Test for r10k failure condition #202
Conversation
|
The r10k test isn't failing. Hmm. |
|
Actually, we already have a pretty in-depth service test, and that never popped before. I wonder if it's not installing the version we think it is. |
nmburgan
force-pushed
the
r10k_service_tests
branch
2 times, most recently
from
2063170 to
da6164f
Compare
|
It might have something to do with r10k being in Bolt, or something getting fast_gettext 2.4 in there before 4.1 goes down and it just falls back to 2.4. I verified the agent version is fine. |
But those are different rubies I think? Bolt isn't installed into the openvox Agent Ruby. |
|
So the deal here is that when we ONLY install 8.23.0, then fast_gettext 4.1.0 is the only thing installed, and it is installed at When we install r10k with pre-8.23.0 installed, then when r10k is installed it sees it 2.4.0 is already present at |
nmburgan
force-pushed
the
r10k_service_tests
branch
from
da6164f to
7326d6d
Compare
nmburgan
changed the title
nmburgan
force-pushed
the
r10k_service_tests
branch
2 times, most recently
from
51472c0 to
b7975f4
Compare
|
Should work now. Passed with 8.22.1 |
nmburgan
force-pushed
the
r10k_service_tests
branch
from
b7975f4 to
e7eb074
Compare
Version 8.23.0 of the agent updated fast_gettext to 4.1.0, which broke r10k since it depends on gettext-setup, which pins fast_gettext to ~> 2.1. This simulates this upgrade condition. However, we should more robustly test upgrades at some point in the future.
nmburgan
force-pushed
the
r10k_service_tests
branch
from
e7eb074 to
9649de5
Compare
Version 8.23.0 of the agent updated fast_gettext to 4.1.0, which broke
r10k since it depends on gettext-setup, which pins fast_gettext to ~>
2.1. This simulates this upgrade condition. However, we should more
robustly test upgrades at some point in the future.
This test will fail with 8.23.0 and should pass on 8.23.1 when it’s ready