Bug 6797 – Fake changes to enum array

Status
RESOLVED
Resolution
INVALID
Severity
normal
Priority
P3
Component
dmd
Product
D
Version
D2
Platform
x86
OS
Windows
Creation time
2011-10-08T19:17:03Z
Last change time
2023-02-11T07:52:56Z
Keywords
accepts-invalid
Assigned to
No Owner
Creator
bearophile_hugs

Comments

Comment #0 by bearophile_hugs — 2011-10-08T19:17:03Z
Maybe this was reported already, because it looks like a basic bug. This compiles and runs with no errors (DMD 2.056head): void main() { enum int[][] array = [[0, 1], [2, 3]]; foreach (r; array) r[0] *= 10; foreach (ref r; array) r[0] *= 10; assert(array == [[0, 1], [2, 3]]); } This is a big hole. I think the implementation of compile-time constants need a lot of debugging.
Comment #1 by lovelydear — 2012-05-01T15:08:48Z
This simpler case actually fails: void main() { enum int[] array = [2, 1]; foreach (r; array){ r *= 10; // should be rejected writeln(array); } foreach (ref r; array){ r *= 10; // should be rejected writeln(array); } assert(array == [20, 10]); } Changing enum to immutable gives the following compilation errors: bug.d(8): Error: variable bug.main.r cannot modify immutable bug.d(12): Error: variable bug.main.r cannot modify const But this void main() { enum int[] array = [2, 1]; array[0] = 5; } is correctly caught by the compiler: bug.d(7): Error: constant [2,1][0] is not an lvalue
Comment #2 by andrej.mitrovich — 2012-12-02T10:33:38Z
There is no "hole" here, what you get here are separate arrays which are allocated for you wherever you use them. For example: import std.stdio; void main() { enum int[] array = [1, 2]; foreach (i, _; array) writeln(&array[i]); foreach (i, ref _; array) writeln(&array[i]); } Prints: 972FE0 972FD4 972FB0 972FA4 Those are two separate arrays.