http://bugzilla.spamassassin.org/show_bug.cgi?id=3376





------- Additional Comments From [EMAIL PROTECTED]  2004-05-13 01:39 -------
Subject: Re:  connfailed =  spamc: connect(AF_INET)

> However, you are right...testing it on other platforms before doing a
> checkin is a good idea. I keep repeating myself on this until maybe
> someday you will hear me: Holding up other developers because you
> check in buggy code is not a good way to do development.

If you want to maintain the infrastructure needed to do this (and donate
the large number of platforms we don't have), we'd love your
contribution.  To date, to test other platforms (which I have done, for
example to test the SHA1 code), I have to use the SourceForge shell
machines which is extraordinarily painful since you can't ssh directly,
you can't use the web, and you can't use svn either.  Cut and paste
works (gzip+uuencode is so much fun).

> At CollabNet, it was considered major bad juju to check in code the
> broke the product because it hampered and slowed down development for
> everyone which was quite frustrating all around.

This hasn't slowed down our development at all.  It's a test.  Tests can
fail sometimes.  It's allowed, especially in the development trunk.
It's considered major good juju here to check in tests, even if they
completely fail initially.
 
> I guess what really bothers me most is that this issue doesn't even
> seem to be a platform specific problem. So, I'm not sure why you are
> saying that you will test it across platforms. Are you saying that
> when you ran 'make test' on your box before doing that checkin and
> that it worked for you?

He's probably just lying.
 
> Just to troll a bit and which I don't really expect a reply...If you
> guys have so many platform specific issues to deal with, why the heck
> are you using Perl/C? Maybe another language(s) that has better
> platform cross compatibility would be a better idea? Of course I'm
> sure that will start a flame war or two...

Blah blah blah blah.
 
> Based on all the constant test breakages, it seems that it would be a
> good idea to setup a few runtime boxes that constantly build the
> latest svn head to catch these problems right away and not depend on
> outside sources such as myself to constantly tell you that you can't
> write stable code the first time ;-) ;-).

I'm quite familiar with the concept (we did similar things at Transmeta)
and I'm also a strong proponent of the idea.  That's why contributions
such as setting up a cross-platform automated test environment would be
quite welcome.

The winkies don't really help with your lame attitude where all you can
do is complain and nag.  If you really want to help, contribute code.
Otherwise, please just submit bugs without the dripping condescension
because we can self-flagellate just fine in private.

> You project is with the ASF now. There is an OSX box for use that is
> hosted at Apple and managed by Wilfredo. Send infrastructure AT apache
> D O T org an email and ask for an account on that box. I'm sure that
> will help shut me up and stop annoying you so much. =)

It doesn't sound like it's really ready if we need more than one
account.  We should wait for officially supported ASF resources that
don't require each of us to request an account, etc. etc. etc.





------- You are receiving this mail because: -------
You are on the CC list for the bug, or are watching someone who is.

Reply via email to