Aggregate issue for LintDiff reliability improvements

[konrad-jamrozik]

Note: the list below is competing for priority with

• Aggregate issue for kojamroz's Dilithium semester backlog #7795

Recently identified set of LintDiff issues needing addressing:

• [LintDiff] Add information about rule at fault if originating from openapi-validator-plugin-func.ts (native ruleset) #7612
• LintDiff ARM linter false negative on PatchBodyParametersSchema rule #7701
• LintDiff is not recognizing AutoRest errors, thus incorrectly reporting success #7706
• LintDiff is being bypassed for TypeSpec conversion PRs, while it shouldn't. #7707
• LintDiff (most likely) incorrectly concludes in several cases it should not run its validation #7730
• LintDiff throws AutoRest error when a specs PR moves README / renames dir #7748
• API version changes are not processed by LintDiff (and likely other tools, too) if the corresponding AutoRest README was unchanged #7759
• If a new spec is not added to README, LintDiff won't validate it (related to Avocado). #7761
• LintDiff check should fail if it ran no validations #7776
• LintDiff is apparently hanging and timing out after 1 hour #7780
• LintDiffToolException with code: ERESOLVER #7781
• LintDiff throws AutoRest exception when default tag is not defined in README; error is too general and Avocado does not report failure #8230

Nice-to-haves:

• [LintDiff] Pin version of @microsoft.azure/openapi-validator-rulesets #7613
• [LintDiff] simplify CI/CD deployment / change management #7619
• [LintDiff][PR Workflow] Handle cases where check result is too large to be processed by our infrastructure #7457
• [LintDiff] Upgrade autorest to latest #7755

This work item is a prerequisite for:

• [PR Workflow] Automatic ARM sign-off for TypeSpec-based PRs  #7352 (comment)

Note there is also other related feature work:

• [LintDiff] Need separate ruleset for TypeSpec-generated Swaggers #7732