"John Machin" <[EMAIL PROTECTED]> wrote: 8<-------------------
> > data = struct.unpack("!H4BH20BHI", strMessage) > > > > (top, ip1, ip2, ip3, ip4, messageCounter, ackRequired, > > dataType, utc1, utc2, utc3, utc4, utc5, utc6, utc7, utc8, > > utc9, utc10, utc11, utc12, st1, st2, st3, st4, st5, st6, > > numberOfLabels, dataWord) = data > > > > Those utc1, ..., utc12 etc inspire a suggestion: a mild addition to the > syntax to allow specifying that the following repeat count should be > interpreted as the expected dimension of a tuple (unpack) or any > indexable object with a __len__ method (pack): > > (top, ip, messageCounter, ackRequired, > dataType, utc, st, numberOfLabels, dataWord, > ) = struct.unpack("! H /4B H B B /12B /6B H I", strMessage) > > Apropos of a recent thread: I don't need pointing at the suggestion box > and the patch submission gizmoid on Sourceforge -- I'd be happy to do a > patch, just wondering if anybody's interested in having that, besides > me and Pat Malone :-) just a comment - it took me a while to figure out what you were on about. The relationship between the sources and targets are not immediately obvious (not to me, at least) I wonder if it would not be better to leave the struct stuff alone, and to "dimension" the left hand side - something like ip(4) and utc(12) ? or ip*4, or whatever... - Hendrik -- http://mail.python.org/mailman/listinfo/python-list