| > The idea had been suggested to me that it would be nice to have
| > the lifelines parser optionally flag variables that were used before
| > being set (in analogy to use before declaration), to help catch
| > typos that otherwise slip by as uses of new empty variables,
| Yes, this sounds like a good idea. Of course, it might be sticky to
| implement. I haven't looked at the coding, but I imagine the various
| iterators might be implemented such as to cause problems with this sort
| of thing.
Hm, I hadn't thought of them -- interesting point.
But they should be ok - they are all put directly into the current local
symbol table, I think. And sometime this year I modified the code to
remove them from the local symbol table when the loop is over (used to
be they were just left there).
I'm thinking that upon a reference to an identifier (probably
if the explicit flag is on, check that it is in either the local or global
and if not, print a warning or something.
| > but I think it would most naturally belong in the report program
| I agree.
| John Chandler
| Lifelines-dev mailing list