I updated the code tree and regression tests to use the new syntax for
path and polygon geometric types. This will require a table update when
upgrading from pre-v6.1 databases, but I think this is preferable to
having two syntax styles lurking in the interpreter. From a previous
message to Hal Snyder:
>... One problem with retaining the ability to read the old notation is
>that, for path and polygon, the interpretation of the input varies
>wildly depending on how many parens you include in your string. For
>example,
>
>(for paths)
>'1,2,3,4,5,6' is interpreted as '((1,2),(3,4),(5,6))'
>'(1,2,3,4,5,6)' is interpreted as '((3,4),(5,6))'
>'((1,2),(3,4),(5,6))' is interpreted as in the first case
>
>(for polygons)
>'1,2,3,4' is interpreted as '((1,2),(3,4))'
>'(1,2,3,4)' is interpreted as '((1,3),(2,4))'
>'((1,2),(3,4))' is interpreted as in the first case
The "new syntax" gives a consistant interpretation to all of the above
input variations.
Hal was the only one who responded to my query a couple of weeks ago
regarding this topic, and he concurred on the change (at least on second
thought :)
I think the downside of having to update a few existing databases (not
everyone uses the geometric types) is far outweighed by not having
syntax variations to cope with in the future.
I don't expect these changes to cause big troubles, and I will make sure
that any unanticipated problems are fixed immediately.
I had wanted to put this in earlier, but did not have the time.
- Tom
"idle hands are the devil's playground"...
------------------------------
