NoQ added a comment. In D54557#1301896 <https://reviews.llvm.org/D54557#1301896>, @Szelethus wrote:
> Wasn't the point of this patch to turn off part of this checkers > functionality because it's immature just yet? From what I understand it is > desired, but the FP rate is a little too high. I guess fixing that is the > project. > > Hmm, actually, tinkering with HTML files might be overkill, especially since > sphinx will hopefully end that era. Let's just add a TODO and let me deal > with it later when I reorganize the checker options. Sorry for talking > nonsense :D I would probably fine with shipping this other part as a lint check if we start shipping lint checks. I don't have any specific improvements in mind for that part. If somebody is willing to enforce "don't ever use things after move" policy in his code, this will be a good checker for such person. At the same time, i don't see how to improve that part so much that it can be enabled by default. I think projects that consist of "i believe it's impossible but you can try to come up with something" kind aren't very good for the open projects list because the person who will try to work on them may get disappointed when he realizes that his few at-a-glance solutions are bad and in fact nobody ever had any better solutions in mind. In D54557#1302069 <https://reviews.llvm.org/D54557#1302069>, @xazax.hun wrote: > In D54557#1300654 <https://reviews.llvm.org/D54557#1300654>, @NoQ wrote: > > > In D54557#1299736 <https://reviews.llvm.org/D54557#1299736>, @xazax.hun > > wrote: > > > > > It would be great to have a way to extend the list of (possibly non-stl) > > > types to check. But I do understand that the analyzer does not have a > > > great way to set such configuration options right now. > > > > > > Do you envision room for another attribute here? I.e., a class attribute > > that says "this object is always unsafe to use after move, unless a method > > annotated with `reinitializes` is called"? > > > Exactly :) My only concern is that I doubt users will end up passing such > options using command line options. Having file base configuration options > would be more convenient. They can be easily checked in the repository and > evolve together with the product (like the .clang-tidy files in the LLVM > repos). I mean, attribute, like, annotation in the source code. CHANGES SINCE LAST ACTION https://reviews.llvm.org/D54557/new/ https://reviews.llvm.org/D54557 _______________________________________________ cfe-commits mailing list cfe-commits@lists.llvm.org http://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
