On 2016-03-31 13:39, Wette, Matthew R (3441) wrote:
On Mar 29, 2016, at 6:05 PM, Wette, Matthew R (3441)
<matthew.r.we...@jpl.nasa.gov> wrote:
On Mar 29, 2016, at 10:16 AM, ra...@openmailbox.org wrote:
I think this is a bug in guile relating to "include":
a.scm: (define a 1)
b.scm: (include "a.scm") (display a) (newline)
guile -l b.scm # prints 1
now update
a.scm: (define a 2)
guile -l b.scm # still prints 1
I have noticed this behavior and have learned to live with it. Given
the following from the Guile Reference manual (2.0.11), I think it
is not considered a bug.
"On the other hand, include does have all the disadvantages of early
binding: once the code with the include is compiled, no change to
the included file is reflected in the future behavior of the
including form.”
And I will add that an implementation that checked all includes would
have to carry a tree of source dependencies for each binary .go file.
The alternative seems unattractive to me so I am happy to live with
the current implementation. — Matt
I think you are right and this can't be considered a bug, even though it
may introduce bugs in programming if used wrong.
wingo mentioned to me that LOAD can be used as an alternative to INCLUDE
if we want late binding instead of early.
Thanks!
