[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: Is Sid for broken stuff? Is it too much to ask for testing the packages?

On Thu, Dec 12, 2002 at 12:10:05AM -0500, Matt Zimmerman scribbled:
> On Thu, Dec 12, 2002 at 01:08:37AM +0100, Marek Habersack wrote:
> > Hello everybody,
> > 
> >   First of all - I'm pissed off, so ignore any inflammatory comments should
> > I miss to spot them in what I'm going to write. 
> > 
> >   So, is Sid supposed to be a place where maintainers upload broken packages
> > so that others hit the problem and report them as bugs? And I'm not talking
> > about big problems, no:
> > 
> > http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=172677
> Yes, this is exactly the sort of problem that should be found in unstable.
> The maintainer obviously tested that the package installed correctly for
It is not. See the other mails, I don't want to repeat the arguments again.
The code responsible for the bug was never tested by the maintainer - a
trivial test was needed.

> him; the rest is easily parallelized by unstable users.  By running
> unstable, you are volunteering to participate in this testing, and if
> someone has misrepresented Debian unstable to you as a production-ready
Oh, come on! I'm tired of that argument, really. I'm not as dumb as I
might've seen from my mail - I _do_ realize that unstable can have bugs,
trust me. But I also expect the packaging scripts in the .deb to be tested
in a basic way. Nah, I won't be repeating the arguments again - I would be
too obnoxious I think :)

> operating system, they (and you) were misled.  Maintainers are human and
> make mistakes, and even when they don't, things go wrong that are not the
> maintainer's fault.
First of all, note that I am also a maintainer of several packages, so I
suppose I don't fall in the "someone" category. Second of all, the bugs I
mentioned should _never_ happen, even in experimental. They require trivial,
really trivial and obvious tests. If what you're saing is that people can
fire and forget their packages without testing, then I dread for the
Debian's future.

> In the BTS, you state:
> > No, ustable[sic] is not for breaking people's work machines, sorry.
> If your work is incompatible with the nature of unstable, then consider
> running something else on your work machine.
My work is maintaining the few several packages in Debian and working with
software that requires me to run Linux kernel 2.5.x and the latest software
from Sid. And it has nothing to do with my complaint.

> Also from the BTS:
> > 'unstable' doesn't mean 'broken'
> 'unstable' means 'unstable'.
> From WordNet (r) 1.7 [wn]:
>   unstable
>        adj 1: lacking stability or fixity or firmness; "unstable political
>               conditions"; "the tower proved to be unstable in the
>               high wind"; "an unstable world economy" [ant: {stable}]
>        2: highly or violently reactive; "sensitive and highly unstable
>           compounds"
>        3: affording no ease or reassurance; "a precarious truce" [syn:
>            {precarious}]
>        4: suffering from severe mental illness; "of unsound mind"
>           [syn: {mentally ill}, {unsound}]
>        5: disposed to psychological variability; "his rather unstable
>           religious convictions"
>        6: subject to change; variable; "a fluid situation fraught with
>           uncertainty"; "everything was unstable following the coup
>           [syn: {fluid}]
Thanks, I will try to listen more carefully on my English lessons.

> I think definition 3 is most applicable in this case.
I think you missed the point of my mail. Totally.

> I think you have acted disrespectfully in handling this bug, and more so by
> trying to stir up more conflict on debian-devel.
I am not trying to stir anything. I have made notes about possible
inflammatory comments and I have made it clear that flame is not my
intention. Other posters seem to have noticed that, you seem to have missed
not only the whole point of the mail but also those notes I mentioned.

> By the way, your suggestion:
> > if you remove a file which you aren't sure whether
> > it exists or not, it is nice to do
> > 
> > rm -f /path/to/file || true
> > 
> > instead of
> > 
> > rm -f /path/to/file"
> is silly; the entire purpose of the '-f' option to rm is to ignore errors.
Yeah, I'm silly. Fine. If that was the point of your mail, you scored. 


Attachment: pgpCEVNAtW0P6.pgp
Description: PGP signature

Reply to: