Provide CHECKBOX_RUNTIME environment variable during job execution only (BugFix)

[pieqq]

Description

Providing $CHECKBOX_RUNTIME when executing the snap (in a wrapper script) is hackish and often does not work because of the problems described in #2295 .

Instead, it is now provided by Checkbox during the job execution phase (similarly to $PLAINBOX_SESSION_SHARE and the like).

This means that it cannot be used directly from the Shell (when entering snap run --shell checkbox.checkbox-cli or checkbox.shell), but it should work as expected when running a job that relies on $CHECKBOX_RUNTIME.

Resolved issues

Fix #2295
Fix CHECKBOX-2164

Tests

Apart from adding some unit tests for the get_execution_environment function, I built snaps from this branch and used the checkbox24_amd64 artifact to run a spread test using @Hook25 's new shiny tool, checkbox_snap_spread:

• I modified tests/run-canary/launcher.conf to not exclude anything and run the com.canonical.certification::misc-client-cert-automated test plan (since this one has some fwts-related tests)
• I put checkbox24_7.1.0-dev27_amd64.snap at the root of the directory, and ran image-garden.spread -artifacts=artifacts ubuntu-core-24

When running jobs that run an fwts command, I can see that the CHECKBOX_RUNTIME has been set to the right thing. For example, with the miscellanea/oops job:

----------------------------[ Run FWTS OOPS check ]-----------------------------
ID: com.canonical.certification::miscellanea/oops
(...)
Command: "fwts -j /snap/checkbox24/current/share/fwts -q --stdout-summary -r 
/var/tmp/checkbox-ng/sessions/session_title-2026-03-04T12.46.53.session/session-share/fwts_oops_results.log
oops".
Running tests: oops.

oops: Scan kernel log for Oopses.
--------------------------------------------------------------------------------
Test 1 of 1: Kernel log oops check.
PASSED: Test 1, Found no oopses in kernel log.
PASSED: Test 1, Found no WARN_ON warnings in kernel log.
(...)

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 57.66%. Comparing base (aef672b) to head (b592ed7).
⚠️ Report is 3 commits behind head on main.

Additional details and impacted files

@@            Coverage Diff             @@
##             main    #2362      +/-   ##
==========================================
+ Coverage   57.63%   57.66%   +0.03%     
==========================================
  Files         465      465              
  Lines       47165    47166       +1     
  Branches     8397     8398       +1     
==========================================
+ Hits        27182    27198      +16     
+ Misses      19108    19088      -20     
- Partials      875      880       +5     

Flag	Coverage Δ
checkbox-ng	75.60% <100.00%> (+0.07%)	⬆️

Flags with carried forward coverage won't be shown. Click here to find out more.

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.

[pieqq]

@tomli380576 Here is the patch I was talking about yesterday. I've built some snaps (check the PR description) and ran some tests with a UC24 VM, and it looks like it's working.

I've attached the submission.json from that run here if you want to investigate, but I'd appreciate if you could run the snaps on your testing environment and report back. One thing to note is that CHECKBOX_RUNTIME is now populated by Checkbox itself when execution a job, so it's not available anymore in the Checkbox Shell by default (since that was poorly handled by the wrapper scripts and was removed).

[Hook25]

I've been wanting to purge that findmnt monstrosity for a while. Thanks.

Please fix black formatting but this lgtm

[tomli380576]

ubuntu desktop:

USB: https://certification.canonical.com/hardware/202503-36476/submission/476198/
Suspend and reboot tests: https://certification.canonical.com/hardware/202503-36476/submission/476204/ (the failed test is a real fwts failure)

I'll do another test with ubutu core on a real machine next Monday but I think the changes look good!

[tomli380576]

lgtm

[Hook25]

Thanks

[tomli380576]

miscellanea/oops also passed on ubuntu core: https://certification.canonical.com/hardware/202503-36476/submission/476473/test/58145/result/56513541/