[RFA] Fix djgpp gdb-7.4 (and probably trunk) build failure in readline

[RFA] Fix djgpp gdb-7.4 (and probably trunk) build failure in readline

[Pierre Muller]

  Jan Kratochvil suggested me to submit my patch directly to bug-readline
mailing list.

The bug appeared while trying to compile the nw GDB release 7.4
version on a Dos DJGPP target, that does not have
SIGWINCH signal. GDB now uses readline version 6.2,
so this patch is against this version of readline library.

>>>
>>> Starting here is the original post to gdb-patches:
>>>

DJGPP has no SIGWINCH
but despite this being checked by configure,
the code in readline/signal.c
is incompatible with a system that doesn't define SIGWINCH.
I found this problem while trying to compile GDB-7.4 sources for DJGPP.
Trunk source seems to be the same...

I don't know what is the master readline repository...
I am also unsure about proper ChangeLog entry for this...

Pierre Muller
GDB pascal language maintainer

ChangeLog entry:
2012-01-27  Pierre Muller  <muller@ics.u-strasbg.fr>

        * signals.c (_rl_block_sigwinch, _rl_release_sigwinch): Add
        conditional SIGWINCH around functions.

--- readline/signals.c  2011-05-12 01:38:39.000000000 +0200
+++ readline/signals.c.new      2012-01-27 17:47:06.837454200 +0100
@@ -580,6 +580,7 @@ _rl_release_sigint ()
   sigint_blocked = 0;
 }

+#ifdef SIGWINCH
 /* Cause SIGWINCH to not be delivered until the corresponding call to
    release_sigwinch(). */
 void
@@ -627,6 +628,7 @@ _rl_release_sigwinch ()

   sigwinch_blocked = 0;
 }
+#endif /* SIGWINCH */

 /* **************************************************************** */
 /*                                                                 */

PING [RFA] Fix djgpp gdb-7.4 (and probably trunk) build failure in readline

[Pierre Muller]

  I have got no reaction on bug-readline mailing list
regarding this issue...
  What should I do next?


Pierre Muller
GDB pascal language maintainer



> -----Message d'origine-----
> De : gdb-patches-owner@sourceware.org [mailto:gdb-patches-
> owner@sourceware.org] De la part de Pierre Muller
> Envoyé : samedi 28 janvier 2012 19:15
> À : bug-readline@gnu.org
> Cc : gdb-patches@sourceware.org
> Objet : [RFA] Fix djgpp gdb-7.4 (and probably trunk) build failure in
> readline
> 
>   Jan Kratochvil suggested me to submit my patch directly to bug-readline
> mailing list.
> 
> The bug appeared while trying to compile the nw GDB release 7.4
> version on a Dos DJGPP target, that does not have
> SIGWINCH signal. GDB now uses readline version 6.2,
> so this patch is against this version of readline library.
> 
> >>>
> >>> Starting here is the original post to gdb-patches:
> >>>
> 
> DJGPP has no SIGWINCH
> but despite this being checked by configure,
> the code in readline/signal.c
> is incompatible with a system that doesn't define SIGWINCH.
> I found this problem while trying to compile GDB-7.4 sources for DJGPP.
> Trunk source seems to be the same...
> 
> I don't know what is the master readline repository...
> I am also unsure about proper ChangeLog entry for this...
> 
> Pierre Muller
> GDB pascal language maintainer
> 
> ChangeLog entry:
> 2012-01-27  Pierre Muller  <muller@ics.u-strasbg.fr>
> 
>         * signals.c (_rl_block_sigwinch, _rl_release_sigwinch): Add
>         conditional SIGWINCH around functions.
> 
> --- readline/signals.c  2011-05-12 01:38:39.000000000 +0200
> +++ readline/signals.c.new      2012-01-27 17:47:06.837454200 +0100
> @@ -580,6 +580,7 @@ _rl_release_sigint ()
>    sigint_blocked = 0;
>  }
> 
> +#ifdef SIGWINCH
>  /* Cause SIGWINCH to not be delivered until the corresponding call to
>     release_sigwinch(). */
>  void
> @@ -627,6 +628,7 @@ _rl_release_sigwinch ()
> 
>    sigwinch_blocked = 0;
>  }
> +#endif /* SIGWINCH */
> 
>  /* **************************************************************** */
>  /*                                                                 */
> 

Re: PING [RFA] Fix djgpp gdb-7.4 (and probably trunk) build failure in readline

[Pedro Alves]

On 02/24/2012 02:16 PM, Pierre Muller wrote:
>   I have got no reaction on bug-readline mailing list
> regarding this issue...
>   What should I do next?

Chet, can we get something like this into readline mainline?  We'd rather
have less local readline patches than more.  :-)

-- 
Pedro Alves

> 
> 
> Pierre Muller
> GDB pascal language maintainer
> 
> 
> 
>> -----Message d'origine-----
>> De : gdb-patches-owner@sourceware.org [mailto:gdb-patches-
>> owner@sourceware.org] De la part de Pierre Muller
>> Envoyé : samedi 28 janvier 2012 19:15
>> À : bug-readline@gnu.org
>> Cc : gdb-patches@sourceware.org
>> Objet : [RFA] Fix djgpp gdb-7.4 (and probably trunk) build failure in
>> readline
>>
>>   Jan Kratochvil suggested me to submit my patch directly to bug-readline
>> mailing list.
>>
>> The bug appeared while trying to compile the nw GDB release 7.4
>> version on a Dos DJGPP target, that does not have
>> SIGWINCH signal. GDB now uses readline version 6.2,
>> so this patch is against this version of readline library.
>>
>>>>>
>>>>> Starting here is the original post to gdb-patches:
>>>>>
>>
>> DJGPP has no SIGWINCH
>> but despite this being checked by configure,
>> the code in readline/signal.c
>> is incompatible with a system that doesn't define SIGWINCH.
>> I found this problem while trying to compile GDB-7.4 sources for DJGPP.
>> Trunk source seems to be the same...
>>
>> I don't know what is the master readline repository...
>> I am also unsure about proper ChangeLog entry for this...
>>
>> Pierre Muller
>> GDB pascal language maintainer
>>
>> ChangeLog entry:
>> 2012-01-27  Pierre Muller  <muller@ics.u-strasbg.fr>
>>
>>         * signals.c (_rl_block_sigwinch, _rl_release_sigwinch): Add
>>         conditional SIGWINCH around functions.
>>
>> --- readline/signals.c  2011-05-12 01:38:39.000000000 +0200
>> +++ readline/signals.c.new      2012-01-27 17:47:06.837454200 +0100
>> @@ -580,6 +580,7 @@ _rl_release_sigint ()
>>    sigint_blocked = 0;
>>  }
>>
>> +#ifdef SIGWINCH
>>  /* Cause SIGWINCH to not be delivered until the corresponding call to
>>     release_sigwinch(). */
>>  void
>> @@ -627,6 +628,7 @@ _rl_release_sigwinch ()
>>
>>    sigwinch_blocked = 0;
>>  }
>> +#endif /* SIGWINCH */
>>
>>  /* **************************************************************** */
>>  /*                                                                 */
>>
> 
> 

Re: [Bug-readline] PING [RFA] Fix djgpp gdb-7.4 (and probably trunk) build failure in readline

[Chet Ramey]

On 2/24/12 9:34 AM, Pedro Alves wrote:
> On 02/24/2012 02:16 PM, Pierre Muller wrote:
>>   I have got no reaction on bug-readline mailing list
>> regarding this issue...
>>   What should I do next?
> 
> Chet, can we get something like this into readline mainline?  We'd rather
> have less local readline patches than more.  :-)

The fix is already in the mainline readline source, though in a different
form.  I don't want to hold your release up waiting for the readline
release process, hence my recommendation that you move forward with the
patch in its current form.

-- 
``The lyf so short, the craft so long to lerne.'' - Chaucer
		 ``Ars longa, vita brevis'' - Hippocrates
Chet Ramey, ITS, CWRU    chet@case.edu    http://cnswww.cns.cwru.edu/~chet/

RE: [Bug-readline] PING [RFA] Fix djgpp gdb-7.4 (and probably trunk) build failure in readline

[Pierre Muller]

> -----Message d'origine-----
> De : gdb-patches-owner@sourceware.org [mailto:gdb-patches-
> owner@sourceware.org] De la part de Chet Ramey
> Envoyé : samedi 25 février 2012 20:41
> À : Pedro Alves
> Cc : Pierre Muller; bug-readline@gnu.org; gdb-patches@sourceware.org;
> chet.ramey@case.edu
> Objet : Re: [Bug-readline] PING [RFA] Fix djgpp gdb-7.4 (and probably
trunk)
> build failure in readline
> 
> On 2/24/12 9:34 AM, Pedro Alves wrote:
> > On 02/24/2012 02:16 PM, Pierre Muller wrote:
> >>   I have got no reaction on bug-readline mailing list
> >> regarding this issue...
> >>   What should I do next?
> >
> > Chet, can we get something like this into readline mainline?  We'd
rather
> > have less local readline patches than more.  :-)
> 
> The fix is already in the mainline readline source, though in a different
> form.  I don't want to hold your release up waiting for the readline
> release process, hence my recommendation that you move forward with the
> patch in its current form.

  I searched for this mainline readline source, but couldn't find the
location...
Is this accessible via cvs,svn, git or anything else,
or is it just private?
  At least the sourceware web site didn't give any link to 
such a repository.

 
Pierre Muller

PS: As I already committed my patch, it will probably not matter much,
but in other similar occasion, it might be better to
commit the changes already on mainline instead of GDB specific ones...
 
> --
> ``The lyf so short, the craft so long to lerne.'' - Chaucer
> 		 ``Ars longa, vita brevis'' - Hippocrates
> Chet Ramey, ITS, CWRU    chet@case.edu
http://cnswww.cns.cwru.edu/~chet/

Re: [Bug-readline] PING [RFA] Fix djgpp gdb-7.4 (and probably trunk) build failure in readline

[Chet Ramey]

On 2/25/12 3:07 PM, Pierre Muller wrote:

>> The fix is already in the mainline readline source, though in a different
>> form.  I don't want to hold your release up waiting for the readline
>> release process, hence my recommendation that you move forward with the
>> patch in its current form.
> 
>   I searched for this mainline readline source, but couldn't find the
> location...
> Is this accessible via cvs,svn, git or anything else,
> or is it just private?

The git repository for the distribution sources is available on savannah.
There doesn't seem to be much to be gained by maintaining a `devel' branch
for readline, since all of the source code changes are captured in the
`devel' bash branch in its savannah git tree.

I try to update the bash devel branch at least every month with the weekly
snapshots.

Chet
-- 
``The lyf so short, the craft so long to lerne.'' - Chaucer
		 ``Ars longa, vita brevis'' - Hippocrates
Chet Ramey, ITS, CWRU    chet@case.edu    http://cnswww.cns.cwru.edu/~chet/

Re: [Bug-readline] PING [RFA] Fix djgpp gdb-7.4 (and probably trunk) build failure in readline

[Chet Ramey]

On 2/24/12 9:16 AM, Pierre Muller wrote:
>   I have got no reaction on bug-readline mailing list
> regarding this issue...

The bug and fix are obvious.

>   What should I do next?

This will be changed in the next readline release.  In the meantime,
the gdb version of readline can be changed to include the fix.

Chet

-- 
``The lyf so short, the craft so long to lerne.'' - Chaucer
		 ``Ars longa, vita brevis'' - Hippocrates
Chet Ramey, ITS, CWRU    chet@case.edu    http://cnswww.cns.cwru.edu/~chet/

RE: [Bug-readline] PING [RFA] Fix djgpp gdb-7.4 (and probably trunk) build failure in readline

[Pierre Muller]

  Pedro,

  may I interpret Chet's answer as an official approval 
for that patch to current gdb cvs?
(As he is not a global GDB maintainer, I would rather get it directly from you...)
Or do you still prefer to postpone this?

Pierre

> -----Message d'origine-----
> De : gdb-patches-owner@sourceware.org [mailto:gdb-patches-
> owner@sourceware.org] De la part de Chet Ramey
> Envoyé : vendredi 24 février 2012 15:41
> À : Pierre Muller
> Cc : bug-readline@gnu.org; gdb-patches@sourceware.org; chet.ramey@case.edu
> Objet : Re: [Bug-readline] PING [RFA] Fix djgpp gdb-7.4 (and probably trunk)
> build failure in readline
> 
> On 2/24/12 9:16 AM, Pierre Muller wrote:
> >   I have got no reaction on bug-readline mailing list
> > regarding this issue...
> 
> The bug and fix are obvious.
> 
> >   What should I do next?
> 
> This will be changed in the next readline release.  In the meantime,
> the gdb version of readline can be changed to include the fix.
> 
> Chet
> 
> --
> ``The lyf so short, the craft so long to lerne.'' - Chaucer
> 		 ``Ars longa, vita brevis'' - Hippocrates
> Chet Ramey, ITS, CWRU    chet@case.edu    http://cnswww.cns.cwru.edu/~chet/

Re: [Bug-readline] PING [RFA] Fix djgpp gdb-7.4 (and probably trunk) build failure in readline

[Pedro Alves]

On 02/24/2012 03:59 PM, Pierre Muller wrote:

>   may I interpret Chet's answer as an official approval 
> for that patch to current gdb cvs?
> (As he is not a global GDB maintainer, I would rather get it directly from you...)
> Or do you still prefer to postpone this?

Go ahead and apply your patch to gdb cvs.

-- 
Pedro Alves

RE: [Bug-readline] PING [RFA] Fix djgpp gdb-7.4 (and probably trunk) build failure in readline

[Pierre Muller]

> -----Message d'origine-----
> De : gdb-patches-owner@sourceware.org [mailto:gdb-patches-
> owner@sourceware.org] De la part de Pedro Alves
> Envoyé : vendredi 24 février 2012 17:09
> À : Pierre Muller
> Cc : 'Chet Ramey'; gdb-patches@sourceware.org
> Objet : Re: [Bug-readline] PING [RFA] Fix djgpp gdb-7.4 (and probably trunk)
> build failure in readline
> 
> On 02/24/2012 03:59 PM, Pierre Muller wrote:
> 
> >   may I interpret Chet's answer as an official approval
> > for that patch to current gdb cvs?
> > (As he is not a global GDB maintainer, I would rather get it directly from
> you...)
> > Or do you still prefer to postpone this?
> 
> Go ahead and apply your patch to gdb cvs.

  OK, thanks, committed after rechecking that it didn't break compilation
on cygwin target.

Pierre