PR# 14018 Evaluation of create {SPECIAL [INTEGER]}.make (10) does not yield right thing in the debugger

Problem Report Summary
Submitter: manus_eiffel
Category: Debugger
Priority: Medium
Date: 2008/02/22
Class: Bug
Severity: Serious
Number: 14018
Release: 6.2.7.2407
Confidential: No
Status: Analyzed
Responsible: manus_eiffel
Environment: Mozilla/5.0 (Windows; U; Windows NT 5.2; en-US; rv:1.8.1.12) Gecko/20080201 Firefox/2.0.0.12
Synopsis: Evaluation of create {SPECIAL [INTEGER]}.make (10) does not yield right thing in the debugger

Description
Evaluating `create {SPECIAL [INTEGER]}.make (12)' I was expecting to expand the area and see zeros in the value column, but instead I see `Void' and NONE for type.

This is classic.
To Reproduce

										
Problem Report Interactions
From:jfiat_es    Date:2009/09/24    Download   
1) and 2) were due to another bug in runtime
fixed by rev#80855

And 3) remains an issue.

from runtime, since INTERNAL has the same trouble


From:alexk_es    Date:2009/07/21    Download   
Depending on creation expression I get different kind of behaviour in rev#79743:
1. create {ANY}
   EiffelStudio error: The Eiffel debugger daemon is dead,
   If you were debugging, the session is about to be terminated.
2. create {ARRAY [INTEGER]}.make (1, 2) or
   create STRING_8.make_empty or
   <<1, 2, 3>>
   [1, 2, 3]
   EiffelStudio silently disappears
3. create {SPECIAL [INTEGER]}.make_filled (5, 5)
   Can not instanciate type {SPECIAL [INTEGER_32]} : {SPECIAL} is not yet supported. - Evaluation of creation expression for this type is not supported.

But
   create {SPECIAL [INTEGER]}.make_empty (3)
shows some data in the watch tab, though there are no associated items (that might be expected - I do not know).

From:jfiat_es    Date:2008/04/11    Download   
Any new comment on this issue ?
I mean any solution from the runtime/INTERNAL ?

From:jfiat_es    Date:2008/02/25    Download   
The code violates a precondition.
For expression evaluation, the assertions are disabled, so no violation.

Any idea to create new instance of SPECIAL [INTEGER] for instance?

From:manus_eiffel    Date:2008/02/25    Download   
Yes, this only creates SPECIAL of ANY, not SPECIAL of expanded. I'm surprised that no precondition is violated. Does the code works without precondition violation?

From:jfiat_es    Date:2008/02/25    Status: Analyzed    Download   
The debugger calls more or less the following code


local
    int: INTERNAL
    a: ANY
do
    create int
    i := int.dynamic_type_from_string ("SPECIAL [INTEGER_32]")
    a := int.new_special_any_instance (i, 12)

So it seems, we would need a new_special_XYZ_instance (...)

any idea ?