>> needs-spec
>> needs-design
>> needs-patch
>> (needs-testing?)
>> needs-review
>
> Actually, it would need a different name, because "needs-testing" is for
> patches that were already committed and need testing on the staging server.
> (Sez Mark.) Perhaps "early testing" or "mass testing"?

Alternately, we could rename this to needs-verification.  So the idea is:

1) feature is requested
2) needs-spec? = spec is designed
3) needs-design? = design is made based on spec
4) needs-patch? = both of the above are ready (or not needed), coder GO
5) needs-testing? = new proposed step for someone who wants external
testing (optional?)
6) needs-review? = ready to review
7) needs-commit? = ready to commit
8) needs-verification? = is committed, and someone should verify the
behavior on stage (this is like needs-testing but later?)
9) needs-merge? = all's good, let's merge this for the next live release

Wow, that's a lot of steps.  Too many?  Let's optimize.  How can we
make sure this process is NOT overwhelming for everybody?

>> (needs-review I'm taking to be Mark officially reviewing the patch for
>> committing into the codebase.
>
> It includes that, but I don't see any reason why it should be limited to
> that. (Eg, if someone else is working in an overlapping area of the code,
> they may want to have a look. Or if it creates a dependency with another
> bug, maybe because it uses DW::Request in a way that doesn't work yet, but
> that will work when another bug is fixed.)

+1, the idea is for other people to become reviewers and hopefully
committers in their particular areas.  At this point, I don't think we
need to spread out yet, but we're getting pretty close.  I've been
contemplating when to ask one or two people to start handling reviews.
 Then I can wait for the needs-commit? and go from there.  (And
eventually, things will get reviewed and committed by
volunteers/staff, but we'll probably maintain one merge point for the
forseeable future.  Even the Linux kernel operates this way, and it's
a huge project!)

>> I think that could work, but I worry that without any way to notify
>> interested parties explicitly, the patch could just sit there and be
>> ignored.
>
> True, but that's also true of the current needs-testing.

There should be a list of "here, if you're interested in $x, use this
Bugzilla query and bookmark it" sort of things.  I have a few right
now, one is "review?" and it shows all OPEN bugs with needs-review? on
an attachment.

>> Doing a shout-out on IRC, or over here in DW-discuss could be better
>> for pinging potential testers, but possible problems: I think that not
>> everyone interested hangs out in IRC, and that it could potentially
>> get spammy if posted on dw-discuss. It's less spammy if only  the
>> original shout-out is posted to the list, and all other discussion is
>> done either through one-to-one correspondence, or on the bugzilla
>> report itself, but I'm still worried that the additional volume could
>> force other subscribers to start tuning out all mail from dw-discuss.
>
> Maybe another mailing list for discussing code? Or one for coding
> specifically, and one for testers?

We've talked about having a dw-dev or something.  I'm still leaning
towards "not enough traffic to warrant a separate list".  Although I
guess there are 280-ish people on dw-discuss...

If anybody has any strong opinions about separating out development
chatter, feel free to speak up!


-- 
Mark Smith / xb95
[EMAIL PROTECTED]

_______________________________________________
dw-discuss mailing list
[email protected]
http://lists.dwscoalition.org/cgi-bin/mailman/listinfo/dw-discuss

Reply via email to