Re: [Cooker] Re: [Maintainers] A plan for automake

[Frederic Crozat]

On Wed, 18 May 2005 15:27:25 +0200, Guillaume Rousse wrote:

> Frederic Crozat wrote:
>> On Wed, 18 May 2005 13:40:28 +0200, Guillaume Rousse wrote:
>> 
>> 
>>>Frederic Crozat wrote:
>>> >>>Your plan will break a lot of package (I don't know how, but I think
>>>
>>>>>>that all old package will be broken, perhaps we can remove them, why
>>>>>>not...)
>>>>>
>>>>>A maximum of 190. Is this really too much ?
>>>>
>>>>
>>>>No, this is only for packages which are currently requiring auto* calls,
>>>>if new patches are needed for other packages, this number might increase.
>>>
>>>You're supposing that package not currently using auto* calls may need
>>>some corrections in the future, and being based on automake 1.7, not
>>>having it in the distribution at that time would make patching more
>>>difficult, right ? With such reasoning, we should not even get rid of
>>>automake 1.6...
>> 
>> 
>> And frankly, it would not be a problem for me to have automake1.6 package
>> if needed.
> Actually, I'm not opposed to keeping this particular package in the
> distribution, if needed. I'm opposed to keep versioning everything, and
> hardcoding uneeded version ad vitam eternam. We should have an
> unversioned autotools branch, and use this one by default.

The versioning of automake and autoconf has been done by automake authors
for a reason : being able to parallel install automake install and be able
(if needed) to break backward compatibility.

>>>Anyway, let's first see how much difficult it is to get rid of current
>>>actual use of automake 1.7 before taking such hypothetical case in
>>>consideration. Using an automatic rebuilder as suggested should allow
>>>to evaluate this.
>> 
>> 
>> So, you'll take care of fixing packages in the future which will not
>> build with new version of automake. Thanks for volonteering.
> I'm volonteering for clearing the current situation, based on fact
> rather than suppositions. Nothing more.

And you'll let maintainers deal with the mess you'll create. Great.
That is why I'm opposed to your plan.

>>>>And yes, it is too much, since auto* knowledge is low among most
>>>>packagers (this is not a insult to anybody, it is a fact :), it won't
>>>>get fixed easily.
>>>
>>>The real long-term solution is to share the knowledge among the
>>>maintainers, and to lower the learning curve, instead of keeping the
>>>current insane situation IMHO.
>> 
>> 
>> The real problem is upstream, not in the distro.
>> 
>> And frankly, I think you are loosing your time on this particular
>> subject.
> I don't think so, from the regular questions occuring among maintainers
> about which autotools package to use.

To know which version to use is very simple : just look at Makefile.in, it
states which version of automake was used by upstream authors. Use the
same and everything is fine, no need to worry about anything else.

-- 
Frederic Crozat
Mandriva