On Thu, Oct 28, 2010 at 9:55 PM, Alex Rousskov <rouss...@measurement-factory.com> wrote: > On 10/28/2010 01:09 PM, Kinkie wrote: > >> Then, if I remember it correctly, that code is broken and needs to be >> fixed :( > > Isn't that what you were doing? If you want to fix the code _and_ follow the > cbdata API rules, use CbcPointer instead of a raw queue pointer. The rules > were not required for C structures to work and are difficult to implement > correctly for complex C++ classes. CbcPointer solves that problem.
Well, my aim is a very modest "let the damn thing build". I do not yet understand the intricacies of cbdata, and thus I am not able to understand when it is abused and when the abuse is benign. > Keep in mind that a lot of cbdata-using code violates these very good rules, > and most violations are, like this one, benign from code logic point of > view. That is why these violations are still being copied and added to > Squid. Hopefully, we will not make the same mistake twice when cbdata API is > updated. I hope! I see lots of thought being spent in making sure things are done at least sensibly :) -- /kinkie
