ARRRRGH! Guys, who's breaking -STABLE's GMIRROR code?!

[Greg Byshenk]

On Fri, Sep 15, 2006 at 03:41:04PM -0300, Marc G. Fournier wrote:

> But, I'm just curious here ... for all of the talk going around about this 
> whole issue, how many ppl have truly ever been bitten by an unstable 
> -STABLE?  And for those that have, how long did it take to get help from a 
> developer to get it fixed?

I run -STABLE on a number of production machines.

I have twice been "bitten by an unstable -STABLE" -- but "bitten" in a 
very small way.

When we build a new -STABLE (on average perhaps once per month), we
build it on a test machine, so that we can be sure that it actually
works. Once it is tested and we know it works, then we can roll it out
to the production machines without undue concern.

I note that we follow the same process with out Linux machines, our
Irix machines, and our Windows machines.  Blindly rolling out updates
or patches to critical production machines is unwise and dangerous (at
least IMO).

I will add that I have never even needed to contact a maintainer.
When there has been a problem, I checked the lists.  In one case the
fix was already committed, in the other there was already an "I'm
working on it" message and a fix was commited in less than 24 hours.
In the interim, my test machine had a problem -- but that's what a 
test machine is for.


> In the case that started this thread, it seems to be that the developer 
> fixed his mistake fairly quickly, which is what one would expect ... it 
> shouldn't be so much that he *broke* -STABLE (shit happens, do you want 
> your money back?), but it should be 'was he around to reverse his mistake 
> in a reasonable amount of time?' ... ?

Exactly.


-- 
greg byshenk  -  gbyshenk at byshenk.net  -  Leiden, NL