[TAG] chown -- giving files away

[Jay R. Ashworth]

On Sat, Jun 12, 2004 at 02:51:19PM -0400, Ben Okopnik wrote:
> > > There are too many "magic" factors here, Jay. If you want to tell us
> > > about the whole situation, great. Otherwise, I'm just making guesses
> > > that are going to be shot down everytime, by "oh, forgot to tell you:
> > > this skateboard is supposed to fly at escape velocity" problems.
> > 
> > Ben, I've given every available detail in damned near every posting.  
> 
> Like
> 
> + Nope, cause that is not going to cause the menu porogram *not to try*
> + to chown the file, nore will it cause that *to suddenly work*.
> + MenuMaestro actually *checks* the return from the function call, and
> + pukes if it didn't work.  Fancy that.  *Good* error checking
> + programming is making my life hell...
> 
> which now fills in (by *implication*, that is) the missing details?
> Jay, I'm not there to see this stuff. Perhaps I've missed that it was
> chown(2) in your earlier emails, which had me guessing that your menu
> system is a script that you can tweak, but a deep problem like this
> requires lots of info - perhaps even more than you expect, even from a
> 2-decade perspective.

I went back and looked; I didn't see any germane information that was
missing from my inital posting...

> Whatever, I didn't feel like I had enough relevant info to go any
> further. Getting huffy instead of just giving up the info is, shall we
> say, suboptimal; you know the standard choices.

Sorry, but as far as I could see, you were asking questions I had
already (gone out of my way to) answer.

> > I
> > don't *do* 'magic' error reports; I've been doing this for damned near
> > 2 decades.  If you think I've left something out, you tell me where.
> 
> Uh, no. That's not the way it works. I don't know the scope of the
> problem; you do. It's up to you to describe the entire thing including
> pertinent external info, and up to me to ask for the details *within*
> that scope; finding out that you omitted part of the overall description
> restarts the entire process. What you gave, from my perspective, were
> some details but left the entire picture hidden, and I don't have much
> patience for pulling worms out of their holes unless I'm getting paid
> for it.

We will, I guess, have to agree to disagree; I really do feel like I'd
provided all the information necessary to answer the question "is there
a way to make linux chown(2) permit giving away files", and to make
clear that that was the question I actually needed to get an answer to.

> > But I've just spent 10 hours at a site chasing down what turned out to
> > be enemy action against someone's carrier configuration on their
> > telemarketing system, and I'm less than thrilled to be told I'm not
> > giving clear enough error reports.
> 
> And your frustration gets carried over onto my trying to help you
> because?... Suboptimal behavior mode, as I've said.

My frustration causes me to be less tolerant of people asking
questions I've made a point to already have answered.  Or so it seemed
at the time.

> > chown(2) won't allow the giving away of files as it's supposed to within
> > Linux-ABI emulation when called from a SCO format binary; how opaque
> > *is* that?
> 
> It isn't, if it's a wishlist item. If you're asking about how to resolve
> a real-life problem by getting around it, it requires a *complete*
> *description* *of* *the* *problem*, including which part of the call
> you're trying to hack. 

And oddly, if you look, you'll see that I provided that in my initial
message.

> At this point, it's sounding like you need to hack libc and LD_PRELOAD
> it. I know it's not much help, but that's what it looks like from here.

"There is no defined way to make chown(2) permit the giving away of
files."  That was really enough answer.  I don't object to "you can't
get there from here."

> Ask on the libc developers list, maybe; I'd imagine they had a
> discussion about it before implementing a relatively major change of
> that sort. Perhaps they even left some sort of an option for invoking
> the old behavior.

Naw; if I gotta hack something, I'll hack lABI to do what it's supposed
to.

Thanks anyway, dude.

Cheers,
-- jra
-- 
Jay R. Ashworth                                                jra at baylink.com
Designer                          Baylink                             RFC 2100
Ashworth & Associates        The Things I Think                        '87 e24
St Petersburg FL USA      http://baylink.pitas.com             +1 727 647 1274

                2004 Stanley Cup Champion Tampa Bay Lightning