1 |
On Mon, 2 Nov 2020 22:44:07 -0500 |
2 |
Matt Turner <mattst88@g.o> wrote: |
3 |
|
4 |
> The catalyst-auto automation scripts live in a repo separate from |
5 |
> catalyst. That increases the difficulty of changing catalyst's |
6 |
> interface, and it doesn't seem to offer any advantages otherwise. |
7 |
> (Keeping build specs in a separate repo allows them to be updated |
8 |
> independent of catalyst and that is valuable). Additionally, since the |
9 |
> primary way catalyst is used is via this automation, it makes sense to |
10 |
> support this workflow in catalyst directly. |
11 |
> |
12 |
> But to get there, there are some changes to catalyst that I think are |
13 |
> improvements on their own and simplify the path to integrating |
14 |
> automation capabilities directly into catalyst. That's what I'd like |
15 |
> to discuss here. |
16 |
> |
17 |
|
18 |
|
19 |
I have been thinking for the past few years that the automation could |
20 |
benefit from using a buildbot to run the stages. In that way it |
21 |
would set the required variables, run the stages in sequence. Upon |
22 |
failure, buildbot makes it easy to re-run failed steps from where they |
23 |
left off. Or initiate unscheduled builds. It also makes it easy to see |
24 |
detailed logs (by anyone with a browser if the buildbot is public |
25 |
viewable) of the various steps for debugging, etc.. |
26 |
|
27 |
|
28 |
Perhaps with your new spec file, you add a varialbe that lists the |
29 |
stages to run in sequence. In that way it would preserve the old |
30 |
capability of running single stages independantly or a full sequence. |
31 |
Or perhaps a cli option to override the setting on an unedited spec |
32 |
file. |
33 |
|
34 |
ie: |
35 |
|
36 |
[build.stages] |
37 |
stage1 |
38 |
stage2 |
39 |
stage3 |
40 |
livecd1 |
41 |
livecd2 |
42 |
|
43 |
|
44 |
sorry, not familiar with toml |
45 |
|
46 |
In that way a spec file could be edited simply to restart from any |
47 |
point with a single variable edit without removing unused [build.???] |
48 |
definitions for the full run. This would be particularly useful when |
49 |
troubleshooting hidden/delayed stage build fails. |
50 |
|
51 |
|
52 |
Overall, I do agree that the releng automation scripts capabilities |
53 |
should be part of the caltlyst repo in order for them to be up-to-date |
54 |
with catalyst code. |
55 |
|
56 |
I have limited time and resources lately, so can't help out much until |
57 |
I get back home (probably Xmas) largely due to covid... I only have |
58 |
this small laptop and my eyes are not that good to be doing lots of |
59 |
work on a tiny 14 inch screen and a stage run to take hours instead |
60 |
of minutes... (yes, I got spoiled with 2 28 inch 4K displays, 16-core |
61 |
128k ram system at home...) |
62 |
|
63 |
Brian Dolbec |
64 |
<dolsen@g.o> |