On Sat, Jul 10, 2010 at 6:31 AM, Balbir Singh <[email protected]> wrote: > * Dhaval Giani <[email protected]> [2010-07-10 05:57:05]: > >> > Dhaval, >> > >> > Thanks, I cloned the code and looked at it. Couple of quick questions >> > >> > 1. Why do we call the old code deprecated? We leave it to the users to >> > use which version they want and provide extensive documentation around >> > thread fork()/exec() issues with the older API >> >> Because I don't think anyone of us will be maintaining that code :-). >> > > I'll promise to maintain it. It is a well abstracted small piece of > code, so I am up for maintaining it. >
I would rather deprecate it. I prefer folks using the newer API, since it is a bit more sane with terms of locking and thread handling. >> > 2. Have you thought about how we can keep cgroup_context_s extendable >> > without breaking too much. Ideally it is opaque so we have no problems >> > in terms of API, but ABI can be an issue. One way to force it would >> > be to pack the structure, so that there is no scope for padding and >> > getting offsets mixed up later. >> > >> >> Honestly, I have not given it much thought as of now. I will get back >> to you in a day or two and let you know how to minimize ABI breakage. >> > > Sure, think about it. I think we need to create a -next branch in > libcg and merge this stuff there, get more feedback and push it in. > I don't want to go that way. We are not that big a project that it justifies so many processes. I would rather post patches, merge them and get more testing feedback. We can always fix bugs later on. Thanks Dhaval ------------------------------------------------------------------------------ This SF.net email is sponsored by Sprint What will you do first with EVO, the first 4G phone? Visit sprint.com/first -- http://p.sf.net/sfu/sprint-com-first _______________________________________________ Libcg-devel mailing list [email protected] https://lists.sourceforge.net/lists/listinfo/libcg-devel
