Add tests for NotificationListHistory

[zappolowski]

... and while at it also fix all memory leaks from this method.

[codecov-commenter]

Codecov Report

All modified and coverable lines are covered by tests ✅

Comparison is base (8196755) 64.48% compared to head (9a54a70) 65.22%.

❗ Your organization needs to install the Codecov GitHub app to enable full functionality.

Additional details and impacted files

@@            Coverage Diff             @@
##           master    #1274      +/-   ##
==========================================
+ Coverage   64.48%   65.22%   +0.74%     
==========================================
  Files          48       48              
  Lines        7918     7952      +34     
==========================================
+ Hits         5106     5187      +81     
+ Misses       2812     2765      -47     

Flag	Coverage Δ
unittests	65.22% <100.00%> (+0.74%)	⬆️

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.

[fwsmit]

Could you explain the memory leak?

[zappolowski]

From what I've seen it comes from g_variant_new_from_bytes (which is not necessary here and can be avoided by using typed versions like I did) and g_variant_type_new, which even states, that it transfers ownership and thus should be freed.

[zappolowski]

I've added some minor clean up:

• avoid heap allocation of the outer builder - just leave it on the stack as management is easier this way
• inline answer just for conciseness - g_dbus_method_invocation_return_value takes care of the floating variant created by g_variant_new already

PS: I cannot reproduce the pipeline failure. That image ran successfully on my machine. Pushed again without changes ... now it works.

@fwsmit If you agree with this, I'd like to include the fixup into the first commit before merging.

[bynect]

looks good. Are you done with the changes or something is still pending?

Also friso is away at the moment