git summary messages

[Ralf Corsepius]

On 02/23/2012 08:49 AM, Sebastian Huber wrote:
> On 02/23/2012 04:36 AM, Ralf Corsepius wrote:
>> On 02/22/2012 07:09 PM, Joel Sherrill wrote:
>>> Please use proper git summary messages.
>>>
>>> Using the first line of ChangeLog with just a date
>>> and author for the git summary is absolute rubbish
>>
>> Well, you view. IMO, depending on a VCS's internal implementation
>> details is
>> absulte rubbish.
>
> Why use a VCS at all?

Sebastian, now you are behaving infantile, to say the least!
Nobody but _you_ questions the need to use a VCS!


The point is: How far to tie yourselves to a tool, how far to tie your 
users to a tool and how far to make your sources dependent on a tool.

Think about people who are used to hardcoding SCCS, RCS, emacs and vi 
meta information into their sources ... whether you like this or not we 
are still struggling with this.

... and if you think this to an end in git-kiddy fashion: Who needs 
releases, who needs tarballs?

... who will be able to generate today's git ChangeLogs several years 
ahead, when RTEMS will switch to the "foobar"-VCS, because RTEMS git 
will be victim of a hash-clash? The old ChangeLogs will still be around 
and readable, ...

>> I guess you are aware it's possible to forge git authors and dates by
>> accident
>> or by violence, rsp. that is easily possible for git to loose authors
>> and dates?
>
> Every complex software has bugs, but I don't think its easy to forge
> something protected by the SHA1 at the moment (maybe the NSA guys think
> differently).

Oh, it's much easier than you think ... I guess, you just provoked me 
sufficiently to challenge it.

Ralf