When to expect bugs fixed?

When to expect bugs fixed?

[Victor Porton]

I reported some Ada compiler bugs in the GCC bugzilla. Will these bugs fixed 
only after the appearance of GNAT 2015, or is there a hope that they will be 
fixed sooner?

And when to expect GNAT GPL 2015?

I work on a non-commercial open source project, but my work is paused now 
due GNAT bugs.

-- 
Victor Porton - http://portonvictor.org

Re: When to expect bugs fixed?

[Ludovic Brenta]

Victor Porton writes on comp.lang.ada:
> I reported some Ada compiler bugs in the GCC bugzilla. Will these bugs
> fixed only after the appearance of GNAT 2015, or is there a hope that
> they will be fixed sooner?

If past experience can serve as a guideline, these bugs will be fixed
when AdaCore has no more pressing business to do, or when some unpaid
volunteer dives into them and finds a solution.  The fixes will appear
first in the Subversion repository of FSF GCC, in the trunk, and
possibly they will be backported into a release branch (e.g. GCC 4.9 or
5.0).

> And when to expect GNAT GPL 2015?

From past experience, around May or June 2015.  But I don't speak for
AdaCore and I cannot guarantee there will even be a GNAT GPL 2015.

> I work on a non-commercial open source project, but my work is paused
> now due GNAT bugs.

I suggest you try to work around these bugs and not wait.  I've been
working on mission-critical software for years now, with a support
contract, but we've never been blocked by compiler bugs.  There is
always a way around, sometimes a temporary one until the bug is fixed.

-- 
Ludovic Brenta.

Re: When to expect bugs fixed?

[Victor Porton]

Ludovic Brenta wrote:

> Victor Porton writes on comp.lang.ada:
>> I reported some Ada compiler bugs in the GCC bugzilla. Will these bugs
>> fixed only after the appearance of GNAT 2015, or is there a hope that
>> they will be fixed sooner?

I expect this be fixed no later than in GNAT GPL 2015, because it would be 
damaging for AdaCore reputation to release a compiler with known internal 
compiler error.

> If past experience can serve as a guideline, these bugs will be fixed
> when AdaCore has no more pressing business to do, or when some unpaid
> volunteer dives into them and finds a solution.  The fixes will appear
> first in the Subversion repository of FSF GCC, in the trunk, and
> possibly they will be backported into a release branch (e.g. GCC 4.9 or
> 5.0).
> 
>> And when to expect GNAT GPL 2015?
> 
> From past experience, around May or June 2015.  But I don't speak for
> AdaCore and I cannot guarantee there will even be a GNAT GPL 2015.
> 
>> I work on a non-commercial open source project, but my work is paused
>> now due GNAT bugs.
> 
> I suggest you try to work around these bugs and not wait.  I've been
> working on mission-critical software for years now, with a support
> contract, but we've never been blocked by compiler bugs.  There is
> always a way around, sometimes a temporary one until the bug is fixed.

-- 
Victor Porton - http://portonvictor.org

Re: When to expect bugs fixed?

[Anh Vo]

On Saturday, November 1, 2014 10:55:18 AM UTC-7, Victor Porton wrote:
> Ludovic Brenta wrote:
> 
> > Victor Porton writes on comp.lang.ada:
> >> I reported some Ada compiler bugs in the GCC bugzilla. Will these bugs
> >> fixed only after the appearance of GNAT 2015, or is there a hope that
> >> they will be fixed sooner?
> 
> I expect this be fixed no later than in GNAT GPL 2015, because it would be 
> damaging for AdaCore reputation to release a compiler with known internal 
> compiler error.
> 
Based on my experience, it will be fixed within couple days or less if reported by a customer. Thus, I do not think reputation of the vendor will be damage at all.

Anh Vo

Re: When to expect bugs fixed?

[Simon Wright]

Ludovic Brenta <ludovic@ludovic-brenta.org> writes:

> Victor Porton writes on comp.lang.ada:
>> I reported some Ada compiler bugs in the GCC bugzilla. Will these bugs
>> fixed only after the appearance of GNAT 2015, or is there a hope that
>> they will be fixed sooner?

>> I work on a non-commercial open source project, but my work is paused
>> now due GNAT bugs.
>
> I suggest you try to work around these bugs and not wait.  I've been
> working on mission-critical software for years now, with a support
> contract, but we've never been blocked by compiler bugs.  There is
> always a way around, sometimes a temporary one until the bug is fixed.

As a for-instance, you can work round PR62235 by changing

========================================================================
generic
   type Record_Type; -- It should be an untagged record for C compatibility
package RDF.Auxilary.Handled_Record is

   type Access_Type is access Record_Type
      with Convention=>C;
========================================================================
by
========================================================================
generic
   type Record_Type; -- It should be an untagged record for C compatibility
   type Record_Type_Access is access Record_Type; -- should have convention C
package RDF.Auxilary.Handled_Record is

   subtype Access_Type is Record_Type_Access;
========================================================================

(I guess the new formal_incomplete_type_declaration stuff is a bit
fragile)

Re: When to expect bugs fixed?

[Simon Wright]

Simon Wright <simon@pushface.org> writes:

> generic
>    type Record_Type; -- It should be an untagged record for C compatibility
>    type Record_Type_Access is access Record_Type; -- should have convention C
> package RDF.Auxilary.Handled_Record is
>
>    subtype Access_Type is Record_Type_Access;

Or even

========================================================================
generic
   type Record_Type; -- It should be an untagged record for C compatibility
   type Access_Type is access Record_Type; -- should have convention C
package RDF.Auxilary.Handled_Record is
========================================================================

Re: When to expect bugs fixed?

[Victor Porton]

Simon Wright wrote:

> Simon Wright <simon@pushface.org> writes:
> 
>> generic
>>    type Record_Type; -- It should be an untagged record for C
>>    compatibility type Record_Type_Access is access Record_Type; -- should
>>    have convention C
>> package RDF.Auxilary.Handled_Record is
>>
>>    subtype Access_Type is Record_Type_Access;
> 
> Or even
> 
> ========================================================================
> generic
>    type Record_Type; -- It should be an untagged record for C 
compatibility
>    type Access_Type is access Record_Type; -- should have convention C
> package RDF.Auxilary.Handled_Record is> 
========================================================================

I think, this is wrong: It tempts to use Access_Type which is NOT convention 
C by default!

I'll better wait for a bugfix GNAT release. I have other business not only 
writing in Ada. So I can keep myself busy.

-- 
Victor Porton - http://portonvictor.org

Re: When to expect bugs fixed?

[Simon Wright]

Victor Porton <porton@narod.ru> writes:

> Simon Wright wrote:
>
>> Simon Wright <simon@pushface.org> writes:
>> 
>>> generic
>>>    type Record_Type; -- It should be an untagged record for C
>>>    compatibility type Record_Type_Access is access Record_Type; -- should
>>>    have convention C
>>> package RDF.Auxilary.Handled_Record is
>>>
>>>    subtype Access_Type is Record_Type_Access;
>> 
>> Or even
>> 
>> ========================================================================
>> generic
>>    type Record_Type; -- It should be an untagged record for C 
> compatibility
>>    type Access_Type is access Record_Type; -- should have convention C
>> package RDF.Auxilary.Handled_Record is> 
> ========================================================================
>
> I think, this is wrong: It tempts to use Access_Type which is NOT
> convention C by default!

I would have been (only very slightly) less inclined to suggest this if
you had not already said (in the line above) that Record_Type "should be
an untagged record for C". But it's a *workround*!!

Re: When to expect bugs fixed?

[Tero Koskinen]

1.11.2014 19:24, Victor Porton wrote:
> I reported some Ada compiler bugs in the GCC bugzilla. Will these 
> bugs fixed only after the appearance of GNAT 2015, or is there a hope
> that they will be fixed sooner?

I have had Ada/GNAT bugs open in GCC bugzilla since 2009 [1,2].
So don't expect the bugs to be fixed anytime soon.

If you want to them to be fixed fast(er), report the bugs directly
to Adacore. Once the bugs are fixed in AdaCore's internal GNAT versions,
it takes 0..2 years before stable FSF GCC/GNAT gets the fixes.

Yours,
 Tero

[1] https://gcc.gnu.org/bugzilla/show_bug.cgi?id=38882
[2] https://gcc.gnu.org/bugzilla/show_bug.cgi?id=50502

Re: When to expect bugs fixed?

[Luke A. Guest]

Tero Koskinen <tero.koskinen@iki.fi> wrote:
> 1.11.2014 19:24, Victor Porton wrote:
>> I reported some Ada compiler bugs in the GCC bugzilla. Will these 
>> bugs fixed only after the appearance of GNAT 2015, or is there a hope
>> that they will be fixed sooner?
> 
> I have had Ada/GNAT bugs open in GCC bugzilla since 2009 [1,2].
> So don't expect the bugs to be fixed anytime soon.
> 
> If you want to them to be fixed fast(er), report the bugs directly
> to Adacore. Once the bugs are fixed in AdaCore's internal GNAT versions,
> it takes 0..2 years before stable FSF GCC/GNAT gets the fixes.

This is another reason why Ada programming sucks sometimes. AdaCore should
use the FSF's bug tracker a swell as their own or every time a bug is
submitted to the gnat front end the Fsf bug tracker should send it to
AdaCore. 5+ years for a bug to be sitting around isn't good.

Luke