You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
NB It is crucial that the predicate remains a quosure—it must remember its original scope.
As such, the error message won't be properly interpreted by the current implementation of err_invalid_input(), because that function relies on the flag dot_as_expr which is set by parse_check(), and is invisible in the output, above.
Consequently, the mechanism by which the proper scope of dot-interpolation is identified will need to be adapted to the above behavior of localize().
Solution 1: One possibility would be substitute the function of the boolean flag dot_as_expr with a special (prefix/suffix) “marker” in the error message, which is interpreted by err_invalid_input().
Solution 2: Alternatively, a simpler, but arguably more “hacky,” approach would be endow localized checkers with an additional “global name” attribute that would be interpreted by parse_check() as overriding the value of the argument msg. In this case, it would suffice for chkrs to behave accordingly:
This issue refers to the development branch.
Example:
should produce the list of two functions such that
NB It is crucial that the predicate remains a quosure—it must remember its original scope.
As such, the error message won't be properly interpreted by the current implementation of
err_invalid_input()
, because that function relies on the flagdot_as_expr
which is set byparse_check()
, and is invisible in the output, above.Consequently, the mechanism by which the proper scope of dot-interpolation is identified will need to be adapted to the above behavior of
localize()
.Solution 1: One possibility would be substitute the function of the boolean flag
dot_as_expr
with a special (prefix/suffix) “marker” in the error message, which is interpreted byerr_invalid_input()
.Solution 2: Alternatively, a simpler, but arguably more “hacky,” approach would be endow localized checkers with an additional “global name” attribute that would be interpreted by
parse_check()
as overriding the value of the argumentmsg
. In this case, it would suffice forchkrs
to behave accordingly:No change would then be required elsewhere, and the logic of
localize()
would be simplified as well, for it would not require any naming logic.The text was updated successfully, but these errors were encountered: