1 |
While reading tosay's PMS draft, i've seen the following (§9.2.1), which |
2 |
reminded me this old discussion: |
3 |
|
4 |
> = Exactly equal to the specified version. Special exception: if the |
5 |
> version specified has an asterisk immediately following it, a string |
6 |
> prefix comparison is used instead. When an asterisk is used, the |
7 |
> specification must remain valid if the asterisk were removed. (An |
8 |
> asterisk used with any other operator is illegal.) |
9 |
|
10 |
This ratifies the current substring matching semantics, and excludes the |
11 |
"=foo/bar-1.2.*" syntax addition that has been suggested in this thread. |
12 |
|
13 |
Re-reading the thread, i'm not sure what the conclusion was, if any. |
14 |
If there was a consensus on this, then it's all good. But if not, then |
15 |
maybe it's still time to change it without waiting for EAPI-1. |
16 |
|
17 |
Imho, it would be a good thing to drop the substring matching and |
18 |
replace it with a "1.2* is 1.2(.Y)*(_suffix)?(-rY)?" semantics. As |
19 |
already pointed, only a tenth of ebuilds would need to be fixed if |
20 |
this semantics restriction was adopted, so it sounds doable. But sure, |
21 |
that's just my opinion, and i have no strong argument but that it sounds |
22 |
closer to what most people need, and intuitivly expect. |
23 |
|
24 |
</friendly reminder> |
25 |
|
26 |
-- |
27 |
TGL. |
28 |
-- |
29 |
gentoo-portage-dev@g.o mailing list |