RE: [External] : Re: Learning Edebug

[Drew Adams <drew.adams@oracle.com>]

> instrumenting code is basically what you've been 
> doing with print statements. When edebug 
> instruments the code and inserts its own debugging
> snippets, you can execute the code normally and
> the instrumentation that edebug inserted will call 
> the edebug functionality.

In Elisp the equivalent of a print statement
is a `message' call:
  (message "foo value at position 42: %S" foo)

You can insert calls to `message' in your code 
anywhere, to provide trace-like behavior and
not invoke any debugger.

You can also invoke the (classic) Lisp debugger
by inserting breakpoints in your code anywhere.
These are just calls to function `debug'.

In the debugger, you can use the key `e' to
evaluate any Lisp expression _within the
context of the stack_, i.e., the current
control context.

As a shortcut, you can insert a `debug'
breakpoint that immediately evaluates a sexp
you pass to function `debug' as an argument,
saving you from typing it and hitting `e'.
If you just immediately hit `c', to continue
(exiting) the debugger that was just entered
(or `q' to quit), then This is akin to using 
`message' (or `error').

You pass the sexp to evaluate as the second
arg to `debug', using `nil' as the first:

  (debug nil (some-expression))

These are _ad hoc_ ways to get info at
particular places.  Easy to control (on/off
for particular breakpoints by commenting out 
their `debug' calls).