Re: 3.0 tests depend on libc >= 2.3.5?

On Mon, 2005-07-25 at 16:17 -0600, Matt Taggart wrote:

> Well I think something good came out of this firedrill and that is I think 
> it may be wrong for alien to be adding library dependencies when converting 
> LSB rpms. Adding those dependencies might be correct for the system that 
> alien is being run on, but could be wrong for a different version of that 
> system (example: converted on etch, broken on sarge) or a different system 
> (example: converted on ubuntu, broken on debian). I also think it's more 
> than just versioned depends, but the dependencies themself, since the 
> package namespaces aren't necessarily the same.

If an LSB package can be properly recognized, I agree.  The LSB model,
after all, is that you not depend on individual items but on the overall
LSB-required provides, which are supposed to imply the correct
functionality is present on the system (lsb-core and lsb-graphics).

> > I thought that was determined by seeing the appropriate lsb depend
> > (lsb-core these days).
> 
> Yeah that's how it works (still using "lsb" not "lsb-core" maybe a bug 
> needs to be filed?).

"lsb" was only correct for 1.x, since the module splitting it's 
"lsb-core" at a miniumum, so that would be the thing to check for.