Peter Djalaliev wrote: > When you call GetUniqueIdentifier(), does "unique" imply that the > identifier will be unique for that layer name, or that the identifier > will just be unique among the identities issued so far?
The latter, I think. Each call to GetUniqueIdentifier() returns an ID number that has not previously been returned to any previous caller of that function in the lifetime of this process, regardless of the layer name that is passed in to the function. The layer name string passed into that function seems entirely pointless to me, as things are presently implemented. All layers could pass the same string, and it would behave no differently than if they pass unique strings, as things now work. > I agree that the query-or-set operation would need to be atomic. > However, do you propose using the modified GetUniqueIdentifier() to do > that? Yes. > It it not quite clear to me how this would make it atomic, if > this is what you mean. The function is already atomic. It simply doesn't look for duplicates while holding the lock. I propose to change that. > > Regards, > Peter > _______________________________________________ dev-tech-crypto mailing list [email protected] https://lists.mozilla.org/listinfo/dev-tech-crypto
