Re: RFC:PHP Release Cycle
| From: | Zak Greant | Date: | Thu, 09 Nov 2000 02:29:27 +0000 |
| Subject: | Re: RFC:PHP Release Cycle | ||
| References: | 1 2 | Groups: | php.qa |
| Request: | Send a blank email to php-qa+get-1608@lists.php.net to get a copy of this message | ||
----- Original Message -----
From: "Richard Lynch" <richard@zend.com>
To: <php-qa@lists.php.net>; <zend@zend.com>
Sent: Wednesday, November 08, 2000 6:19 PM
Subject: Re: [PHP-QA] RFC:PHP Release Cycle
Hi Richard,
Thanks for contributing! Me and James were starting to count the
tumbleweeds blowing through the list...
> Just from the traffic on the PHP-QA list, I would suggest that 24 hours
is
> unreasonable for a testing time-frame.
True - I may be dreaming...
> RCs go out and reports come in days later, once in a while even a week or
> two later. Obviously you can only stretch it so far, but I don't think
24
> hours for volunteers using their spare time is a feasible turn-around. I
> also suspect that for some folks, PHP-QA is a weekend-warrior task -- So
a
> Friday-Monday cycle would work but a Monday-Friday would not. I could be
> wrong, though. Maybe most QAers are doing it "at work" or in the
evenings
> on "work" machines, and it's the exact opposite... Worst-case scenario,
is
> there are some of each, and you need a full week for a complete QA cycle.
> This probably seems exceedingly long to OpenSource standards, but it
really
> might be best for the current situation...
If we could only get into the habit of doing a smoke build... I have been
doing a build a day for the last week (experimenting and trying to get
various extensions installed) and find that it is not that bad once you get
used to the process. <dream, dream, dream... stars in eyes...>
> In theory, an RC branch and its bug fixes should be "easy" to merge back
> into the main compared to a more general branching, or perhaps developers
> could even be encouraged to add RC bug-fixes to both branches, but new
> chunks of code only to the main branch... Kind of defeats the purpose of
a
> "branch" in a general sense, but again it seems to fit the need here.
>
> If that's not feasible, having developers not commit anything except RC
> bug-fixes seems the only other option.
>
> But the PHP QA team simply can't be effective if large-scale code-changes
> occur between sub-RC releases.
>
> Disclaimer: Easy for me to spout off -- I'm not the one doing the work
> here. You guys rock!
Hmmm... in many ways neither are we. :) I can't offer much help for the
next few weeks, except for building releases and the occasional potshot
comment on the list.
This discussion is all well and good... perhaps the best thing to do would
be to try and finish the build tracker and QA bug queue?
Perhaps if we actually get some of the groundwork done, the developers will
be more interested.
Just my tired and grumpy 0.02.
BTW as soon as I get my book project mostly wrapped up, I will be back
working on the QA effort - hopefully Sterling and I will be working on the
bug tracker in a 2-4 weeks.
Zak "Too stupid to be a pessimist" Greant