Hi Matti, Matti Karnaattu wrote on Sat, Sep 13, 2014 at 12:25:54AM +0300:
> This also means that there is probably desire to dump GCC Yes, we strongly wish to replace GCC because we are stuck with the last available GPLv2 version, which is ancient by now. Newer GCC is GPLv3, and GPLv3 code won't make it into the tree. > favor of LLVM? I'm not the right person to report on the status of the replacement effort, and whether LLVM will be chosen. > Btw, I have been auditing code and I possible found some smelling code. > If I can confirm bug, I put diff soon to tech@. Usually, when we find potentially dangerous idioms in the tree, we improve the code no matter what and don't waste our time trying to figure out whether the specific instances are actually exploitable or not. Obviously, opinions on what is a dangerous idiom or a good idiom can vary, so i'm not saying what you call "smelling" is actually worth replacing - but it might well be. In some cases, when talking about *very* minor style improvements, there is also the following case: On their own, they may hardly be considered worth committing - making sure they are really correct takes time and testing, and a small risk of introducing bugs remains. However, if somebody is working on more important aspects of the code in question, such very minor improvements may go in in separate commits right before or after, because thorough testing is done, anyway. The only way to find out what we think about the "smelling" you found is to show diffs. Yours, Ingo
