Hello,
I'm puzzled by what is going on here:
scheme@(guile-user)> (eval-string "`(foo ,b)")
ERROR: In procedure %resolve-variable:
Unbound variable: b
Entering a new prompt. Type `,bt' for a backtrace or `,q' to continue.
scheme@(guile-user) [1]> ,q
scheme@(guile-user)> (define b 'bar)
scheme@(guile-user)> (eval-string "`(foo ,b)")
$1 = (foo bar)
scheme@(guile-user)> (null-environment 5)
$2 = #<interface (#{ g298}#) 7f4447df16e0>
scheme@(guile-user)> (eval-string "`(foo ,b)" $2)
$3 = (foo (unquote b))
scheme@(guile-user)> (eval-string "`(foo ,b)" (current-module))
$4 = (foo bar)
scheme@(guile-user)> (module-ref (current-module) 'b)
$5 = bar
scheme@(guile-user)> (module-ref $2 'b)
ERROR: In procedure scm-error:
No variable named b in #<interface (#{ g298}#) 7f4447df16e0>
Entering a new prompt. Type `,bt' for a backtrace or `,q' to continue.
scheme@(guile-user) [1]> ,q
scheme@(guile-user)> (module-define! $2 'b 'bar)
scheme@(guile-user)> (module-ref $2 'b)
$6 = bar
scheme@(guile-user)> (eval-string "`(foo ,b)" $2)
$7 = (foo (unquote b))
Why does b get evaluated in the default environment (current module), but
gets unquoted for one passed as an argument to eval-string? What is
different about these environments? Is there a correct way to use a
different environment, such that the result is the same?
