| 1 |
On Wed, Nov 25, 2009 at 04:49:17PM -0800, Zac Medico wrote: |
| 2 |
> Ciaran McCreesh wrote: |
| 3 |
> > On Wed, 25 Nov 2009 23:59:45 +0100 |
| 4 |
> > Ulrich Mueller <ulm@g.o> wrote: |
| 5 |
> >> Real examples would be issues like bugs 83877 [1] or 263387 [2]. |
| 6 |
> >> Nothing that could be easily dismissed or worked around. Both issues |
| 7 |
> >> are fixed with Portage since a long time. |
| 8 |
> > |
| 9 |
> > Yes, those are examples of packages relying upon something that is |
| 10 |
> > undefined behaviour, and that behaves differently depending upon the |
| 11 |
> > Portage version you use. |
| 12 |
> > |
| 13 |
> >> I don't know of any example where non-preservation of nanosecond |
| 14 |
> >> timestamps would cause problems. |
| 15 |
> > |
| 16 |
> > Not non-preservation. Partial and inconsistent corruption. |
| 17 |
> |
| 18 |
> Wouldn't "loss of precision" be a more accurate description? Of the |
| 19 |
> known packages which require timestamp preservation, do any of them |
| 20 |
> use sub-second precision in their timestamp comparisons? |
| 21 |
|
| 22 |
This discussion in generall is daft. No package can rely on |
| 23 |
nanonsecond resolution for installation because the most common FS out |
| 24 |
there (ext3) does *second* level resolution only. As such, I can |
| 25 |
pretty much gurantee there is *zero* packages out there that require |
| 26 |
nanosecond resolution for installation. |
| 27 |
|
| 28 |
It's an academic discussion, and pointless. We don't mandate the |
| 29 |
filesystems PMS implementations are run on- as such we cannot make a |
| 30 |
gurantee to ebuilds that nanosecond resolution is available. It's |
| 31 |
daft to encode in the spec NS resolution when it's essentially |
| 32 |
impossible to gurantee it- I'm honestly not sure why we're having this |
| 33 |
discussion beyond the "python sucks" angle. |
| 34 |
|
| 35 |
~brian |
Archives