Files Added To Git

Chris Johns chrisj at rtems.org
Fri Jul 27 20:57:36 UTC 2012


On 28/07/12 12:28 AM, Ralf Corsepius wrote:
> On 07/27/2012 02:43 PM, Joel Sherrill wrote:
>> Hi
>>
>> I am on holiday but I don't recall any discussion or agreement that
>> automatically generated files should be in git.
>
> I can't help you, if you ignore my emails. I mentioned this several
> times, and had started to add them gradually.

This is incorrect and unfair. You offered no specific detail. Comments 
have been made and you decided to ignore those. I had not seen any 
agreement for your changes, only opposition. No one has agreed with you.

>
> That said, I feel you are trying to make a lot of noise about nothing.
>

He is not and I am not either. There was no acceptance of this large 
change. You offered no public review be it posted patches or in this 
case a repo where they could be reviewed, discussed and tested in 
detail. Review after commit for changes like this are unacceptable and 
an abuse of your open commit access.

>
>> They could have been added to cvs so we could have discussed this
>> anytime in the past 17 years.
>
> Correct. It's what e.g. gdb, newlib and binutils have been doing for
> decades, and what GCC had done before they switched to svn.
>

We are not one of those projects and because they do does not make them 
an ideal model to follow. This is RTEMS and what GNU projects do is 
their business. If you think it does then raise this and discuss it and 
I will disagree. I do not wish RTEMS to be locked into processes and 
models those projects cannot avoid. They have so much cruft and 
complexity hidden in Makefiles and configure scripts they cannot change. 
Coping that complexity is a mistake. As developers age and leave those 
project that knowledge leaves and they cannot change.

Chris



More information about the devel mailing list