Expand HQRM tests to run on class-based resources

[dan-hughes]

As per title.

IMO it should be run against the built module or source files to check all private, public scripts and classes.

[johlju]

Thought we ran it on the builtModule? Certin are run on source? 🤔

[dan-hughes]

I did too, but locally or in CI only script-based DSC resources are being checked in the HQRM task.

Or in SqlServerDsc the .Common module.

[dan-hughes]

After some investigation, this is back to the checking of source or built module. PSSA tests run against source.

Depending on the module, the Module.psm1 either does not exist or is empty in the source location.

With the direction of DSCv3 favoring class-based resources, I feel it will be beneficial to decide some 'rules' for HQRM on how/what and where checks are made on the different resources, functions etc.

There is the issue with localization for class-based resources, which has a bunch of edge cases that need to be handled too.

[dan-hughes]

Current status of the v5 tests:

Name	source	module	covers classes	covers public/private func
ExampleFiles	✓		N/A	N/A
FileFormatting	✓	✓	✓	✓
Localization	✓	✓	TODO	TODO
MarkdownLinks	✓		N/A	N/A
ModuleManifest		✓	TODO	TODO
ModuleScriptFiles	✓	✓	builtOnly	builtOnly
PSSAResource	✓		TODO	TODO
PublishExampleFiles	✓		N/A	N/A
RelativePathLength		✓	N/A	N/A
ResourceSchema	✓		N/A	N/A

I believe three need attention:

• Localization
• PSSAResource

For the above two the where and what that need to be tested is more complex. e.g. for script-based DSC resources, source is fine, but class or functions need to be checked once the module is assembled therefore the test should use builtModule.

• ModuleManifest

I have identified some gaps e.g. checking if functions and aliases are declared. There is also the PS7 issue where it does not read DscResources in the manifest with the current configuration. A check should be added to confirm CmdletsToExport = '*' for example.

Is it worth tracking all of this in this as a top level issue and track each one under some sub issues?

[johlju]

I think I have resolved the ModuleManifest tests in PR #167. It should now validate correctly against built module's manifest.

[dan-hughes]

@johlju,

Can you add #163, #165 as child issues to this please.

[johlju]

Added.