On Wed, 2005-01-19 at 19:44 -0600, Daniel Goller wrote:
> what is unprofessional is that we acknowledge that poor QA can lead to 
> breakage that is hard to resolve if one is not a dev with access to all 
> previous versions.

viewcvs... everyone has access to it.

> i think it is unprofessional to acknowledge that it can happen and to 
> reply "then they file a bug" when it could be avoided before hand

No.  Breakage is a bug.  Period.

Now, the fact that it could have been prevented or not does not change
the FACT that breakage is a bug.  You seem to be confusing QA
pre-release with problems encountered after release.

> it is not much more professional (albeit helpful on the technical side) 
> to make a suggestion and right out say "but i will not do it" which 
> again shows no care for what happens to the users

You're talking about SpanKY and you have "professional" in the same
sentence.  You haven't been here long, have you?  :P

> wolf said proper QA would avoid what i outlined, and it would, but we 
> are all humans, and noone is perfect, all i am after is a safety net

Versioned eclasses, as I said and you so magnificently twisted my words
to fit your ideas of what you think I said, will not resolve this
problem.  Bad quality is bad quality, no matter how many versions you
stick on it.

> kind of like building a nuclear reactor with the rods pulled out from 
> top, so in case of power loss they will sink into the core instead of 
> allowing a tschernobil style meltdown (yes im saying we are building a 
> core with rods pulled out from bottom, for analogy sake)

Actually, we're building a nuclear core, and you're discussing the
pounds per square inch of force in your latest potato gun.  There is
absolutely no correlation, yet you're trying to make one.  As I said
before, bad QA is bad QA.  No amount of versioning will solve this.  I
am only repeating this because I want to make it very clear what I mean
and not to have my words twisted.

BAD QA IS BAD QA AND NO AMOUNT OF VERSIONING WILL FIX THAT.

> it was not my intention to say this is how it is done, my intention was 
> to talk about versioned eclasses, while i only achieved to have people 
> pick on an example, if anything, this thread was helpful to choose topic 
> and initial post more wisely in the future
> 
> it should allow me to avoid misunderstandings and keep it closer to the 
> issue, it will be asking for input on how it could be achieved, rather 
> than sounding like i had a soultion i wish to push
> 
> and my summarizing came from the thread overall sounding like noone 
> cares that it can happen, and that something should be done to minimize 
> the risk, that i find unprofessional, to toy with a users system, cause 
> an eclass that can be used across many archs is not tested first, but 
> shoved into the tree instead, and please correct me if i am wrong, 
> stable gcc-3.3.5-r1 inherits toolchain.eclass too, now if it is supposed 
> (lets pretend for example sake) to be that it works entirely independent 
> on gcc-3.3 than on gcc-3.4 i could still respond, that nothing was 
> supposed to change for gcc-3.4, so how do we know it didnt affect 
> gcc-3.3.5 too? then we are facing an eclass change affecting ~arch and 
> arch users, which in my eyes is not professional

What?  I'm sorry, but I had a real hard time following that paragraph.

> im not needing people to agree with me to find them professional, i just 
> would like to see people take their QA seriously, which is what this in 
> the end boils down to

Judging by your past posts in this thread, I call BS on this.  You've
pretty much called anyone that didn't agree with you unprofessional.

Guess what?  We aren't professionals.

Many of us are not developers by trade.  Quite a few of us are still in
school, maybe even high school.  There's nothing wrong with that.  I
would bet many of us have never been in a job where they were held up to
"professional development standards" of any kind.

I know that this will probably piss off some people, but there are
reasons why Gentoo is viewed as a hobbyist distribution.  One of those
reasons is because to most of our developers, THIS IS A HOBBY.

-- 
Chris Gianelloni
Release Engineering - Operational/QA Manager
Games - Developer
Gentoo Linux

Attachment: signature.asc
Description: This is a digitally signed message part

Reply via email to