| 1 |
On Wed, 2005-16-03 at 21:03 -0800, Robin H. Johnson wrote: |
| 2 |
> On Thu, Mar 17, 2005 at 01:29:23PM +0900, Georgi Georgiev wrote: |
| 3 |
> > Which is one of the reasons why I am behind the idea of *unique* package |
| 4 |
> > names, even if that involves some redundand naming. It's for the greater |
| 5 |
> > good. The package name being tied to its category is evil and it also |
| 6 |
> > means that we'll never see multiple categories per package, or more |
| 7 |
> > descriptive category names, or any other of these goodies. Ever. |
| 8 |
> And what about packages that have the same name upstream, and yet do different |
| 9 |
> things? It's a nice concept, but not practical. |
| 10 |
> |
| 11 |
> find $(<profiles/categories) -maxdepth 1 -mindepth 1 ! -name CVS -printf '%h %f\n' | sort -k2 | uniq -f1 -dD |
| 12 |
> Shows we have 145 packages with non-unique names. |
| 13 |
> |
| 14 |
> We've even got a few cases where there are 3 applications with the same name |
| 15 |
> upstream: |
| 16 |
> app-arch/par |
| 17 |
> app-text/par |
| 18 |
> dev-util/par |
| 19 |
> (2 other examples of the same thing as well). |
| 20 |
|
| 21 |
Don't we already have rules that forbid two packages having the same |
| 22 |
name ? |
| 23 |
|
| 24 |
-- |
| 25 |
Olivier Crête |
| 26 |
tester@g.o |
| 27 |
x86 Security Liaison |
Archives