1 |
On Feb 16, 2015 8:01 AM, "Patrick Lauer" <patrick@g.o> wrote: |
2 |
> |
3 |
> Right now repoman is relatively permissive - it whines about many things, |
4 |
but |
5 |
> treats many issues as warning. |
6 |
> The result is that many ebuilds get committed with 'minor' cosmetic issues |
7 |
> which then someone more OCD than the original committer cleans up, making |
8 |
> pretty much everyone involved more unhappy. |
9 |
> |
10 |
> There's no reason to not error out on, for example, an invalid RESTRICT. |
11 |
Just |
12 |
> printing a message is relatively useless. |
13 |
> |
14 |
> |
15 |
> Thus I suggest making the following warnings proper errors: |
16 |
> |
17 |
> (Taken from current repoman 'qawarnings' set) |
18 |
> |
19 |
> "changelog.missing", |
20 |
> "changelog.notadded", |
21 |
> "digest.assumed", |
22 |
> "digest.unused", |
23 |
> "ebuild.notadded", |
24 |
> "ebuild.nesteddie", |
25 |
> "DESCRIPTION.toolong", |
26 |
> "RESTRICT.invalid", |
27 |
> "ebuild.minorsyn", |
28 |
> "ebuild.badheader", |
29 |
> "metadata.warning", |
30 |
> "LIVEVCS.stable", |
31 |
> "LIVEVCS.unmasked", |
32 |
> |
33 |
> Most of these have few or no occurrences in the current tree, so changing |
34 |
the |
35 |
> default from warn to error won't get in the way of the normal workflow. |
36 |
> |
37 |
> (A few of them, like DESCRIPTION.toolong, still have about a dozen |
38 |
leftovers, |
39 |
> but that should be easy to fix) |
40 |
> |
41 |
> Have fun, |
42 |
> |
43 |
> Patrick |
44 |
> |
45 |
|
46 |
I would also like to put forward the idea of a pedantic flag, like with |
47 |
GCC, To make all warnings fatal. Or at minimum, make a lesser flag where |
48 |
certain flags that are not normally fatal, would become fatal. |