Failed to make uClibc 0.9.26

Christian MICHON christian.michon at gmail.com
Wed Dec 20 23:36:40 UTC 2006


On 12/20/06, Rob Landley <rob at landley.net> wrote:
> > there could be many reasons:
> > 1) faster build
>
> You benchmarked it then?

oh yes. way before 0.9.26 actually. I guess Erik was actually a
recipient of some of my results (around 2001/2002 ?)

> > 2) less bloat
>
> The point of the config options is so the new versions can be about as small
> as the old ones.  If they're not (with the same set of features enabled),
> this is a bug we'd probably like to know about.

you do not get me right on this. "locale" is a nice to have, but it's still
way too big (hence bloat).

> > 3) smaller size
>
> Do you know this for a fact, or are you guessing?  And isn't this a repeat of
> #2?

a bit of both. but if you consider my clue about "locale", it definitely should
be smaller.

> > 4) code more stable (at least less prone to change)
>
> Any given snapshot won't change unless you update it, but 0.9.26 was
> incredibly buggy and lots of packages are known not to work against it, and
> most of the known bugs have since been fixed.

sorry, I've not been active on this list since my last contributions to uclibc
(since then, life has been crazy, not very much time for coding to spare:
this is not my main job, unlike you). What I meant is when the addition
of new features is frozen (a release number is given), it easier to fix bugs
than to keep adding features at the same time.

> > most important:
> > what if he's trying to fix a bug in an existing product linked with
> > this particular version ?
>
> Then that would have answered my question.  I asked _why_ he was doing it.

I'm surprised that even though you're an adept of scm and released products,
and on top of that virtualization (we're actually members of many common
dev-lists), you're missing my point here.

When we do return to an old version, in the "real" world (understand: industry),
it's because we've little choice. But maybe Jack the original poster should
answer.

> > who are you to judge everyone so harshly recently ?
>
> You're the one who interpreted a question as an attack.  To answer your
> question: I'm a volunteer, participating here for free, and I owe you
> nothing.  However, since it seems to bother you: welcome to my spam filter.
> I'll never even see one of your messages again (even sent directly), so you
> won't be bothered by replies from me.

I was referring to the recent events related to busybox and Bruce Perens
(hence lead me to your development notes and blog).

You owe one thing to the policy of this list: being courteous and helpful
especially to newbies. I know for sure there never was a split between
uclibc developers and users mailing lists. We're all volunteers here too.
And as I say, your tone has changed over the last few months...

Your spam filtering measure falls flat on me. I'll keep a watch at least
on your toybox (for I'm not revengeful and know we're actually sharing
a lot of common aims).

-- 
Christian



More information about the uClibc mailing list