[Developers] Accessing current source code

[Ian Taylor]

  Hi,
Thanks for this discussion about the source code.

I realize that there might be reasons for wanting 
separate development code from what's publicly 
accessible. However, I think the general public 
deserves to know that there's a difference. That 
is, if you go to 
http://code.google.com/p/admb-project/source/checkout
and check out the code or browse around the files, 
I think you deserve to be informed somehow that 
this isn't the code that is under current 
development. If Arni forgets to do svn checkout, 
that's his fault. But how about a warning for the 
folks who don't forget, but never see any changes 
because they're using the wrong repository?

Finally, any ideas about how long it will be 
before those of us who don't know the secret 
handshake will be able to access the bleeding edge 
code?

-Ian

On 8/19/2010 2:57 PM, John Sibert wrote:
> Ben, Arni, et al -
> Yes there is a big caveat emptor to be invoked 
> for those who use the bleeding edge code (or as 
> the practice has come to be known, eating your 
> own dogfood). I routinely use the development 
> code - checking it out and building it about 
> once a week. It is seldom broken, because the 
> test suite is part of the repository and you get 
> it when you check out the code. It is the 
> developer's responsibility to rebuild the 
> libraries and run "make test" before committing 
> code to the repository.
>
> Testing is an integral part of the routine build 
> process under the control of the buildbot. Every 
> time someone commits code to the repository, the 
> buildbot knows about it and automatically 
> triggers a build and test (see 
> http://www.admb-project.org/buildbot/grid for 
> all the gory details).
>
> Cheers,
> John
>
> On 08/19/2010 11:03 AM, Ben Bolker wrote:
>> On 10-08-19 03:12 PM, Arni Magnusson wrote:
>>> Hi Ben, these are valid points.
>>>
>>> I'm not sure how old the 'public' SVN version 
>>> is. It may date back to the
>>> last formal release of ADMB 9.1 (31 Dec 2009), 
>>> equivalent to downloading
>>> and unzipping admb-9.1-src.zip. The reported 
>>> SVN revision 74 makes no
>>> sense, as version 9.1 corresponds to revision 
>>> 496.
>>>
>>> Like you point out, this can be frustrating 
>>> and undermines user
>>> participation in the development of ADMB. 
>>> Similar free software projects
>>> have anonymous read-only access to the actual 
>>> source repository, so anyone
>>> can view the code in real time.
>>>
>>> That level of open source enforces development 
>>> discipline; at any given
>>> time, build errors should be very rare and 
>>> quickly fixed.
>>>
>>> My guess is that ADMB has not reached that 
>>> point, or has reached it very
>>> recently. If I checkout the source code right 
>>> now, I wouldn't assume that
>>> it builds.
>>    A minor comment:
>>
>>    in my opinion it's OK to have a publicly 
>> accessible development SVN
>> repository that is frequently broken. That's 
>> what 'bleeding-edge' means.
>> Not broken is nicer, of course, and if 
>> developers get embarrassed by
>> having their dirty laundry on display and start 
>> being more careful not
>> to break the development version that's great, 
>> but people who choose to
>> use the development version have been warned: 
>> they should be willing to
>> take what they get and/or should have the 
>> technical expertise to revert
>> to a version that isn't broken.
>>
>>    cheers
>>      Ben
>>
>> _______________________________________________
>> Developers mailing list
>> Developers at admb-project.org
>> http://lists.admb-project.org/mailman/listinfo/developers 
>>
>>
>