| 1 |
On Tue, Aug 08, 2006 at 12:55:28PM +0200, Enrico Weigelt wrote: |
| 2 |
> * Roy Marples <uberlord@g.o> schrieb: |
| 3 |
> > On Tuesday 08 August 2006 09:56, Enrico Weigelt wrote: |
| 4 |
> > > If you want an dhcp client, install "dhcp-client", if you |
| 5 |
> > > want an dhcp server, install "dhcp-server". Could it be simpler ? |
| 6 |
> > |
| 7 |
> > Maybe you missed the part of the discussion where we thought that |
| 8 |
> > maintaing 3 ebuilds vs 1 ebuild was a bad idea. Yes we would need |
| 9 |
> > 3 due to the way that the dhcp builds and installs. |
| 10 |
> |
| 11 |
> Okay, but they're maintained at the same time. |
| 12 |
> |
| 13 |
> Let's see where the extra work could come from: |
| 14 |
> |
| 15 |
> + changes in build options. okay, have to type some things twice. |
| 16 |
> adds 5mins |
| 17 |
> + three packages have to be tested now. today one package has to |
| 18 |
> be tested in three variants. is there really more work ? |
| 19 |
> |
| 20 |
> The 3rd package is mostly copy-and-paste, since doesn't actually |
| 21 |
> do anything. It's just rdeps based on useflags. Just an multiplexer. |
| 22 |
> |
| 23 |
> On the other hand I see some more changes on an split: |
| 24 |
> Let's say, in a newer version, there's an interesting improvement |
| 25 |
> in the server, but an bad bug in the client. Currently the client |
| 26 |
> would block the server, just for buerocrativ reasons. |
| 27 |
> After a split, both part-packages can evolve separately. |
| 28 |
|
| 29 |
To do that, you have to seperate any libs used between the two. In |
| 30 |
such a pkg, there *should* be a common lib- so you're suggesting |
| 31 |
either static linkage of said code (disk but more importantly mem |
| 32 |
bloat), or so renaming (further divergance from upstream, more issues |
| 33 |
in glsa handling). |
| 34 |
|
| 35 |
Yet *more* manual work. |
| 36 |
|
| 37 |
You want this, implement it in an overlay. |
| 38 |
|
| 39 |
You get what you want, and if you manage to make it not suck the big |
| 40 |
one, hey, maybe you might convince a few people. |
| 41 |
|
| 42 |
Either way, people aren't going to yield- put in the work to prove |
| 43 |
them wrong rather then just trying to talk them into the ground |
| 44 |
please. |
| 45 |
|
| 46 |
Besides... pushing this hard for something, you better be willing to |
| 47 |
do the work yourself- can't expect others to do what you want when |
| 48 |
they disagree. |
| 49 |
|
| 50 |
~harring |
Archives