--- In [EMAIL PROTECTED], Ron Jeffries
<[EMAIL PROTECTED]> wrote:
...
> No ... "Everyone" is somewhere else, both physically and mentally,
> coding away. A failed build interrupts everyone, or everyone

Re: Interrupts everyone
As it should.

> implicated, at their desk. But the whole theory of your idea is that
> they have moved on. No one is ready, no one is engaged in the build
> process. It's a fire alarm.

What happens in practice...

1. Pair codes task using TDD.  Tests related to task run in < 2 seconds.
2. Task completed in ~20 minutes or less.  Update from repository. 
Execute pre-commit build which should take < 5 minutes (I prefer < 2
minutes)
3. At this point, we are saying that we are confident that the state
of the code is good enough to commit to the repository, so we do that.
4. Asynchronous, more exhaustive, post-commit build is triggered and
should complete in 10 minutes or less.
5. If the post-commit build broke, then yes, there is a fire alarm
because there is a fire.  Our confidence that the pre-commit build was
good enough was misplaced.  Fix the build and restore the confidence
by adding to and/or adjusting the pre-commit build.

At any point in the above, there may be reflection on what has been
done and/or discussion on what should be done next.

So the theory of why this works...

- Post-commit build failures should be rare.  Rare enough such that
the time saved by not waiting each time more than pays for any loss of
context.
- The team will learn how to keep post-commit build failures rare.
- Keeping post-commit build times low will also keep the loss of
context low
- Keeping task times low limits the amount of changes to deal with
- Not waiting for longer post-commit builds encourages developers to
commit smaller tasks

> It also seems a bit nose to the grindstone, now that I think about
> it.

I don't find it to be.  It can turn that way if failed builds are used
to beat people down.

> I don't know, I've never tried it. It might be wonderful. But it's
> not where I'd start. It might be where I'd wind up.

I suspect that you don't believe that's where you'd wind up?

I don't believe that serial CI is where I'd wind up nor would it be
where I'd start.  Granted, I've never tried it.





To Post a message, send it to:   [EMAIL PROTECTED]

To Unsubscribe, send a blank message to: [EMAIL PROTECTED]

ad-free courtesy of objectmentor.com 
Yahoo! Groups Links

<*> To visit your group on the web, go to:
    http://groups.yahoo.com/group/extremeprogramming/

<*> To unsubscribe from this group, send an email to:
    [EMAIL PROTECTED]

<*> Your use of Yahoo! Groups is subject to:
    http://docs.yahoo.com/info/terms/
 



Reply via email to