1 |
I don't know if the discusion of the recently begun portage v2 rewrite |
2 |
deserves it's own list or not, but I thought I'd start discussing it |
3 |
here, and we can decide later if we need to move it. |
4 |
|
5 |
For any gentooers haven't heard about v2, please be aware that it is |
6 |
only just beginning. I'm sure the current portage will go through many |
7 |
new releases before the v2 codebase is ready. |
8 |
|
9 |
Anyway, I just read portage-core.py and wanted to discuss it some: |
10 |
|
11 |
I think we should ditch the non-validating codepath, at least for now. |
12 |
I'm sure it's probably slightly faster, but duplicating this code seems |
13 |
like a bad idea with regards to maintenance. If we need the speed |
14 |
later, we can optimize, perhaps by duplicating code, but preferrably |
15 |
some other way. |
16 |
|
17 |
I think I like the proposed new dependency syntax. It doesn't seem as |
18 |
flexible as another that was proposed earlier on this list (by karltk? |
19 |
sorry, don't remember). Are there good reasons to use this less |
20 |
flexible scheme, such as readability or something? Or is my premise |
21 |
incorrect? |
22 |
|
23 |
Is there a reason to not use "1.0-*" instead of introducing another |
24 |
special char in the syntax, "~1.0"? |
25 |
|
26 |
Ok, I guess that's all I have at the moment. I suppose we should start |
27 |
discussing how to split up the work that needs to be done. Maybe |
28 |
drobbins will just tell us each what to do. :-) |
29 |
|
30 |
Oh, I think we should start writing test code for these new classes |
31 |
immediately, and maintain them as we go along. I think python has some |
32 |
good support for regression test sorts of things, but I'm not very |
33 |
familier with it yet. Better go read... :-) |
34 |
|
35 |
--Chouser |