Nick Ing-Simmons <nick@ing-simmons.net> wrote > Treating a tied ARRAY as a hash should trigger the pseudo hash code. > (I gave this a passing thought when doing TIEARRAY fixes.) But that doesn't work, as my original example showed. The problem is that TIEARRAY gets called instead of TIEHASH. What I presume is happening, without having done much research, is that when you do tie %$avhv, ... the % dereference correctly understands about the pseudo-hash case, and stacks the AV instead of an HV. But it leaves no indication that this reference should "really" be thought of as a hash. So the tie code just tests the reference type and calls TIEARRAY. > Likewise a tied HASH should be usable as the index hash. > Now whether either _works_ is a different story - anyone want to > write a t/*.t type testcase? The latter already has a test case in t/op/avhv.t, presumably by yours truly. I only discovered this after writing my own test. :-( Mike GuyThread Previous | Thread Next