[Framework-Team] scope of reviews was: Re: [Plone-developers] Updated PLIP review deadline

Raphael Ritz raphael.ritz at incf.org
Fri Feb 15 12:12:15 UTC 2008


Tom Lazar wrote:
> On Feb 7, 2008, at 8:52 AM, Martin Aspeli wrote:
>
>> For 3.0, we had one main reviewer for each bundle who'd report back for
>> a quick vote on the list. People with concerns could then dig into the
>> code in more detail or request more specifics.
>
> interesting point. maybe that's what's holding me back personally 
> (only finished one review up until now, buildout of the second one 
> running as i'm writing this). i was under the impression that a review 
> consisted of 'serious QA', ie. hunting down the individual revisions 
> that make up the plip and looking at the diffs; doing manual 
> clicktests on FF, Safari, Win IE 6/7, including migrations then 
> discussing any bugs that come up with the author and see their fixes 
> through until resolution (with possibly bouncing unacceptable fixes 
> back to the author once or twice).
>

That would be the ideal situation, yes.

> that stuff can amount to some serious time, i can tell you. but OTOH 
> i've read things like, "it usually only takes 20-30 minutes" for a 
> review, what gives?
>

only a very first course overview

> maybe i'm just taking this *too* seriously and thereby (ironically) 
> actually holding the whole process *back*.
>
> anyway, i'd be interested to hear some feedback on this, in the 
> meantime i'll be doing another review ;-)
>

It's all about a reasonable balance I think.
No two PLIPs are the same and some are easier to handle than others.
E.g., in my case the one I've spend the most time on I didn't
even comment yet (the WebDAV stuff) while others went real
quick indeed.
What I usually do try in all cases is to think about edge cases
and whether they work or not and sure enough this is where
issues are hidden. That's how I figured out that the kupu formlib
widget is broken with FCK editor or that the GS export handlers for
the portlets don't handle unicode/non-ascii correctly - not by
reading the code in detail (though I do browse through it).

Just my 2 cents,

    Raphael


> cheers,
>
> tom
>
>>
>>
>> Martin
>>




More information about the Framework-Team mailing list