Debugging and the execution path

Dick Kriesel dick.kriesel at mail.com
Wed Dec 21 13:52:47 EST 2005


On 12/21/05 9:30 AM, "Jerry Daniels" <jerry at daniels-mara.com> wrote:

> Assuming you could have a calling path in your debug environs, what
> would it look like?

IMHO, since objects of the same type can have the same name, and objects in
different stacks can have the same abbreviated id, we need the calling path
to identify long ids.  We need to toggle the display of the calling path to
show objects' names and/or ids.  We need to see handler names.  Since within
a script functions and commands can have the same name, and getProps and
setProps can have the same name, we need a handler type, or a line number.
Since one handler can invoke another from multiple lines, we need a line
number.  Since a line can have multiple statements delimited by semicolons,
and since a statement can invoke a handler more than once, we need a
character offset within the line.

Since execution can depend on the target, not just on the calling path, we
need to know the target (during debugging, typing "put the target" into the
message box does not respect the executionContexts).

So it seems to me the minimum entry for each call is the long id of me, the
character offset in the script of me, and the long id of the target. A
handler which displays the path to the scripter can determine object names,
handler names, and handler types.

-- Dick





More information about the Use-livecode mailing list