From: Andreas L. <no...@sb...> - 2004-09-28 16:15:34
|
On Tue, 2004-09-28 at 17:47 +0200, Eric Bezault wrote: > Andreas Leitner wrote: > > I believe there is a bug in ET_REGULAR_INTEGER_CONSTANT.compute_value. > > It is supposed to set `value' but the procedure never touches that > > attribute. I think the following should fix this bug: > > This is now fixed in CVS. Thanks for the quick response and fix. Just out of curiosity, I saw in the CVS-log that you kept `v' as a temporary variable and only at the end of the procedures `v' gets assigned to `value'. Is this a style issue or does it have performance reasons (access to an attribute should be slower in general, because it needs to be dereferenced with `Current' -- me thinks)? thanks in advance, Andreas |