GNU bug report logs - #11323
automake-1.11.4 regression

Previous Next

To add a comment to this bug, you must first unarchive it, by sending
a message to control AT debbugs.gnu.org, with unarchive 11323 in the body.
You can then email your comments to 11323 AT debbugs.gnu.org in the normal way.

Toggle the display of automated, internal messages from the tracker.

Message #5 received at submit <at> debbugs.gnu.org (full text, mbox):

From: Ralf Corsepius <ralf.corsepius <at> rtems.org>
To: bug-automake <at> gnu.org
Subject: automake-1.11.4 regression
Date: Tue, 24 Apr 2012 09:50:29 +0200

Hi,

With automake < 1.11.4 it was possible to create empty directories this way:

Makefile.am:
mystatedir = $(pkglocalstatedir)
mystatedir_DATA =


With automake-1.11.4, this doesn't work anymore and silently breaks 
existing Makefile.ams.

Ralf

Message #8 received at 11323 <at> debbugs.gnu.org (full text, mbox):

From: Stefano Lattarini <stefano.lattarini <at> gmail.com>
To: Ralf Corsepius <ralf.corsepius <at> rtems.org>
Cc: 11323 <at> debbugs.gnu.org
Subject: Re: bug#11323: automake-1.11.4 regression
Date: Tue, 24 Apr 2012 10:34:56 +0200

tags 11323 notabug
close 11323
thanks

Hi Ralf.

On 04/24/2012 09:50 AM, Ralf Corsepius wrote:
> Hi,
> 
> With automake < 1.11.4 it was possible to create empty directories this way:
> 
> Makefile.am:
> mystatedir = $(pkglocalstatedir)
> mystatedir_DATA =
> 
> 
> With automake-1.11.4, this doesn't work anymore and silently breaks
> existing Makefile.ams.
> 
Not a regression, but an intended change, also documented in the NEWS
file:

  - An empty declaration of a "foo_PRIMARY" don't cause anymore the
    generated install rules to create an empty $(foodir) directory

and in the relevant  news entry on Savannah:

  <http://savannah.gnu.org/forum/forum.php?forum_id=7175>

See also:

  <http://debbugs.gnu.org/cgi/bugreport.cgi?bug=11030>

Regards,
  Stefano

Message #15 received at 11323 <at> debbugs.gnu.org (full text, mbox):

From: Ralf Corsepius <ralf.corsepius <at> rtems.org>
To: 11323 <at> debbugs.gnu.org, GNU Automake mailing list <automake <at> gnu.org>
Subject: Re: bug#11323: automake-1.11.4 regression
Date: Tue, 24 Apr 2012 10:49:21 +0200

On 04/24/2012 10:34 AM, Stefano Lattarini wrote:
> tags 11323 notabug
> close 11323
> thanks
>
> Hi Ralf.
>
> On 04/24/2012 09:50 AM, Ralf Corsepius wrote:
>> Hi,
>>
>> With automake<  1.11.4 it was possible to create empty directories this way:
>>
>> Makefile.am:
>> mystatedir = $(pkglocalstatedir)
>> mystatedir_DATA =
>>
>>
>> With automake-1.11.4, this doesn't work anymore and silently breaks
>> existing Makefile.ams.
>>
> Not a regression, but an intended change,
I once more have to disagree with you:

This change is a critical behavioral regression, causes defective 
packages and forces packagers to heavily hack their packages.

>   also documented in the NEWS
> file:
>
>    - An empty declaration of a "foo_PRIMARY" don't cause anymore the
>      generated install rules to create an empty $(foodir) directory
>
> and in the relevant  news entry on Savannah:
>
>    <http://savannah.gnu.org/forum/forum.php?forum_id=7175>
>
> See also:
>
>    <http://debbugs.gnu.org/cgi/bugreport.cgi?bug=11030>
>
-EFAILMAINTAINER

a) This kind of changes is inappropriate within a release series.
b) You don't seem to be aware about the harmfulness of this change of yours.

Please comprend that this change was a mistake and revert it!


Ralf

Message #18 received at 11323 <at> debbugs.gnu.org (full text, mbox):

From: Ralf Corsepius <ralf.corsepius <at> rtems.org>
Cc: 11323 <at> debbugs.gnu.org, GNU Automake mailing list <automake <at> gnu.org>
Subject: Re: bug#11323: automake-1.11.4 regression
Date: Tue, 24 Apr 2012 11:04:51 +0200

On 04/24/2012 10:49 AM, Ralf Corsepius wrote:
> On 04/24/2012 10:34 AM, Stefano Lattarini wrote:

>> <http://savannah.gnu.org/forum/forum.php?forum_id=7175>
>>
>> See also:
>>
>> <http://debbugs.gnu.org/cgi/bugreport.cgi?bug=11030>
>>
> -EFAILMAINTAINER
>
> a) This kind of changes is inappropriate within a release series.
> b) You don't seem to be aware about the harmfulness of this change of
> yours.

Let me elaborate. previously it was possible to do something of this kind:

if FEATURE0
mystatedir = $(pkglocalstatedir)
mystate_DATA =

if FEATURE01
mystate_DATA += xxx
endif
if FEATURE02
mystate_DATA += yyy
endif
endif

Now you are forcing devs to manually add rules to achieve the same 
effect. Worse, if automake on packags which have not been manually added 
such manually written rules, these package end up broken (e.g with build 
errors in Linux distro-buildsystems).

Ralf

Message #21 received at 11323 <at> debbugs.gnu.org (full text, mbox):

From: Miles Bader <miles <at> gnu.org>
To: Ralf Corsepius <ralf.corsepius <at> rtems.org>
Cc: 11323 <at> debbugs.gnu.org, GNU Automake mailing list <automake <at> gnu.org>
Subject: Re: bug#11323: automake-1.11.4 regression
Date: Tue, 24 Apr 2012 18:08:41 +0900

Ralf Corsepius <ralf.corsepius <at> rtems.org> writes:
> -EFAILMAINTAINER

No need to be obnoxious.

> a) This kind of changes is inappropriate within a release series.
> b) You don't seem to be aware about the harmfulness of this change of yours.
>
> Please comprend that this change was a mistake and revert it!

I think some actual data might be in order...  _which_ packages break?
In what way?  [I mean: what's the actual effective on the package's use?
In some cases a failure to create an empty directory might be a change
in behavior, but not particularly critical...]

Thanks,

-Miles

-- 
Bore, n. A person who talks when you wish him to listen.

Message #24 received at 11323 <at> debbugs.gnu.org (full text, mbox):

From: Stefano Lattarini <stefano.lattarini <at> gmail.com>
To: Ralf Corsepius <ralf.corsepius <at> rtems.org>
Cc: 11323 <at> debbugs.gnu.org, GNU Automake mailing list <automake <at> gnu.org>
Subject: Re: bug#11323: automake-1.11.4 regression
Date: Tue, 24 Apr 2012 11:12:11 +0200

On 04/24/2012 10:49 AM, Ralf Corsepius wrote:
>
>> On 04/24/2012 09:50 AM, Ralf Corsepius wrote:
>>> Hi,
>>>
>>> With automake < 1.11.4 it was possible to create empty directories
>>> this way:
>>>
>>>  Makefile.am:
>>>  mystatedir = $(pkglocalstatedir)
>>>  mystatedir_DATA =
>>>
>>> With automake-1.11.4, this doesn't work anymore and silently breaks
>>> existing Makefile.ams.
>>>
>> Not a regression, but an intended change,
>
> I once more have to disagree with you:
> 
> This change is a critical behavioral regression, causes defective packages
> and forces packagers to heavily hack their packages.
>
s| hack | modify not to rely on an undocumented behaviour |

>>   also documented in the NEWS file:
>>
>>    - An empty declaration of a "foo_PRIMARY" don't cause anymore the
>>      generated install rules to create an empty $(foodir) directory
>>
>> and in the relevant  news entry on Savannah:
>>
>>    <http://savannah.gnu.org/forum/forum.php?forum_id=7175>
>>
>> See also:
>>
>>    <http://debbugs.gnu.org/cgi/bugreport.cgi?bug=11030>
>>
> -EFAILMAINTAINER
> 
> a) This kind of changes is inappropriate within a release series.
>
This might be true; I should have been less eager to install such a
change.  Oh well, no point in crying over spilt milk, but I'll have to
be more careful next time.

> b) You don't seem to be aware about the harmfulness of this change
>    of yours.
> 
> Please comprend that this change was a mistake and revert it!
>
Have you read the discussion I have linked above, in automake bug#11030
and the references given there?  If you want this change reverted,
you'll have to counter the points made there in is favour.

Regards,
  Stefano

Message #27 received at 11323 <at> debbugs.gnu.org (full text, mbox):

From: Stefano Lattarini <stefano.lattarini <at> gmail.com>
To: Ralf Corsepius <ralf.corsepius <at> rtems.org>
Cc: 11323 <at> debbugs.gnu.org, GNU Automake mailing list <automake <at> gnu.org>
Subject: Re: bug#11323: automake-1.11.4 regression
Date: Tue, 24 Apr 2012 11:26:53 +0200

On 04/24/2012 11:04 AM, Ralf Corsepius wrote:
> 
> Let me elaborate. previously it was possible to do something of this kind:
> 
> if FEATURE0
> mystatedir = $(pkglocalstatedir)
> mystate_DATA =
> 
> if FEATURE01
> mystate_DATA += xxx
> endif
> if FEATURE02
> mystate_DATA += yyy
> endif
> endif
>
To obtain what exactly?  I can't understand your use case if you don't
explain me what you expect to obtain from the given input ...

> Now you are forcing devs to manually add rules to achieve the same effect.
>
Which effect?

> Worse, if automake on packages which have not been manually added such
> manually written rules, these package end up broken (e.g with build
> errors in Linux distro-buildsystems).
>
New versions of these package should adapt to the new automake behaviour
once (anf if) they upgrade from automake < 1.11.4 to automake >= 1.11.4.
And if you are re-bootstrapping an older version of these package (after
you've applied distro-specific patches I guess), you should be prepared
to use the same autotools versions they used in the corresponding release
tarball.

Regards,
  Stefano

Previous Next

GNU bug tracking system
Copyright (C) 1999 Darren O. Benham, 1997,2003 nCipher Corporation Ltd, 1994-97 Ian Jackson.