add unit test for cli

[chloeyin]

Signed-off-by: zaihaoyin zaihaoyin@microsoft.com

This PR adds unit test for some cli commands, only to make sure the flags are parsed correctly.
This is the last PR of the whole series(#255).

[codecov-commenter]

Codecov Report

Merging #274 (eebbd39) into main (1f858d0) will decrease coverage by 20.62%.
The diff coverage is n/a.

@@             Coverage Diff             @@
##             main     #274       +/-   ##
===========================================
- Coverage   50.00%   29.37%   -20.63%     
===========================================
  Files          11       27       +16     
  Lines         344     1685     +1341     
===========================================
+ Hits          172      495      +323     
- Misses        164     1176     +1012     
- Partials        8       14        +6     

Impacted Files	Coverage Δ
cmd/notation/plugin.go	0.00% <0.00%> (ø)
cmd/notation/push.go	24.61% <0.00%> (ø)
cmd/notation/manifest.go	0.00% <0.00%> (ø)
cmd/notation/logout.go	54.16% <0.00%> (ø)
cmd/notation/registry.go	0.00% <0.00%> (ø)
cmd/notation/common.go	100.00% <0.00%> (ø)
cmd/notation/verify.go	19.00% <0.00%> (ø)
cmd/notation/main.go	0.00% <0.00%> (ø)
cmd/notation/key.go	23.52% <0.00%> (ø)
cmd/notation/sign.go	29.88% <0.00%> (ø)
... and 6 more

📣 Codecov can now indicate which changes are the most critical in Pull Requests. Learn more

[JeyJeyGao]

LGTM

[binbin-li]

Left some tiny comments. Just wanna raise some questions out of scope.

1. Looks like in this PR we focus on the main use cases, successful case and a common failed case which should cover most of cases. But users can fail a command in a lot ways (a combination of flags or args). It's kind of time-wasting to test all of them, just wanna confirm we would do the same to the new PRs. @shizhMSFT
2. This PR tests the conversion from command input to the internal Opts object. For the business logics after it, is there a good way to test it? Maybe just run cmd.ParseFlags() like this PR does? Or move the logics to other places for easier testing?

[binbin-li]

lgtm overall

[shizhMSFT]

LGTM

[shizhMSFT]

Left some tiny comments. Just wanna raise some questions out of scope.

1. Looks like in this PR we focus on the main use cases, successful case and a common failed case which should cover most of cases. But users can fail a command in a lot ways (a combination of flags or args). It's kind of time-wasting to test all of them, just wanna confirm we would do the same to the new PRs. @shizhMSFT
2. This PR tests the conversion from command input to the internal Opts object. For the business logics after it, is there a good way to test it? Maybe just run cmd.ParseFlags() like this PR does? Or move the logics to other places for easier testing?

We are still figuring out a better way to test CLI.