| 1 |
On Tue, Jan 15, 2013 at 8:44 AM, Ian Stakenvicius <axs@g.o> wrote: |
| 2 |
> I wonder if it might be pertinent for future portage's to install an |
| 3 |
> alias command, "emerge-system-update" or similar, that would wrap the |
| 4 |
> standardly accepted emerge update command more or less everyone |
| 5 |
> already runs.. easier end-user experience as they don't need to |
| 6 |
> learn/remember all the little fiddly options, plus portage dev's can |
| 7 |
> control the default (probably we make it over-ridable via a make.conf |
| 8 |
> var or something) so that if things change with newer EAPIs or portage |
| 9 |
> features the default flags can be adjusted too... |
| 10 |
|
| 11 |
I'm fairly confident that this was already discussed a while back, and |
| 12 |
generally had positive feedback. |
| 13 |
|
| 14 |
I'd make it easy-to-remember though - either a one-letter option, or |
| 15 |
something short. |
| 16 |
|
| 17 |
To avoid rehashing the same arguments over again a few things to note: |
| 18 |
1. The settings would be reasonably conservative - intended for |
| 19 |
newbies and such and unlikely to break things. |
| 20 |
2. We would not re-use any existing parameter - no changes in behavior/etc. |
| 21 |
3. Users could still throw alphabet soup at portage if they already |
| 22 |
know the one-true-way (TM). |
| 23 |
4. Script writers would be warned up-front that the behavior of this |
| 24 |
feature would be subject to change without much notice. |
| 25 |
|
| 26 |
Is there any reason that this can't just be done? It would only be an |
| 27 |
alias, so it shouldn't really require development per-se. |
| 28 |
|
| 29 |
Some open questions: |
| 30 |
1. What is the correct use-flag behavior - -N, or --reinstall=changed-use? |
| 31 |
2. What is the correct --with-bdeps behavior? |
| 32 |
3. What is the correct --deep behavior? |
| 33 |
|
| 34 |
Now let the bikeshedding commence... |
| 35 |
|
| 36 |
Rich |
Archives