
