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

Reality check.

On Tue, Oct 07, 2014 at 10:09:10PM +0200, Matthias Urlichs wrote:
> Adam Borowski:
> > > The only acceptable concrete value for 'extremely few' is Zero.
> > 
> > I'd say losing patience is quite understandable in this case
> Probably. However, the context of this thread was not at all about a
> maintainer who refused to apply a perfectly sensible patch. Getting
> confronted with language like that, out of the blue, is no longer
> something I'm willing to accept here.

Perfectly reasonable; I entirely concur his language was unacceptable.
He clearly is not a native speaker, but I must confess that I have
probably wound him up a bit too many times about "debilian" (he winds me
up in turn about "craptoo", mind;) and recently he's found it much more
difficult to respond as he sees it lose its broad-based applicability.

We're both embarassed about the lack of POSIX.2 support by default, since
it's hardly onerous by comparison to what else is getting pulled in
nowadays. OFC there's more control over that with Gentoo, ime; on the
flipside it's not as convenient.

> A decade or so ago, this kind of thing was regarded as normal on our
> mailing lists. We mostly-succeeded in civilizing the place, and frankly I
> like them much better that way.
> Slippery slope, and all that.

Speaking of slippery slopes, one would hope that a community software
project would take the opportunity to _address the concern_ raised.
In this case, that would be an unreasonable restriction, such that
/usr/bin is not in PATH, despite it being mounted in initramfs. This
leads to the problem that the POSIX.2 compliance of the "Universal" OS
is not even a second-class citizen, let alone a default, as one might
reasonably expect.

It's no good just pontificating about language, and ignoring the concern,
any more than it is to say "this happened in 2009, get over it," a
position seemingly out of touch with the reality that the maintainer is
forced to deal with this nonsensical restriction every time he deals
with debian.

For all Linux distros try to mock POSIX, it's both what got you your OS
in the first place, and also what every professional programmer MUST know
how to code to, as a basis. Do you really think all those companies
spend so much time and money on compliance for the sake of it?

I hope you don't throw out the baby with the bathwater, since mirabilos
has an excellent reputation as a programmer, is responsive to users and
holds to high standards both professionally and personally. I'm not sure
what the rest of you do; ime most distro "developers" are in fact
script-writers, and not very good ones. The senior ones usually learnt
from TLDP/ABS, widely panned in #bash and across IRC, except by newbs.

Certainly your "senior people" seem to be focussed on entirely the wrong
sets of things as far as professional development is concerned; to wit:


Because he apparently does not like the upstream (as I was told in #awk,
before being given that bug url, in relation to a convi about cclasses in
mawk), the only thing your "professional" "developer" responds to is
comments about the delay after 5 years, and then only to be snide.
This has "been going on since 2008." Extraordinary.

Still, it's a good thing it's on his todo list.
So that's all right, then.  /s

Feel free to flame me if that makes anyone feel better. It won't change
the reality that your distro is losing focus. (I'm going on holiday
tomorrow in any case, and regularly ignore email for weeks on end in
any case.) Pretty soon all that will be left is a bunch of oh-so-polite
corporate employees who all work/ed for Red-Hat, but have no passion
for the craft, just the stock-option.

Then you can turn it into fedora-lite, and become a monoculture,
enforcing dbug to get round the GPL by classifying localised RPC as
IPC, and _pretending_ that's not an API call.[1]

Oh wait.

Have fun arguing about who sits where as the ice flows nearer, and
the POSIX lifeboats float off into the distance.


PS @mira: And you wonder why I call it the "Unipolar OS"?

[1] One can see the pattern in "GPL" libs that _only_ provide a dbus
interface, and the "API" is in fact a dbus wrapper. Hence the
much-vaunted "success" with the embedded industry, who ofc *love*
Till the question of how a practioner views that call comes before
a Judge, at least.
#friendly-coders: We're friendly, but we're not /that/ friendly ;-)

Reply to: