1 |
Neil Bothwick <neil <at> digimed.co.uk> writes: |
2 |
|
3 |
> |
4 |
> On Sun, 26 Apr 2015 15:49:19 +0200, Alan McKinnon wrote: |
5 |
|
6 |
> Yes, I agree it needs a full overhaul of the way it communicates what is |
7 |
> happening. |
8 |
|
9 |
|
10 |
One simple fix is this blockage of packages that do not need to be blocked. |
11 |
When I wait more than 5 or 6 days to update, I often get many packages to |
12 |
build, running something like 'emerge -uDNvp world'. Yes I know I can sit |
13 |
there and go through every package with --ask, but that is a huge |
14 |
waist of time, particularly if you poll user desires. So I find all |
15 |
the packages (ugly little cut|awk|sed|paste script) and install everything |
16 |
that is not really block. Source && update and then concentrate on the |
17 |
actual blockage. I'd be surprised if this is anything other than a trivial |
18 |
fix to portage (hello no, I'm not working on portage!; Zac is very smart and |
19 |
capable and look how he struggles...). |
20 |
|
21 |
|
22 |
But our reticent developers think that is good for the rank and file |
23 |
gentoo users; other see this sort of situation as obstinate. |
24 |
|
25 |
So a simple improvement is to allow what can install, to install, in a |
26 |
semantic where folks do not have to sit and watch the screen to keep the |
27 |
updates rolling along; then leave the true blockage for close attention. |
28 |
I.E. blockage should only *ever* effect the minimal of what has to be |
29 |
blocked during an update. |
30 |
|
31 |
|
32 |
YMMV_hth, |
33 |
James |