feat(gnovm): add 'gno test -print-events' + cleanup machine between tests

[moul]

• add gno test -print-events flag for unit tests.
  Props to @r3v4s for his work on feat(gno): use verbose flag to print emitted events #2071
• add // Events: support in _filetests.gno.
• cleanup gno.Machine between unit tests (\o/) .

Fixes #1982
Closes #2071
Addresses #2007

[codecov]

Codecov Report

Attention: Patch coverage is 59.42029% with 28 lines in your changes missing coverage. Please review.

Project coverage is 63.40%. Comparing base (d961785) to head (3409b96).
Report is 1 commits behind head on master.

Files with missing lines	Patch %	Lines
gnovm/tests/file.go	46.93%	23 Missing and 3 partials ⚠️
gnovm/cmd/gno/test.go	90.00%	1 Missing and 1 partial ⚠️

Additional details and impacted files

@@            Coverage Diff             @@
##           master    #2975      +/-   ##
==========================================
+ Coverage   63.39%   63.40%   +0.01%     
==========================================
  Files         565      565              
  Lines       79390    79454      +64     
==========================================
+ Hits        50326    50377      +51     
- Misses      25674    25682       +8     
- Partials     3390     3395       +5     

Flag	Coverage Δ
contribs/gnodev	60.57% <ø> (ø)
contribs/gnofaucet	14.82% <ø> (ø)
gno.land	67.37% <ø> (ø)
gnovm	67.87% <59.42%> (+0.01%)	⬆️
misc/genstd	79.72% <ø> (ø)
tm2	62.47% <ø> (+0.03%)	⬆️

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.

[leohhhn]

There's a PR for this, although it hasn't been merged for some time.

#2071

[thehowl]

This is no longer required, no?

[thehowl]

I'm not sure about this. What if we had something like testing.Events() to get the list of events which were emitted, which then can be printed by the test?

This way, we also have a way to programmatically verify the tests, rather than having to pass -print-events.

[leohhhn]

See consideration at bottom of #2007

This would definitely be something needed, it's standard in other blockchain ecosystems. Ideally, we could actually do type checking on event objects down the line, but for now I guess testing.Events() would return a string of events or similar.

[moul]

testing.Events appears to be a valuable enhancement, particularly for unit testing. However, print-events allows for event printing in existing flows, regardless of whether they have been tested or modified. In my case, I wanted to understand what was happening in some multiphase test flows. At another time, I aimed to identify all the tests that triggered a specific std.Emit, including those from realms that import the emitter indirectly.

I believe it makes sense to keep both options. The one in this PR functions like a specialized logging engine. Additionally, events serve as an alternative stdout in some ways.

Once we have the testing.Events() helper, I'm definitely open to removing the -print-events flag for the sake of simplification. Let's discuss when it will take place.

[leohhhn]

I wonder if maybe it would be nicer to have a more human-readable format when printing. 🤷

[moul]

The current mode is to provide the same JSON output on the WebSocket, during the maketx call, and in unit tests. Essentially, events are to be intercepted by machines, so I believe it is beneficial to have the exact same representation, which is also copy-paste friendly.

We can reconsider.