fix: apt cache performance

[stevenh]

Use a single call to apt-cache to reduce the time needed to lookup package versions.

Also:

• Added millisecond details to log timing so slow operations can be more easily identified.
• Perform apt update before determining package versions.

Fixes #103

[stevenh]

In a simple test case with 7 packages this reduced the processing time from 17s to 4s for the cached case.

This makes it quicker to use this action than to install the packages natively, which previously wasn't the case.

[awalsh128]

Thank you so much for this PR. Added some minor feedback.

Can you redo the PR into the dev branch? That way I can run it against the CI.

[stevenh]

Thanks for the review @awalsh128 I believe I've addressed the feedback.

If you're happy with how it now looks I can do PR against develop, just wanted to update this one first to make sure I catch all the feedback.

[stevenh]

Quick nudge to see if we can get this one in @awalsh128

[bonlime]

@awalsh128 what stops this from being merged?

[richarddd]

Any update?

[awalsh128]

Sorry for the really long wait. It has now been merged into master. I'll do a release today or tomorrow.

[awalsh128]

This is now included in release v1.3.1.

Although it looks like it is causing 2 regressions on the CI.

• https://github.com/awalsh128/cache-apt-pkgs-action-ci/actions/runs/6697244906/job/18196709301
• https://github.com/awalsh128/cache-apt-pkgs-action-ci/actions/runs/6697244906/job/18196712208

The most critical is the multiline block one. If I can't resolve in a couple of days I will need to pull the change for now. Could you help diagnose the issue introduced with the latest change (verified it is in this PR)?

[stevenh]

This is now included in release v1.3.1.

Although it looks like it is causing 2 regressions on the CI.

• https://github.com/awalsh128/cache-apt-pkgs-action-ci/actions/runs/6697244906/job/18196709301
• https://github.com/awalsh128/cache-apt-pkgs-action-ci/actions/runs/6697244906/job/18196712208

The most critical is the multiline block one. If I can't resolve in a couple of days I will need to pull the change for now. Could you help diagnose the issue introduced with the latest change (verified it is in this PR)?

Could you describe how I can validate this, so I can investigate @awalsh128 ?

[awalsh128]

The easiest to tackle is https://github.com/awalsh128/cache-apt-pkgs-action-ci/actions/runs/6697244906/job/18196712208 which tests that two package lists of different order should be the same in terms of a cache hit.

I think the issue here is that the returned normalized package list no longer sorts. Below is the difference now between the pre-hashed value.

xdot=1.2-2 rolldice=1.16-1build1 @ 6697244906-1-standard_workflow 1
rolldice=1.16-1build1 xdot=1.2-2 @ 6697244906-1-standard_workflow 1

BTW, received a report today where it is failing on the function. #116
Isn't open source fun? 😆