| 1 |
On Tue, 6 Aug 2013 20:44:57 +0100 |
| 2 |
Ciaran McCreesh <ciaran.mccreesh@××××××××××.com> wrote: |
| 3 |
|
| 4 |
> On Tue, 6 Aug 2013 15:31:14 -0400 |
| 5 |
> Alexis Ballier <aballier@g.o> wrote: |
| 6 |
> > Well, ok, but this doesn't relate to what I was writing. Subslot, or |
| 7 |
> > slot emulators or whatever, in their current usage with := |
| 8 |
> > dependencies, are not fine grained enough for some use cases. Those |
| 9 |
> > cause regressions if used improperly. |
| 10 |
> |
| 11 |
> There is no regression. Previously, packages sometimes broke when |
| 12 |
> doing an upgrade. Now, packages do not break when doing an upgrade. |
| 13 |
|
| 14 |
The regression is the useless rebuild. Without preserve-libs, packages |
| 15 |
break even more: cf the libc example. |
| 16 |
|
| 17 |
> > > You just make the ebuilds install different bits. In effect you |
| 18 |
> > > emulate a simple subset of how parts would do it. |
| 19 |
> > |
| 20 |
> > Which needs patching to be done properly... unless you are |
| 21 |
> > suggesting to build it twice and throw away whats not needed just |
| 22 |
> > to workaround subslots limitations. |
| 23 |
> |
| 24 |
> It's up to the relevant developers to decide how much work they're |
| 25 |
> willing to put in to save some users a bit of CPU time. |
| 26 |
|
| 27 |
It's up to those that want to force this on developers to do the |
| 28 |
relevant work so that it can be done properly. |
| 29 |
|
| 30 |
> > Your argumentation is basically 'Other parts are doing it wrong so |
| 31 |
> > it's ok to add some more to it'... We're back a dozen emails back, |
| 32 |
> > aren't we? |
| 33 |
> |
| 34 |
> It's not adding more to it. It's avoiding eliminating a tiny portion |
| 35 |
> of it. Even if you subscribe to the notion that unnecessary rebuilds |
| 36 |
> are a relevant problem, there's no point in caring about the |
| 37 |
> occasional unnecessary rebuild due to overly strict dependencies when |
| 38 |
> most unnecessary rebuilds are caused by something else entirely. |
| 39 |
|
| 40 |
And here we go again... |
| 41 |
|
| 42 |
> > It was meant as an example and has nothing to do with dependency |
| 43 |
> > resolution. The above exercise is something extreme but that we have |
| 44 |
> > to solve; preserve-libs has proven to be correct enough. You have |
| 45 |
> > yet to show a correct, in your sense, solution. |
| 46 |
> |
| 47 |
> The correct solution is heavy slotting. And I'd hardly consider |
| 48 |
> "intermittently introduces invisible security holes and causes |
| 49 |
> unbootable systems" to be "correct enough"... |
| 50 |
|
| 51 |
There is no difference between heavy slotting and preserve-libs in this |
| 52 |
case. |
Archives