compute service hash with a default DeployConfig

[ndeloof]

What I did
as we compute service hash to compare actual vs expected state, always consider a default DeployConfig as passing --scale service=N will create one, even if the compose.yaml config doesn't define such a section

how to test:

services:
  test:
    image: alpine
    command: ping localhost

 $ docker compose up -d
[+] Building 0.0s (0/0)                                                                                                                                
[+] Running 1/1
 ✔ Container truc-test-1  Started                                                                                                                 0.2s 
 $ docker compose up --scale test=2 -d
[+] Building 0.0s (0/0)                                                                                                                                
[+] Running 2/2
 ✔ Container truc-test-1  Recreated                                                                                                             10.2s 
 ✔ Container truc-test-2  Started 

container test-1 MUST not be re-created

with this PR:

 $ docker compose up -d
[+] Building 0.0s (0/0)                                                                                                                                
[+] Running 1/1
 ✔ Container truc-test-1  Started                                                                                                                 0.2s 
 $ docker compose up --scale test=2 -d
[+] Building 0.0s (0/0)                                                                                                                                
[+] Running 2/2
 ✔ Container truc-test-1  Running                                                                                                                0.0s 
 ✔ Container truc-test-2  Started 

[ndeloof]

While this would be a significant effort, I wonder we should stop using service hash, but use container inspect to compare the actual vs expected state

[codecov]

Codecov Report

Patch coverage: 100.00% and project coverage change: -0.02 ⚠️

Comparison is base (42cd961) 59.57% compared to head (d3f145a) 59.55%.

Additional details and impacted files

@@            Coverage Diff             @@
##               v2   #10621      +/-   ##
==========================================
- Coverage   59.57%   59.55%   -0.02%     
==========================================
  Files         107      107              
  Lines        9437     9438       +1     
==========================================
- Hits         5622     5621       -1     
- Misses       3239     3241       +2     
  Partials      576      576              

Impacted Files	Coverage Δ
pkg/compose/hash.go	80.00% <100.00%> (+1.42%)	⬆️

... and 1 file with indirect coverage changes

☔ View full report in Codecov by Sentry.
📢 Do you have feedback about the report comment? Let us know in this issue.

[glours]

container

This is something we should consider, I agree

[glours]

LGTM

[glours]

Much better with a e2e test 😉🙏