Test both collateral configurations on CI

[gkaracha]

Prerequisite for #213.

[github-actions]

Gas costs: No change.

Entrypoint sizes	d6ea806	4263df3	Diff
touch	56749	56556	-193

Test coverage	d6ea806	4263df3	Diff
price.ml	76.92	100	23.08
TOTAL	92.47	92.57	0.09999999999999432

[gkaracha]

This seems to be breaking the message bot; reverting back to a draft until we resolve this.

[gkaracha]

Undrafting; the bot seems to be working after all. I guess writing the output twice is problematic in general though 🤔 (and perhaps a little flaky; statistics can be different for different configurations)

[dorranh]

Ah now that is an unforeseen consequence - perhaps we should start plotting histograms instead 😅

[dorranh]

In all seriousness, I think that with the current approach we are actually overwriting the coverage, gas cost, etc. artifacts. We might want to split them out per workflow (e.g. with a different suffix) and update the comment bot accordingly.

[gkaracha]

Yeah, true. My initial thought was to only allow writing gas costs if ${{ matrix.configuration_file }} == "checker-tez.yaml" but I agree with you. It would be nice to be able to catch regressions for either deployment. Let me turn this again into a draft until this is implemented correctly.

[dorranh]

My initial thought was to only allow writing gas costs if ${{ matrix.configuration_file }} == "checker-tez.yaml"

You know, this might actually be a nicer immediate solution. That way we can go ahead and start testing both and then add support for metrics for FA2 checker in a separate PR.

[github-actions]

Gas costs: No change.

Entrypoint sizes	cf1f913	adc6c48	Diff
touch	56749	56556	-193

Test coverage	cf1f913	adc6c48	Diff
price.ml	76.92	100	23.08
TOTAL	92.47	92.57	0.09999999999999432

[gkaracha]

My initial thought was to only allow writing gas costs if ${{ matrix.configuration_file }} == "checker-tez.yaml"

You know, this might actually be a nicer immediate solution. That way we can go ahead and start testing both and then add support for metrics for FA2 checker in a separate PR.

More incrementally, eh? 🤔 That sounds good to me; at least we can already safely know that both collateral configurations work, even if we don't get extra gas info about the FA2 one (gas costs can still not go overboard though, because if that happens I expect the corresponding e2e test would fail, so that's another safety net). Let's see how adc6c48 does

[gkaracha]

I'll go ahead and merge this one so that both collateral configurations are always tested and we can always add support for metrics for collateral=FA2 checker in a separate PR; currently the artifacts stored only capture gas costs etc. for collateral=TEZ.