Refactor FBC caching

[joelanford]

This commit improves maintainability and extensibility of the FBC caching used to serve the GRPC API.

It:

• introduces a new cache subpackage and interface, which defines primitives for checking cache integrity, building the cache, and loading the cache.
• moves the existing registry.Querier to an implementation of the cache interface, called cache.JSON

This refactor also resolves two outstanding issues:

1. The current code contains a bug that causes panics when certain failures occur loading the cache.
2. The current code is hardcoded to always rebuild the cache if it is unreadable or its digest doesn't contain the expected value.

This commit:

1. Refactors and removes the cases the led to the panic situation
2. Adds a new flag, --cache-enforce-integrity, that causes opm serve to exit with a failure if the cache is unreadable or its digest doesn't contain the expected value. This is useful in production contexts when one always expects the cache to be pre-populated correctly.

Closes #1049
Closes #1050

Signed-off-by: Joe Lanford joe.lanford@gmail.com

Reviewer Checklist

• Implementation matches the proposed design, or proposal is updated to match implementation
• Sufficient unit test coverage
• Sufficient end-to-end test coverage
• Docs updated or added to /docs
• Commit messages sensible and descriptive

[joelanford]

Updated coverage report is here: https://app.codecov.io/github/operator-framework/operator-registry/commit/712627b1d32deeeb29773629724dbb785321d68d

Trying to track down what I did to make coverage go down... 🤔

[codecov]

Codecov Report

Merging #1051 (ca8b014) into master (0080ea0) will decrease coverage by 0.08%.
The diff coverage is 59.45%.

❗ Current head ca8b014 differs from pull request most recent head 174fad2. Consider uploading reports for the commit 174fad2 to get more accurate results

@@            Coverage Diff             @@
##           master    #1051      +/-   ##
==========================================
- Coverage   52.06%   51.98%   -0.08%     
==========================================
  Files         102      104       +2     
  Lines        9218     9216       -2     
==========================================
- Hits         4799     4791       -8     
- Misses       3505     3514       +9     
+ Partials      914      911       -3     

Impacted Files	Coverage Δ
pkg/cache/tar.go	61.11% <ø> (ø)
pkg/cache/cache.go	32.35% <32.35%> (ø)
pkg/cache/json.go	51.40% <51.40%> (ø)
alpha/declcfg/load.go	76.52% <70.37%> (-2.13%)	⬇️
pkg/cache/pkgs.go	70.40% <70.40%> (ø)
pkg/sqlite/query.go	64.61% <100.00%> (+0.08%)	⬆️

Help us with your feedback. Take ten seconds to tell us how you rate us. Have a feature suggestion? Share it here.

[everettraven]

These changes look good and seem reasonable to me. I like the approach of making the cache an interface which will make it nice if we ever feel we need to extend to support different types of caches.

[openshift-ci]

[APPROVALNOTIFIER] This PR is APPROVED

This pull-request has been approved by: everettraven, joelanford

The full list of commands accepted by this bot can be found here.

The pull request process is described here

Needs approval from an approver in each of these files:

• OWNERS [joelanford]

Approvers can indicate their approval by writing /approve in a comment
Approvers can cancel approval by writing /approve cancel in a comment

[grokspawn]

/lgtm