bug#11371: [Patch] fix silent-rules option

[Stefano Lattarini]

tags 11371 wontfix
thanks

Hi Eric, Marc-Antoine.

On 04/28/2012 07:48 PM, Eric Blake wrote:
> On 04/28/2012 11:25 AM, Marc-Antoine Perennou wrote:
>> The AM_SILENT_RULES macro needs an argument to be enabled.
>> Currently, if the silent-rules option is set in AM_INIT_AUTOMAKE, the
>> action is an AC_REQUIRE on the macro, which runs it without any arg,
>> resulting in an useless action.
>> The only way for now to enable silent-rules is to call
>> AM_SILENT_RULES([yes]), the silent-rules option has no effect at all.
>> This patch solves it, calling the macro with the right arg
> 
> NACK.  This would change the default behavior.  The 'silent-rules'
> listing in AM_INIT_AUTOMAKE is only to specify whether silent-rule
> support is turned on
>
Exactly.

> (and with a recent patch, this support is _always_ turned on now
> that we have a solution that works with POSIX make),
>
This is incorrect: such a change has been applied to Automake-NG only
(but if someone thinks it would be worth to "backport" such a change to
mainstream automake, I'll gladly accept a patch).

> and NOT whether it defaults to silent or verbose.
>
This is correct, and an important point to understand.

> AM_SILENT_RULES is the only documented way of changing the default
> between silent and verbose.
> 
>>  ])
>> -_AM_IF_OPTION([silent-rules], [AC_REQUIRE([AM_SILENT_RULES])])dnl
>> +_AM_IF_OPTION([silent-rules], [AM_SILENT_RULES([yes]))dnl
> 
> But this patch would change things so that the default is now verbose if
> you list silent-rules in AM_INIT_AUTOMAKE, which is not appropriate (at
> least, not without a NEWS entry and tests to validate the new behavior).
> 
Eric analysis is right, we shouldn't change the default with any valid
reason, so I'm marking this bug as "wontfix".

Marc-Antoine, in case you proposal has stemmed from a misunderstanding
of the documentation, maybe caused by poor wording or lack of clarity,
feel free to point out the problematic passages, so that we can improve
our manual and avoid further misunderstandings.

Thanks,
  Stefano