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
I know there has been a blog discussion on equality, and some new proposals are forthcoming to (re-)define =, == and =? (plus a fourth one perhaps), but I cannot help feeling uneasy with the simplest and most forgiving of these: equality (as in =) applied to the numerical types (number! and money!) which are similar enough to be freely convertible
see the code for observed behaviour; BTW it works with any integer number, not only 0
>>0=0.0; true (also 0.0 = 0% and 0.0 = $0)>>0!=0%; true (also 0 != $0, but 0% = $0);; thus, e.g.>>0>0%; false>>0<0%; false>>0=0%; false;; none of the three yields true, this might be awkward;; in case statements which rely on exactly one of them being true
In order to get define all possible combinations, we need a series of tests that specify those combinations. That's the only way we can get this done right.
Rebolbot commented on Jul 15, 2009:
Submitted by:meijeru
This ticket can now be closed (I could do it myself, but that would be abusing a loophole in Curecode (see ALTME posting by BrianH).
Rebolbot commented on Jul 15, 2009:
Submitted by:BrianH
Determined the build where this was fixed (it could have been 69 or 70) and marked as tested.
We finally have a policy for when to mark a ticket as complete, and for Bug tickets that is when the conditions in question have tests added to verify them in the standard test suite. Since the standard test suite isn't quite standardized yet (we're working on it), this ticket shouldn't be marked as closed. For that matter, some of the other tickets that have already been marked as closed are not appropriately so. Reviewing the tested and closed tickets is on the todo list.
There's more to reviewing than just flipping a few settings :)
Submitted by: meijeru
I know there has been a blog discussion on equality, and some new proposals are forthcoming to (re-)define =, == and =? (plus a fourth one perhaps), but I cannot help feeling uneasy with the simplest and most forgiving of these: equality (as in =) applied to the numerical types (number! and money!) which are similar enough to be freely convertible
Imported from: CureCode [ Version: alpha 66 Type: Bug Platform: All Category: Datatype Reproduce: Always Fixed-in:alpha 69 ]
Imported from: metaeducation#1046
Comments:
Submitted by: BrianH
This should be taken into account when the equality operations are revamped. Changed to a Bug.
Submitted by: meijeru
It is actuall worse than I tought:
Submitted by: meijeru
See also #952
Submitted by: Carl
In order to get define all possible combinations, we need a series of tests that specify those combinations. That's the only way we can get this done right.
Submitted by: meijeru
This ticket can now be closed (I could do it myself, but that would be abusing a loophole in Curecode (see ALTME posting by BrianH).
Submitted by: BrianH
Determined the build where this was fixed (it could have been 69 or 70) and marked as tested.
We finally have a policy for when to mark a ticket as complete, and for Bug tickets that is when the conditions in question have tests added to verify them in the standard test suite. Since the standard test suite isn't quite standardized yet (we're working on it), this ticket shouldn't be marked as closed. For that matter, some of the other tickets that have already been marked as closed are not appropriately so. Reviewing the tested and closed tickets is on the todo list.
There's more to reviewing than just flipping a few settings :)
The text was updated successfully, but these errors were encountered: