launchpad-dev team mailing list archive

[Deryck Hodge <deryck.hodge@xxxxxxxxxxxxx>]

On Mon, Nov 22, 2010 at 11:39 AM, Robert Collins
<robert.collins@xxxxxxxxxxxxx> wrote:
> Man, it sucks that our infrastructure here is so broken [No one needs
> more than 640K].
>
> This plan has some risk in it. Have you considered fixing Windmill? It
> may be unmaintained upstream, but we're still using it.

Yes, I did consider fixing Windmill.  There's an assumption there that
I *can* fix Windmill.  Okay, so I'll give myself the benefit of the
doubt.  Let's say I can. :-)  Then there's the assumption that fixing
the 512k bug will unblock this landing.  We don't know what other
problems might be lying around with so much JavaScript.  We know the
512k bug is the current problem.  I could spend a couple days getting
that fixed and still not be able to land this branch.  I've also still
got to deal with the file size because a download of 1.3 Mb of JS is
ridiculous.  So there's that issue by itself, too.

So I chose this plan instead of trying to fix Windmill because I
*know* it will work and it's the quickest path toward fixing the file
size issue and the testing issue.

>
> I'm tentatively in support of your plan, but I'll take bets *now* that
> we're going to have a brown paper bag event as a result of it.
>
> Its just that I'm happy that with our rollback story now (reactive the
> old symlink ASAP) we can rollback a bad deploy in a matter of minutes.
> That plus all our deploys being initiated by
> developers-taking-responsibility gives me a warm feeling to counter
> the chills that disabling a whole wing of tests gives me.
>

Yes, I feel better about the risks, too, because of the ease at which
we can rollback at bad revno and the requirement of manual qa to
deploy.  We haven't had this in the past.  FWIW, I can't think of a
single failure Windmill has caught of the sort where someone was
working on something unrelated to js and broke widgets across the
site.  This happened a couple times when upgrading lazr-js and some
changes to widgets in the library had site-wide impact.  This would
have been caught by testing or manual qa.  And we shouldn't upgrade
lazr-js or yui in LP after what I'm doing now, until we settle all of
these issues.

Thanks for your tentative support! :-)  I know it makes you nervous,
and I want to make this as painless as possible.

Cheers,
deryck


-- 
Deryck Hodge
https://launchpad.net/~deryck
http://www.devurandom.org/