Pregenerated files: repo freeze and discourse

[Ralf Corsepius]

On 08/01/2012 08:01 AM, Thomas Doerfler wrote:
> Hi,
>
> just my two cents in this discussion:
>
> Am 01.08.2012 05:44, schrieb Ralf Corsepius:
>> On 07/31/2012 07:41 PM, Gedare Bloom wrote:
>>> Hello,
>>>
>>> I would like to request that we freeze all commits to RTEMS and
>>> consider what to do about including generated files in the repository.
>>> Agreement (much less consensus) was not reached and now we have a
>>> large number of automatically generated files in rtems.git that
>>> substantially bloats the size of the repository.
>>
>> The sizes of the repos hardly has any impact on using the repos.
>
> At least the transfer sizes for "git clone" should be affected.

Right, the transfer sizes change once ... but afterwards, almost nothing 
changes.

> If the
> generated files make up a visible amount of the total file sizes, and
> these files will be regenerated very often,

This is untrue - They only change if the corresponding source files 
change or if the versions of the autotools being used change.

I.e. instead of forcing each and every user to "bootstrap" his local 
checkouts, to expose himself to local installation issues (e.g. to be 
using different versions of the autotools), this now is done centralized 
=> Deterministic build infrastructure.

> this will generate a lot of
> versions for each of these files, and each version will be cloned for
> each developer. This would not be a problem if their content stayed
> stable, but obviously it ain't.

Again, no, this is untrue. These files only change when their "source 
counterparts change".

This may introduce a lot of commits, when working on local clones, but 
it does not introduce a "lot of commits" in the public repos.

>>> So far I have not seen very clear explanation why adding these
>>> generated files to the git repository is the best solution long-term
>>> for RTEMS.
>> I already tried to explain this before.
>>
>> The advantages are:
>> - No need for "bootstrap".
>> - Makes bugs in configuration files more visible.
>> - Makes user mistakes in configuration files more visible.
>
> Can you elaborate the last two points further? I don't see the point
> (maybe due to my limited understanding of auto*)

Try the following:

git clone <rtems4.11> rtems4.11
cd rtems4.11
./bootstrap -c -f -f
./bootstrap
git diff

You will see incomplete and bogus Makefile.in's


>>> All developers should feel free to contribute to this discussion and
>>> weigh-in with suggestions about what should be done.
>> With all due respect, I do not see much sense in this. To me this
>> discussion already has become a shit storm, because some
>> semi-knowledgiable folks are spreading FUD.
>
> You are quite fast with this view.

Your freedom to think so - The responses, I received after commiting the 
Makefile.ins/configure scripts spoke a clear and misunderstandable language.


Ralf