o/h/ctlcmd: add a snapctl fail command

[MiguelPires]

Add a snapctl fail <reason> command to rejects changes in a registry transaction in change-view-<plug> hooks.

[zyga]

Should this be some sort of snapctl debug fail <...> command?

Separately I would make it take an argument that somehow implies where the failure is injected. As written it's somewhat generic in naming but very specific in semantics which feels wrong.

[MiguelPires]

Should this be some sort of snapctl debug fail <...> command?

Separately I would make it take an argument that somehow implies where the failure is injected. As written it's somewhat generic in naming but very specific in semantics which feels wrong.

@zyga I don't think it should be a debug subcommand but maybe a registry subcommand?
I agree with the sentiment in general. At some point, I added a --view flag to relate this to registries, since we use it for other commands, but that also wasn't right. Maybe snapctl registry fail <reason> since I think we'll add snapctl registry sub-commands eventually. @pedronis wdyt?

[codecov]

Codecov Report

Attention: Patch coverage is 90.00000% with 3 lines in your changes missing coverage. Please review.

Project coverage is 78.86%. Comparing base (ac897ee) to head (61e31f4).
Report is 28 commits behind head on master.

Files with missing lines	Patch %	Lines
overlord/hookstate/ctlcmd/fail.go	89.28%	2 Missing and 1 partial ⚠️

Additional details and impacted files

@@           Coverage Diff           @@
##           master   #14525   +/-   ##
=======================================
  Coverage   78.85%   78.86%           
=======================================
  Files        1079     1080    +1     
  Lines      145615   145666   +51     
=======================================
+ Hits       114828   114875   +47     
- Misses      23601    23605    +4     
  Partials     7186     7186           

Flag	Coverage Δ
unittests	78.86% <90.00%> (+<0.01%)	⬆️

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.

[ZeyadYasser]

The changes look really good. I agree regarding moving this to a registry sub-command snapctl registry fail <reason>. Also, It would be nice to have a small spread test with a simple test snap.

[MiguelPires]

Sorry, I thought I had commented here yesterday but looks like I forgot to click submit. Samuele's input was that this could be applicable to other hooks (so it would stay at snapctl fail) but to gate this behind the feature flag for now until we can confirm that

The spread test will be added once we have all the required pieces (I have a ticket for it already)

[ZeyadYasser]

LGTM

[andrewphelpsj]

Thanks!