1 |
On Sat, Oct 10, 2015 at 8:16 AM, Brian Dolbec <dolsen@g.o> wrote: |
2 |
|
3 |
> On Sat, 10 Oct 2015 14:24:28 +0200 |
4 |
> Fabian Groffen <grobian@g.o> wrote: |
5 |
> |
6 |
> > On 10-10-2015 14:19:44 +0200, hasufell wrote: |
7 |
> > > > +RDEPEND=" |
8 |
> > > > + !libressl? ( dev-libs/openssl:0 ) |
9 |
> > > > + libressl? ( dev-libs/libressl ) |
10 |
> > > > + sys-libs/zlib |
11 |
> > > > + net-libs/http-parser |
12 |
> > > |
13 |
> > > Please order deps alphabetically (I know I added libressl without |
14 |
> > > reordering, but that was just to keep the diff as small as |
15 |
> > > possible). |
16 |
> > |
17 |
> > Is this a policy or a preference? In case of the first, have repoman |
18 |
> > do it for you. |
19 |
> > |
20 |
> > Fabian |
21 |
> > |
22 |
> > |
23 |
> |
24 |
> Please NO, repoman's code is still a mess, the last thing it needs is |
25 |
> adding a bunch of code messing with ebuilds directly. Reporting is one |
26 |
> thing, modifying ebuilds quite another. |
27 |
> |
28 |
|
29 |
At Google the go team spent a bunch of time building different tools for |
30 |
these purposes (which share code of course.) |
31 |
|
32 |
So there is a: |
33 |
go compiler # compiles your code |
34 |
go cmt # properly formats your code |
35 |
go vet # finds common anti-patterns in your code |
36 |
|
37 |
I don't see why someone could not write an 'ebuild vet' or 'ebuild fmt' |
38 |
tool; it doesn't have to start as a strict component of repoman at all |
39 |
(although ideally it ends up there to be run automatically.) |
40 |
|
41 |
-A |
42 |
|
43 |
|
44 |
> |
45 |
> -- |
46 |
> Brian Dolbec <dolsen> |
47 |
> |
48 |
> |