| 1 |
On Thu, 15 Dec 2005, Ciaran McCreesh wrote: |
| 2 |
|
| 3 |
> On Thu, 15 Dec 2005 00:35:07 -0500 Alec Warner <warnera6@×××××××.edu> |
| 4 |
> wrote: |
| 5 |
> | I've been trying to sit in #gentoo more often ( I figured insanity |
| 6 |
> | would be a good excuse for my crummy grades ) and I am scared by the |
| 7 |
> | fact that people still walk in and try using ACCEPT_KEYWORDS="~blar" |
| 8 |
> | to emerge things. We all know why this is bad, and it's been bad for |
| 9 |
> | quite some time. |
| 10 |
> | |
| 11 |
> | Opinions on adding a warning when users have set this in their |
| 12 |
> | environment? |
| 13 |
> |
| 14 |
> If you're warning on that, you could also warn on far more dangerous |
| 15 |
> and at least equally as common things like silly CFLAGS or having |
| 16 |
> LDFLAGS set. |
| 17 |
|
| 18 |
Well I wouldn't consider those incorrect portage usage. The deal with |
| 19 |
this is we used to recommend ACCEPT_KEYWORDS="~arch" emerge foo. People |
| 20 |
still use this convention; so I hope that a warning will inform people to |
| 21 |
not use it. |
| 22 |
|
| 23 |
I am all for checking if there is a package.keywords file first, and not |
| 24 |
warning them if that file is present. Also it would probably be prudent |
| 25 |
to not warn if --quiet is specified, so that SpanKY can keep his super |
| 26 |
quiet scripts :) |
| 27 |
-- |
| 28 |
gentoo-portage-dev@g.o mailing list |
Archives