| 1 |
>>>>> On Sun, 24 Mar 2013, Micha³ Górny wrote: |
| 2 |
|
| 3 |
>> >> > +multibuild_merge_root() { |
| 4 |
>> >> > + [...] |
| 5 |
>> >> > + if type -P lockf &>/dev/null; then |
| 6 |
>> >> > + # On BSD, we have 'lockf' wrapper. |
| 7 |
>> >> > + tar -C "${src}" -f - -c . \ |
| 8 |
>> >> > + | lockf "${lockfile}" tar -x -f - -C "${dest}" |
| 9 |
>> >> > + else |
| 10 |
>> >> > + [...] |
| 11 |
>> >> > + cp -a -l -n "${src}"/. "${dest}"/ |
| 12 |
|
| 13 |
>> Maybe explicitly testing for USERLAND would be less fragile? Or use |
| 14 |
>> tar everywhere? |
| 15 |
|
| 16 |
> I prefer something that can boom into face and let us know that it |
| 17 |
> is broken rather than silent, poor catch-all. |
| 18 |
|
| 19 |
How about adding a couple of "assert" and "die" to tar and cp then? |
| 20 |
|
| 21 |
Ulrich |
Archives