Re: [cdt-dev] Implicit Destructor Names (AST)

[Doug Schaefer <cdtdoug@xxxxxxxxx>]

That should be defined in the C++ spec, no?

On Tue, Sep 14, 2010 at 12:53 PM, Alena Laskavaia
<elaskavaia.cdt@xxxxxxxxx> wrote:
> And how do I suppose to know the order?
>
> On Tue, Sep 14, 2010 at 12:51 PM, Schorn, Markus
> <Markus.Schorn@xxxxxxxxxxxxx> wrote:
>> The function references in the AST will not occur in the order they are
>> actually executed. A simple example is 'C c= f();', where the function call
>> is executed before the call to the constructor. The same holds basically for
>> any overloaded binary operator.
>> Therefore I don't see the requirement that the implicit names in the ast have
>> to reflect the order of execution.
>> Markus.
>>
>>
>>> -----Original Message-----
>>> From: laskava@xxxxxxxxx [mailto:laskava@xxxxxxxxx] On Behalf
>>> Of Alena Laskavaia
>>> Sent: Tuesday, September 14, 2010 6:42 PM
>>> To: Schorn, Markus
>>> Cc: CDT General developers list.
>>> Subject: Re: [cdt-dev] Implicit Destructor Names (AST)
>>>
>>> Asides from that is intention to add just definition of the
>>> destructor or the implicit call also?
>>> Implicit call should be attached to the end of the block if
>>> case we are building control flow, it should not be at the definition.
>>>
>>> On Tue, Sep 14, 2010 at 12:31 PM, Schorn, Markus
>>> <Markus.Schorn@xxxxxxxxxxxxx> wrote:
>>> > The destructors will be called as a result of returning
>>> from main and
>>> > as a result of calling std::exit.
>>> > Markus.
>>> >
>>> >> -----Original Message-----
>>> >> From: cdt-dev-bounces@xxxxxxxxxxx
>>> >> [mailto:cdt-dev-bounces@xxxxxxxxxxx] On Behalf Of Alena Laskavaia
>>> >> Sent: Tuesday, September 14, 2010 6:30 PM
>>> >> To: Doug Schaefer
>>> >> Cc: CDT General developers list.
>>> >> Subject: Re: [cdt-dev] Implicit Destructor Names (AST)
>>> >>
>>> >> Contructors are, destructors are not called (or called atexit?)
>>> >>
>>> >> On Tue, Sep 14, 2010 at 12:04 PM, Doug Schaefer
>>> <cdtdoug@xxxxxxxxx>
>>> >> wrote:
>>> >> > I thought for globals they were called at the global
>>> >> constructor time
>>> >> > before main?
>>> >> >
>>> >> > On Tue, Sep 14, 2010 at 11:58 AM, Alena Laskavaia
>>> >> > <elaskavaia.cdt@xxxxxxxxx> wrote:
>>> >> >> Disadvantage is it is good to know when destructor is
>>> >> actually called
>>> >> >> for static analysis.
>>> >> >> For globals it never actually called...
>>> >> >> Are we adding definition only or implicit call into the AST?
>>> >> >>
>>> >> >> On Tue, Sep 14, 2010 at 11:53 AM, Schorn, Markus
>>> >> >> <Markus.Schorn@xxxxxxxxxxxxx> wrote:
>>> >> >>> My first thought was, that they should be attached to the
>>> >> compound
>>> >> >>> statement.
>>> >> >>> However, that does not work for global variables.
>>> Therefore, for
>>> >> >>> simplicity I suggest to add them to the declarator. I
>>> don't see a
>>> >> >>> disadvantage with this approach.
>>> >> >>>
>>> >> >>> Markus.
>>> >> >>>
>>> >> >>>> -----Original Message-----
>>> >> >>>> From: cdt-dev-bounces@xxxxxxxxxxx
>>> >> >>>> [mailto:cdt-dev-bounces@xxxxxxxxxxx] On Behalf Of Lukas Felber
>>> >> >>>> Sent: Tuesday, September 14, 2010 11:29 AM
>>> >> >>>> To: CDT General developers list.
>>> >> >>>> Subject: [cdt-dev] Implicit Destructor Names (AST)
>>> >> >>>> Importance: Low
>>> >> >>>>
>>> >> >>>> Hi
>>> >> >>>>
>>> >> >>>> Since CDT 7.0 (I think), in code statements like 'X x;'
>>> >> >>>> (where X is a type containing a constructor declaration) the
>>> >> >>>> declarator x contains an implicit name which refers to X's
>>> >> >>>> constructor. This is a very useful and also necessary feature!
>>> >> >>>>
>>> >> >>>> However, when considering the following code {
>>> >> >>>>   X x;
>>> >> >>>> }
>>> >> >>>> one misses, beside the reference to the constructor
>>> of X, also a
>>> >> >>>> reference to X's destructor, which, for completeness,
>>> >> should also
>>> >> >>>> be part of the AST.
>>> >> >>>>
>>> >> >>>> The question which arises here, is, where this name
>>> >> should best be
>>> >> >>>> contained.
>>> >> >>>>  - The simples, but also a bit incorrect option is to
>>> >> put the name
>>> >> >>>> into the declarator x. This is not really the place where the
>>> >> >>>> destructor is called, but it is the origin which causes the
>>> >> >>>> destructor to be called at the end of the enclosing compound
>>> >> >>>> statement
>>> >> >>>>  - The more precise place to place the name is the end
>>> >> of the end
>>> >> >>>> of the enclosing compound statement (which would mean
>>> to let the
>>> >> >>>> interface IASTCompoundStatement extend
>>> IASTImpliciteNameOwner).
>>> >> >>>>
>>> >> >>>> What is your opinion on this matter?
>>> >> >>>>
>>> >> >>>>
>>> >> >>>> Best regards
>>> >> >>>> Lukas Felber
>>> >> >>>>
>>> >> >>>> _______________________________________________
>>> >> >>>> cdt-dev mailing list
>>> >> >>>> cdt-dev@xxxxxxxxxxx
>>> >> >>>> https://dev.eclipse.org/mailman/listinfo/cdt-dev
>>> >> >>>>
>>> >> >>> _______________________________________________
>>> >> >>> cdt-dev mailing list
>>> >> >>> cdt-dev@xxxxxxxxxxx
>>> >> >>> https://dev.eclipse.org/mailman/listinfo/cdt-dev
>>> >> >>>
>>> >> >> _______________________________________________
>>> >> >> cdt-dev mailing list
>>> >> >> cdt-dev@xxxxxxxxxxx
>>> >> >> https://dev.eclipse.org/mailman/listinfo/cdt-dev
>>> >> >>
>>> >> >
>>> >> _______________________________________________
>>> >> cdt-dev mailing list
>>> >> cdt-dev@xxxxxxxxxxx
>>> >> https://dev.eclipse.org/mailman/listinfo/cdt-dev
>>> >>
>>> >
>>>
>> _______________________________________________
>> cdt-dev mailing list
>> cdt-dev@xxxxxxxxxxx
>> https://dev.eclipse.org/mailman/listinfo/cdt-dev
>>
> _______________________________________________
> cdt-dev mailing list
> cdt-dev@xxxxxxxxxxx
> https://dev.eclipse.org/mailman/listinfo/cdt-dev
>