| 1 |
On Wednesday 26 September 2007, Donnie Berkholz wrote: |
| 2 |
> On 04:51 Wed 26 Sep , Mike Frysinger wrote: |
| 3 |
> > On Wednesday 26 September 2007, Donnie Berkholz wrote: |
| 4 |
> > > for USE in ${IUSE}; do |
| 5 |
> > > [[ ${USE} = linguas_* ]] && use_linguas ${USE} |
| 6 |
> > > done |
| 7 |
> > |
| 8 |
> > you dont want to use the variable 'USE' ... while it would work |
| 9 |
> > correctly if you declared it with 'local', it'd be confusing and |
| 10 |
> > repoman may just barf on it ($USE is supposed to be readonly for |
| 11 |
> > ebuilds) |
| 12 |
> |
| 13 |
> Ahh, reviewing is so fun and educational, it almost feels like I'm |
| 14 |
> watching the Discovery channel. Although I'm not sure why $USE is |
| 15 |
> directly accessible from ebuilds ... seems like it should only be |
| 16 |
> accessible to use(). Implementation detail, or should it be part of the |
| 17 |
> PMS? |
| 18 |
|
| 19 |
if it's all common shell, it's pretty hard to hide ... you could try some |
| 20 |
tricks like rewriting USE to __USE__ so internal code (like the `use` |
| 21 |
function) still work, but in the end, does that really help ? |
| 22 |
-mike |
Archives