D issues are now
tracked on GitHub. This Bugzilla instance remains as a read-only archive.
Summary: |
Docs about order of execution of invariant and pre/post conditions |
Product: |
D
|
Reporter: |
bearophile_hugs |
Component: |
dlang.org | Assignee: |
No Owner <nobody> |
Status: |
RESOLVED
FIXED
|
Severity: |
normal
|
CC: |
bugzilla
|
Priority: |
P2
|
Keywords: |
spec |
Version: |
D2 |
Hardware: |
All |
OS: |
All |
This D2 docs page states:
http://www.digitalmars.com/d/2.0/class.html#invariants
> The invariant is checked when a class constructor completes, at the start of
> the class destructor, before a public or exported member is run, and after a
> public or exported function finishes.
If preconditions, postconditions and invariants are pure functions then they need to refuse the usage of functions like printf() inside them. If they aren't pure (maybe because they must support printing and logging), then printing and logging is a side effect, and their order of execution counts.
Even if preconditions, postconditions and invariants are pure functions, their purpose is to stop the execution of the program when there's a bug, so during debugging it's important to know the order of execuction of those contracts, to understand where the bug is.
Either way, the order of execution of preconditions, postconditions and invariants is important and it needs to be stated in the D documentation (and it needs to be the same in different D implementations).
So I suggest to change the D docs like this:
> The invariant is checked when a class constructor completes, at the start of
> the class destructor, before a public or exported member is run (before its precondition),
> and after a public or exported function finishes (after its postcondition).
Comment 1
bearophile_hugs
2010年10月09日 06:36:54 UTC
See bug 5024
Comment 2
Walter Bright
2012年01月19日 13:57:27 UTC
Fixed, but the order is preconditions, invariant, body, invariant, postconditions.
Comment 3
Walter Bright
2012年01月19日 13:58:09 UTC
*** Issue 5024 has been marked as a duplicate of this issue. ***