petsc-dev directory structure questions

[Barry Smith]

On Dec 13, 2007, at 5:07 PM, Satish Balay wrote:

>
> For one, I think conf/configure.py should be ./configure.py

    How come this wasn't done originally or hasn't been changed to
this already?

>
>
> And conf and config are 2 different things - so they should be in
> separate locations. merging them doesn't make sense to me. The primary
> issue I see is the names "conf" and "config" are extremely close to
> each other.

    Exactly, so how the heck is anyone suppose to know what is in either
one of them?

>
>
> One option is to move config to src/config [this would work well with
> ./configure.py]

    config stuff is not source code so cannot go in there.

>
>
> Satish
>
> On Thu, 13 Dec 2007, Barry Smith wrote:
>
>>
>>  Matt,
>>
>>    This is a rational for the current set up (or perhaps more a
>> rationalization :-))
>> The reason for merging is for others attempting to understand our
>> system; all the little complications add up to a system that is  
>> overwhelming
>> to other people so in my mind anything that reduces the  
>> complications is
>> good thing (so long as it doesn't screw up other things).
>>
>> Does anyone else have comments?
>>
>>  Barry
>>
>>
>> On Dec 9, 2007, at 10:03 AM, Matthew Knepley wrote:
>>
>>> I am cool with the division because
>>>
>>> 1) config is for configuration tools explicitly
>>>
>>> 2) conf is for things that customize the build
>>>
>>> They have completely different things in them, and I do not see  
>>> that we
>>> need to merge them.
>>>
>>> Thanks,
>>>
>>>   Matt
>>>
>>> On Dec 8, 2007 10:13 PM, Barry Smith <bsmith at mcs.anl.gov> wrote:
>>>>
>>>>  Currently petsc-dev has the subdirectories
>>>>
>>>>    bin include src conf config ...
>>>>
>>>> $PETSC_ARCH has the subdirectories
>>>>
>>>>  bin include conf lib ...
>>>>
>>>> conf has
>>>>
>>>>  adic.init           adicmf.init              
>>>> base                    test
>>>>  adicmastercontrol   adicmfb.init            rules
>>>> variables
>>>>
>>>> config has
>>>>
>>>>  PETSc BuildSystem configure.py   +tons of sample configure.py
>>>> files for different systems
>>>>
>>>> ------------------------------------------------------------------------------
>>>>
>>>> I find that having both a conf and a config directory is confusing
>>>> and unneeded and propose:
>>>>
>>>> 1) merging the config and conf directory
>>>> 2) putting the ons of sample configure.py files for different  
>>>> systems
>>>> into a subdirectory called samples
>>>>
>>>> Questions:
>>>>
>>>> 1) Does this make sense?
>>>>
>>>> 2) What should the directory be called? conf or config? Are there  
>>>> any
>>>> standards that dictate/suggest
>>>>   one over the other? If we use config then I would like to change
>>>> the $PETSC_ARCH/conf directory
>>>>   to match it, but would that violate some standard? I'm inclined
>>>> to go with conf even though that breaks
>>>>   our long standing config/configure.py script to become conf/
>>>> configure.py
>>>>
>>>> 3) This is perhaps also a good time to reopen the question of PETSc
>>>> "installs": currently they go into
>>>>   (system independent) prefix/bin prefix/include prefix/conf and
>>>> the system dependent prefix/$PETSC_ARCH/conf
>>>>   prefix/$PETSC_ARCH/include prefix/$PETSC_ARCH/lib If one selects
>>>> a PETSC_ARCH of "" then
>>>>   everything ends up in prefix/bin prefix/include prefix/conf
>>>> prefix/lib.
>>>>
>>>>  The reason for this design is so that the same makefiles with
>>>> includes based on PETSC_DIR and
>>>>  PETSC_ARCH will work in all three cases: no install of PETSc,
>>>> install with PETSC_ARCH and install with
>>>>  PETSC_ARCH ""
>>>>
>>>>  I know we will never get full agreement on everything, but what
>>>> are the concerns with this layout?
>>>>
>>>>
>>>>  Thanks
>>>>
>>>>    Barry
>>>>
>>>> Remember if I don't get comments I may do some crazy thing :-(
>>>>
>>>>
>>>>
>>>
>>>
>>>
>>> -- 
>>> What most experimenters take for granted before they begin their
>>> experiments is infinitely more interesting than any results to which
>>> their experiments lead.
>>> -- Norbert Wiener
>>>
>>
>