On Wed, Sep 14, 2016 at 10:40 AM josvazg <josv...@gmail.com> wrote: > When analyzing the code the model could distinguish "executors", which cold be goroutines (main - the entry one, g1, g2, etc) or pieces of code under a certain lock variable. If a variable is ONLY accessed by a single "executor" (including locks) then it is race free.
Well, that's a particular case which can be successfully analyzed. The problem, in the general case, is that it is not possible to (always) tell when/in which order/if at all is the synchronization performed. Such arrangement is introduced easily by a conditional path where the condition is not statically decidable. The above does not rule out some programs can be statically analyzed. I am skeptical about the share of such programs, but that's just my guess. I would be very interested in knowing the real percentage of a proper corpus, provided someone writes this tool. -- -j -- You received this message because you are subscribed to the Google Groups "golang-nuts" group. To unsubscribe from this group and stop receiving emails from it, send an email to golang-nuts+unsubscr...@googlegroups.com. For more options, visit https://groups.google.com/d/optout.
