1 |
Hi! |
2 |
|
3 |
On Sun, 26 Apr 2015, Jeroen Roovers wrote: |
4 |
> On Sun, 26 Apr 2015 08:04:11 -0400 |
5 |
> Rich Freeman <rich0@g.o> wrote: |
6 |
> > If we're talking about human parsing, can you give an example of how |
7 |
> > variation makes your life more difficult today? I'm just trying to |
8 |
> > understand what we're trying to fix... |
9 |
> |
10 |
> Reading through hundreds of Summaries. If the atoms and the request |
11 |
> variant are always in the same place, parsing by humans is MUCH quicker. |
12 |
|
13 |
Look at security bugs that ask for stabilization with the goal of |
14 |
having a GLSA. The bugs all look the same and I wish all |
15 |
stabilization and keywording bugs adopted the same format. |
16 |
|
17 |
There is one corner case where that format is not enough: |
18 |
multiple ebuilds/versions with non-homogenic archs, i.e.: |
19 |
|
20 |
cat-egory/packageA-1.2.3 amd64 x86 alpha |
21 |
cat-egory/packageB-2.3.9 amd64 alpha |
22 |
cat-egory/packageC-3.99 amd64 x86 ppc64 |
23 |
cat-egory/packageD-10.2.5a alpha |
24 |
|
25 |
The format I used here seemes to be slightly more common than |
26 |
others and it is good enough for me™. |
27 |
|
28 |
Any add-on prose should be _after_ the standardized bit. |
29 |
|
30 |
And while we're talking about ponies: |
31 |
|
32 |
Let's make it trivial for the requester to also specify the |
33 |
_correct and complete_ list of per-arch dependencies that also |
34 |
need to be tested and keyworded. It is one of the most annoying |
35 |
things about stabilization bugs: having to hunt dependencies. |
36 |
|
37 |
Side note: please make sure to include all FEATURES=test gated |
38 |
dependencies, too. The prose should also mention if there are |
39 |
circular ones (I'm looking at you, dev-ruby/*). |
40 |
|
41 |
|
42 |
There, now that's off my chest. |
43 |
|
44 |
Regards, |
45 |
Tobias |
46 |
|
47 |
|
48 |
|
49 |
-- |
50 |
Sent from aboard the Culture ship |
51 |
GCU Sacrificial Victim |