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

Re: dpkg-cross: dh_strip & make-kpkg issues



>> Kernel developers won't change anything unless a HUGE reason exists.
>> They definitly don't (and shouldn't) care about dpkg-cross.
> What I intent to say is that it is not unusual that well known projects do
> not like to accept the conventions. (Please do not blame me while saying
> that!)

I'm afraid there is too many of non-technical issues here. Kernel developers
and FSF have quite different points of view on how things should be
called / done / etc.

>> 'mode', 'scope' and 'package' are orthogonal, not scoped.
> For my understanding. I can write the following cross-compile
> configuration:
> 
>   mode testmodeA:
> 
>   scope makeflags:
>   package pkg10:
>     WHATIAM=var-for-pkg10-for-makeflags-for-testmodeA
> 
>   scope environment:
>   package pkg10:
>     WHATIAM=var-for-pkg10-for-environment-for-testmodeA
> 
>   mode testmodeB:
> 
>   scope makeflags:
>   package pkg10:
>     WHATIAM=var-for-pkg10-for-makeflags-for-testmodeB
> 
>   scope environment:
>   package pkg10:
>     WHATIAM=var-for-pkg10-for-environment-for-testmodeB
> 
> Is it not the way dpkg-buildpackage is working if -M is used ?
> 
> For you 'architecture XXX:' should not be scoped by mode. But this would
> make it less flexible for my point of view.

Currently, none of 'package', 'mode' or 'scope' is scoped by each-other.
These 3 settings define a point in 3-dimension universe, on which
subsequent variable definitions will reside.

When you change one of dimensions, others stay unchanged.
So the above example may be written as

package pkg10:
mode testmodeA:
scope makeflags:
  WHATIAM=var-for-pkg10-for-makeflags-for-testmodeA
scope environment:
  WHATIAM=var-for-pkg10-for-environment-for-testmodeA
mode testmodeB:
  WHATIAM=var-for-pkg10-for-environment-for-testmodeB
scope makeflags:
   WHATIAM=var-for-pkg10-for-makeflags-for-testmodeB

Althouth this may be not a good sequence for understanding, I believe better
examples exist.

>> However, we may postpone this until a real-world example is found.
> Yes, we should wait until a request.

Ok



Reply to: