Re: [Savannah-hackers-public] Pusing to Git repo declined

[Jim Meyering]

On Tue, May 24, 2016 at 10:30 PM, Aljosha Papsch <address@hidden> wrote:
> On 25.05.2016 05:27, Bob Proulx wrote:
>>
>> Assaf Gordon wrote:
>>>
>>> Hello Aljosha,
>>>>
>>>> Aljosha Papsch wrote:
>>>> today I tried pushing to a Git repository via SSH, but it fails:
>>>> ...
>>>> remote: error: hook declined to update refs/heads/master
>>>> ...
>>>> Why does Savannah reject my changes? If it's really about coding style,
>>>> as the messages suggest, I will be happy fixing those. But I can't
>>>> believe
>>>> Savannah is trying to make me push clean files :) Is there some other
>>>> issue?
>>>
>>> Let me check that and hopefully I'll have an answer soon.
>>
>> Honestly I when I set up that repository I simply cloned the coreutils
>> setup.  It has that configuration.  Sometimes the cut-n-paste paradigm
>> can catch up with you.  That is the case here and how I did it.
>>
>> However I personally would consider it.  Most of the issues it was
>> complaining about look like real issues to me.  It was complaining
>> about spaces before TABs.  That is almost always a bad thing.  It
>> complains if the file doesn't end with a newline.  Not ending with a
>> newline causes the file to be treated as a binary file instead of a
>> text file since by definition text files are lines terminated by
>> newlines.  Things like that are actually very important errors.  The
>> trailing whitespace is a much lessor issue but tends to be a problem
>> for gratuitious diffs between different committers.
>>
>> I have removed that hook from the git repository.  You won't get that
>> rejection from it for any content now.  Sorry for the unpleasant
>> interaction from my cloning another setup for you.
>>
>> Bob
>>
> Thanks for the clarification. I will fix these files now. While forcing
> users to push clean files is unexpected for user, I guess it would be great
> if
> these messages were warnings and would not result in a fatal error.
> Users could then decide for themselves whether to clean up files, in the
> sense "You have been warned". Is that possible?

This is a last-resort policy-enforcer set by the package
maintainer(s). One typically has first-level notifiers (editor, tools)
to warn about fundamental policy/cleanliness violations like this, so
that it's only in the unusual event that a violation gets past those
that it will trigger a push rejection.

Of course, if those who contribute to a particular package want a
different policy, someone here can change your instance of the commit
hook to work differently.