Mark S. Miller

Doh! I remembered the wrong name. `test.serial` is the problem. I revised the first message in this thread to correct that.

It seems to be only those two uses of `test.serial`. Several others do fine with ses-ava. Some others have not yet been tried with ses-ava.

The left pane of https://github.com/Agoric/agoric-sdk/pull/2740/files#diff-8845f7584737c882a225a4b88eee42f5064626370ad4b222d3e0319fd6e4cb68 shows that it was previously (prior to this PR at least) using ava, not ses-ava. When was it using ses-ava ?

https://github.com/Agoric/dapp-token-economy/pull/154/files is the only change to ses-ava since your "macro" change. See https://github.com/endojs/endo/commits/master/packages/ses-ava Nothing there but updating dependency on ses to 0.12.6 and one irrelevant test change.

Close as in "it's already fixed, let's close the bug" or "let's fix this"?

Looks like the bug report was provoked by test-message-pattern.js breaking when we tried switching it to ses-ava. If the underlying issue is fixed, can we now switch test-message-pattern.js to ses-ava?...

Looks like we've got over 180 imports of 'ava' rather than 'ses-ava', most of which I expect should be switched if these are no longer problems. Again, not an MN-1...

The current shim behavior --- where an extensible weakmap side table is consulted by the replacement `Function.prototype.toString` --- is unlikely to be something we'd want to propose. Rather, the hardened...

"after lockdown"? How could that possibly be safe? Do you mean, during the (hypothetical) vetted shim phase, that happens after the `repairIntrinsics` phase of lockdown, but before the `hardenIntrinsics` phase?...

What I don't understand is the suggestion to be able to do it after lockdown. This would imply that lockdown would leave `Function.prototype.toString` in a mutable state (writable or configurable...