Roland Schwarz <roland.schwarz@chello.at> writes:
Peter Dimov wrote:
If the windows version conforms to the specification and passes its tests, there's nothing to be concerned about;
Of course. What I wanted to point out is that I am not sure if William also "is there", ready for a switch to HEAD. E.g. I have been told by him that he has implemented a different mutex algorithm than Terekhovs. I am not sure if this algorithm is already at the same level of reliability.
By "William", do you really mean me, Anthony Williams? If so, then most of the code is ready for a switch to HEAD, I believe. The mutex algorithm is fairly simple, but I guess it's the condition variables that you are thinking of. I believe it to be reliable, but that can only be borne out with experience. Putting it on HEAD would allow more people to try it out, and more people would actually look at the code. If it's broken, I'd like to know. The read-write mutexes aren't ready, yet, but the old ones are broken anyway. I have mostly added additional tests, rather than removing old ones. These should be moved to HEAD too.
I do not know if William will revert his changes to Bills code base again, since he already changed a lot more. I also do not know if he already tried to run the regressions on his changes.
I've run the tests on my changes many times, and they all pass, for the bits I've done. For the areas I haven't touched yet (TSS, thread launching, etc.), I'm happy to revert to Bill's code. For the areas I have reimplemented, I would prefer to go with my new code. In some cases, I believe it's an active improvement (e.g. call_once), and in others I think it will make it easier for me to extend, and alter the interface in line with the current standardization proposals. Anthony -- Anthony Williams Software Developer Just Software Solutions Ltd http://www.justsoftwaresolutions.co.uk