Re: Dependency tracking

[Akim Demaille]

Hi Eric,

I have not seen any reaction about this.  Maybe you think I'm
asking you to update the patch and apply it; I don't.  As a
matter of fact, maybe the disclaimers have not been signed, I
don't know.

I was asking why this was not integrated.  I can ask the authors
to sign the disclaimer, and I can update their contribution to
today's sources.

WDYT?

Merry Christmas!

        Akim

> Le 6 déc. 2018 à 05:52, Akim Demaille <address@hidden> a écrit :
> 
> Hi!
> 
> Any reaction about the following message?
> 
> Cheers!
> 
>       Akim
> 
>> Le 27 nov. 2018 à 05:56, Akim Demaille <address@hidden> a écrit :
>> 
>> Hi all!
>> 
>> Bison's backends are implemented in (GNU) M4.  Bison is shipped
>> with default M4 files (called skeletons), which m4_include several
>> bits.  Users may also write their own skeletons.
>> 
>> I regularly have the dependencies wrong in Bison's Makefile, and
>> some tests are not re-run when they should.  Fine tuning the
>> dependencies is tedious and error-prone, so in general I use a
>> superset of the dependencies, and some tests are run more than
>> needed.  Not a big deal, but since M4 knows what files are actually
>> used (bison itself does not know what M4 files are used, it knows
>> only about the top-level M4 file, but then m4 does what it wants),
>> it would make sense for it to tell us.
>> 
>> I made a quick search on this, and found that such a feature was
>> already designed (I found this:
>> https://lists.gnu.org/archive/html/m4-patches/2010-11/msg00001.html),
>> and actually contributed
>> (https://lists.gnu.org/archive/html/m4-patches/2011-02/msg00003.html).
>> 
>> Is there a reason for this work not to be part of M4?  I cannot
>> see any further reference after 2011-02 (but maybe I failed to
>> look for the right words, namazu does not seem very reliable).
>> I can see older messages, including interactions with Eric, but
>> nothing more recent.
>> 
>> Cheers!
>