fix(async-flow): endowment equate bug

[erights]

#endo-branch: master

Staged on #9719

closes: #XXXX
refs: #XXXX

Description

The failure at https://github.com/Agoric/agoric-sdk/pull/9719/files#diff-cbd091458bbd51ddc6640a3783cdb340d7f2588cb7ff27b330ba228cfbd0e18dR64-R87 revealed a bug introduced to async-flow when adding support for endowments. Because of the so-called "unwrapping" of some guests, there can be two guests corresponding to one host, with the host of course only mapping back to one of them -- the outer one. This makes bijection.js more complicated and irregular than an actual bijection.

equate(g, h) had a test for early return, if the g and h were already "equated", i.e., were corresponding guest and host. But the equate test was written before the elaboration of bijection. In fact, it should only test whether this guest g maps to the host h, irrespective of whether h maps back to this g.

Unfortunately:

This failure was not tested by any async-flow test before it failed for a user of async-flow.

• This PR must add such a test before it becomes a candidate for merging

This PR alters the failing test case https://github.com/Agoric/agoric-sdk/pull/9719/files#diff-cbd091458bbd51ddc6640a3783cdb340d7f2588cb7ff27b330ba228cfbd0e18dR64-R87 to see if this fix to async-flow fixes that test case. Unfortunately, it does not. However, it now fails with different symptoms. Someone who understands #9719 better than I should investigate the nature of these new symptoms. If it still looks like it might be an async-flow problem, let me know!

Security Considerations

Scaling Considerations

Documentation Considerations

Testing Considerations

Upgrade Considerations

[cloudflare-workers-and-pages]

Deploying agoric-sdk with    Cloudflare Pages

Latest commit:	883c0bb
Status:	✅  Deploy successful!
Preview URL:	https://76050423.agoric-sdk.pages.dev
Branch Preview URL:	https://markm-fix-async-flow-endowme.agoric-sdk.pages.dev

View logs

[erights]

At https://github.com/Agoric/agoric-sdk/actions/runs/9967848861/job/27542195386?pr=9722#step:5:1588 I see the following. @Chris-Hibbert , does that seem like the same kind of mysterious problem we were looking at earlier today?

RemoteError(error:liveSlots:v9#70001)#4 ERROR_NOTE: Sent as error:liveSlots:v1#70001
Logging sent error stack (RemoteError(error:liveSlots:v9#70001)#4)
not ok 8 - orchestration › contract-upgrade › resume %ava-dur=51124ms
#   start sendAnywhere
#   making offer
#   upgrade sendAnywhere with fix
#   REJECTED from ava test("resume"): (RemoteError(error:liveSlots:v1#70001)#5)
#   RemoteError(error:liveSlots:v1#70001)#5: vat-upgrade failure
#     at kunser (packages/kmarshal/src/kmarshal.js:102:51)
#     at queueAndRun (packages/SwingSet/tools/run-utils.js:40:15)
#     at async evalProposal (packages/boot/tools/supports.ts:1:9163)
#     at async packages/boot/test/orchestration/contract-upgrade.test.ts:1:996

  ---
    name: AssertionError
    message: Rejected promise returned by test
    values:
      'Rejected promise returned by test. Reason:': |-
        Error {
          message: 'vat-upgrade failure',
        }
    at: |-
      kunser (packages/kmarshal/src/kmarshal.js:102:51)
      queueAndRun (packages/SwingSet/tools/run-utils.js:40:15)
      async evalProposal (packages/boot/tools/supports.ts:1:9163)
      async packages/boot/test/orchestration/contract-upgrade.test.ts:1:996
  ...

[erights]

Argh! I see the bad old typescript stack traces with line-number 1. Now that endojs/endo#2355 is merged and endojs/endo#2359 is in review, either we'll get a new endo-release-agoric-sdk-sync soon, or I'll complete #9711 so we have proper typescript line-numbers.

[erights]

As an experiment, I added

sharp-endo-branch: master

to the top of the PR comment and provoked CI to run again. The bad news is I'm getting different problems, which I have yet to diagnose but seem unrelated to the equate bug.

The good news is that these CI errors now show proper TypeScript line-numbers!

#       at runPackageScript (/home/runner/work/agoric-sdk/agoric-sdk/packages/boot/tools/supports.ts:124:25)
#       at buildAndExtract (/home/runner/work/agoric-sdk/agoric-sdk/packages/boot/tools/supports.ts:163:7)
#       at evalProposal (/home/runner/work/agoric-sdk/agoric-sdk/packages/boot/tools/supports.ts:492:29)
#       at <anonymous> (/home/runner/work/agoric-sdk/agoric-sdk/packages/boot/test/orchestration/contract-upgrade.test.ts:26:3)

[erights]

At least some of these failures may be the same as #7937 , explaining them as not having anything to do with the contents of this PR.

[Chris-Hibbert]

does that seem like the same kind of mysterious problem we were looking at earlier today?

Oh yeah, absolutely! I've ruled out a number of possible causes of the issue I'm tracking, but haven't reached resolution yet.

[erights]

Closing in favor of #9736