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

Re: Buddha Flash support?



Christian T. Steigies wrote:
> > I don't see what this libc has to do with Catweasel. My libc6-2.0.7v still
> > sits in HOLD. But that's not sufficient, you need to recompile dpkg (yikes)
>   ^^^^ good!
> > and a few other things to take advantage of the patch. Note that I can't
> > upload
> > that glibc to incoming any better than you can :-(
> catweasel is only supported in 2.1 kernels (and 2.2 ;-) I think its very
> hard to backport it to 2.0.

I disagree. IDE subdrivers aren't hard to port both ways from what I recall
(unless that weasel is a rare broken case). Someone just has to do it.

> In 2.1 we have the lchown problem, the syscall behaviour has changed (no
> expert, but thats what all this glibc discussion was about)
> your libc6 and my new libgc fix this, but unfortunately we have the other
> installed, where I forgot the patch.
> Now, why does is not work, have you ever tried ln -a foo bar running 2.1
> with both foo and bar non existant? It fails (with the unpatched glibc).

No, I can do that as soon as I have the Mac free to test it on.

> I think the same thing happens when you install packages with dpkg, since
> there are often lots of symlinks in the packages. You can fix this, when you
> first touch every non-existant symlink, but this can be a lot of work for
> certain packages, I tried it...

That might be one of the problems, yes, 

> Are you sure we have to recompile dpkg? I thought the lchown patch fixed
> this, maybe we need a newer ln, anyway it should be consistent for all
> archs, I think that was the problem why the patch was accepted only
> reluctantly.

Positive. The dpkg only knows about chown currently, and uses it. The glibc
patch doesn't change the chown semantics, it just modifies lchown to 
try the lchown syscall, ant try chown as a backup if lchown doesn't exist.
That's how I understood it... For that to work, we need to rebuild dpkg
so it knows about lchown in the first place. 
 
> > ROTFL. Wouldn't it be nice to have working boot floppies, for starters?
> > Let me suggest a deal: you build the new libc, rebuild dpkg, cp, ln (and
> > what
> > else uses chown on symlinks now), and I build base2_ when you're done?
> 
> > Waiting ...
> Well, very bad news... I am bitten by the substvars problem, cant build
> packages because first they fail right after unpacking (substvars, no such
> file) and after "fixing" that they fail short before building the package
> with the same error. I dont know it its a bug or a feature, and Im not sure
> where it comes from...

I'm sure it's a feature. So what you say is I should not upgrade my system
to be safe ? :-)

I'll send a packages list for you to compare ...

	Michael


Reply to: