Yitzchak Scott-Thoennes wrote: > My memory is that they used the READONLY flag on the hashes to > indicate a restricted hash only after checking that it had no current > meaning. If that memory is correct, setting it on the HVs in 5.6.1 > was a noop. Did you find that it had any effect? I think it had an effect if you applied it to the hash elements, but I don't think it ever did anything to the hash itself. From looking through the Hash::Utils documentation, I think I probably *do* want to use restricted hashes, as I'm representing a tree of read-only C structs as a tree of hashes, which is why I was fiddling around with SvREADONLY_on in the first place. However, I'm not aware of any documentation on how this should be done from XSUB code. I can always look at how Hash::Utils does it, but that's implemented (mainly) in perl rather than XSUB - I was hoping for some existing XSUB code I could plagarise ;-) -- Alan Burlison --Thread Previous | Thread Next