1 |
On Sat, 2012-09-22 at 09:55 +0200, Michał Górny wrote: |
2 |
> Hello, |
3 |
> |
4 |
> The current dependency syntax: |
5 |
> |
6 |
> [VERSION-OP] PACKAGE-NAME ["-" PACKAGE-VERSION] |
7 |
> |
8 |
> suffers a few problems: |
9 |
> |
10 |
> |
11 |
> 1. It is not really human-friendly. |
12 |
> |
13 |
> People don't say things like: |
14 |
> |
15 |
> I need newer than monkey-1.2. |
16 |
> |
17 |
> They say instead: |
18 |
> |
19 |
> I need monkey, newer than version 1.2. |
20 |
> |
21 |
[snip :/ ] |
22 |
|
23 |
> 4. It follows the syntax used by bash (for conditionals), pkg-config |
24 |
> -- it is more natural in the environment. |
25 |
> |
26 |
|
27 |
The BIG problem with that is bash has nothing to do with evaluating |
28 |
dependencies. All bash does is source the *DEPEND and pass the value to |
29 |
the package manager which does all the processing. And all 3 current |
30 |
package managers are set up to parse those dep strings with a set syntax |
31 |
and whitespace. None of the PM's dependency resolvers are written in |
32 |
bash, two are python based, one C++. This proposal would throw a big |
33 |
monkey wrench into parsing those strings. Introducing lots of bugs, |
34 |
both in the PM and the ebuilds. |
35 |
|
36 |
And this after all the fuss about the unified DEPENDENCIES proposal, |
37 |
which is a small syntax change for the current processing code, easily |
38 |
incorporated into the PM's. AND has definite, measurable advantages. |
39 |
|
40 |
so, -1 |
41 |
-- |
42 |
Brian Dolbec <dolsen@g.o> |