hyc@symas.com wrote:
In reviewing the patch, I see several small details that need to be fixed. E.g., you added a config keyword but re-used an existing OID for your new attribute. OIDs must never be re-used, and certainly you must not use two different attributes with the same OID.
Oh never mind, I see the confusion. You're just defining a new 2nd argument to the existing keyword.
You use strlen/strncat, which are heavily discouraged in our code. (We never use strcat/strncat; we use strlen only when it's unavoidable.) (Some uses of strlen already existed in constraint.c; we just didn't get around to eliminating them.) In general we use bervals instead of flat strings, and there's almost never a reason to count string lengths at runtime.
I got 3 rejects when applying the patch to my source tree (current HEAD).
I think these points can all be cleaned up pretty easily. Let me know if you need any particular pointers.
I'm working on getting this to build now.