[nem-bug] [Nemerle 0000619]: Losing type coersion between short and int

[feedback at nemerle.org]

The following issue requires your FEEDBACK.
======================================================================
<http://nemerle.org/bugs/view.php?id=619> 
======================================================================
Reported By:                kanru
Assigned To:                
======================================================================
Project:                    Nemerle
Issue ID:                   619
Category:                   Compiler (type engine)
Reproducibility:            always
Severity:                   minor
Priority:                   normal
Status:                     feedback
======================================================================
Date Submitted:             02-15-2006 07:27 CET
Last Modified:              02-15-2006 19:53 CET
======================================================================
Summary:                    Losing type coersion between short and int
Description: 
mutable s : short = 0;
s = s + 2;
s += 2;
mutable u : byte = 0;
u = u + 2;
u += 2;

Compiling above code got:
t.n:2:1:2:10: error: expected System.Int16, got int in assigned value:
System.Int32 is not a subtype of System.Int16 [simple require]
t.n:3:1:3:7: error: expected System.Int16, got int in assigned value:
System.Int32 is not a subtype of System.Int16 [simple require]
t.n:5:1:5:10: error: expected System.Byte, got int in assigned value:
System.Int32 is not a subtype of System.Byte [simple require]
t.n:6:1:6:7: error: expected System.Byte, got int in assigned value:
System.Int32 is not a subtype of System.Byte [simple require]

Seems the compiler always treats number_literal as Int32.
======================================================================

----------------------------------------------------------------------
 malekith - 02-15-06 19:53 
----------------------------------------------------------------------
The uint testcase seems like a bug (it also shouldn't work).

As for the other testcases -- C# doesn't allow x = x + 2 too, but it
allows x += 2 for some strange reason. It is to be reconsidered what
behavior we want. But I don't see a reason x += 2 should be allowed and x
= x + 2 not.



Issue History
Date Modified  Username       Field                    Change              
======================================================================
02-15-06 07:27 kanru          New Issue                                    
02-15-06 19:53 malekith       Note Added: 0001086                          
02-15-06 19:53 malekith       Status                   new => confirmed    
02-15-06 19:53 malekith       Description Updated                          
02-15-06 19:53 malekith       Additional Information Updated                    
02-15-06 19:53 malekith       Status                   confirmed => feedback
02-15-06 19:53 malekith       Description Updated                          
02-15-06 19:53 malekith       Additional Information Updated                    
======================================================================