Skip to content
This repository has been archived by the owner on May 4, 2024. It is now read-only.

Pushing Move-on-aptos upstream #1113

Merged
merged 306 commits into from
May 1, 2024
Merged

Pushing Move-on-aptos upstream #1113

merged 306 commits into from
May 1, 2024

Conversation

wrwg
Copy link
Member

@wrwg wrwg commented May 1, 2024

This PR contains commits from aptos-core over the last 9 months and brings the aptos-main branch up-to-date to get ready for forking.

The Cargo.toml has been updated to contain workspace dependencies as now used by the Move code.

There are some fixes in the code which had to be done to make things compile, specifically an invalid reference to a prover prelude file in aptos-core from the prover code

gregnazario and others added 30 commits April 30, 2024 13:39
* [move-cli] Remove package info command

This command was using `ptree` for printing out the dependency
tree, but the dependency is out of date.  The command can be
added later, but this is to prevent from being blocked by a
future Rust update.

* [deps] Update jemalloc for future Rust version

* [move] fix formatting issue

* fix broken expected test

GitOrigin-RevId: 831d52555bec9f7fc7783dd86e82c4a5a9b71e80
GitOrigin-RevId: bc2cd13f8e81e4248e8c2c9a3617156ddb2227f6
Fix Issue 6907: A verifier failure related to inline functions

Description
provide 2 test cases related to Gerben's bug:
gerbens_test.move: ICE in hlir/translate.rs if lambda is provided to unknown or non-inline
gerbens_test1.move: byte verifier error when (&mut vector) actual is passed to (&vector) param
fix both errors:
generate a useful diagnostic error instead of the ICE
retain formal parameter types when inlining, but also
substitute all type parameters appropriately in them
Test Plan
Added test cases to illustrate bug. Any collateral damage should show up in existing test cases.

Fixes
#6907

GitOrigin-RevId: a656a1fa1c75e1542572796e9ab322b7f233b1ab
This fills in the `used_modules` field in a module, enabling the `module_env.get_used_modules` and related functionality, to unblock work on the vscode plugin which depends on this.

GitOrigin-RevId: 19de206193c1d46d475770469b8c3c66db9fe2c2
…… (#10001)

add MOVE_COMPILER_DEBUG and MOVE_TEST_DEBUG env vars to add more info to outputs for debugging

Description
VMError outputs are deliberately suppressed in unit and transactional tests to make comparison with "golden" expected outputs easier. But for debugging an unexpectedly failing test, the full VMError output is useful to see. By setting the environment variable MOVE_TEST_DEBUG=1, the developer can see the full test output. If RUST_BACKTRACE=1 is also set, then this should include the VM or Loader stack for better debugging. This is also now available by adding --verbose to the publish command in transactional tests, see PR#10000.

To debug the Move Compiler, it is most useful to see the compiler IR between passes. By setting the envronment variable MOVE_COMPILER_DEBUG=1 when running the compiler (or full build), the developer can see that IR after each pass. Note that this will be too verbose to be useful unless just a small program is being compiled.  Bytecode is not shown, but can be shown by other means (disassemble or --print-bytecode from test framework).

Test Plan
Testing that without these environment variables set, nothing changes. Manual testing for useful output with environment variables set.

GitOrigin-RevId: e7afa45778e124a466e4bda957cb586a0c43b8c1
GitOrigin-RevId: d8dd7422341991dd6aa4dd7a5c2744d6926e8bab
* [move] Add err msg for BACKWARD_INCOMPATIBLE_MODULE_UPDATE

* [move] Add err msg for global operations

* [move] Add err msg for arithmetic errors

* Fixing tests

* Fix lint

* Fix comments, update tests

GitOrigin-RevId: dd3010887c2e942c80b80d3e217b978fc3db3744
* [move-vm] Log stacktrace for natives

* Add test case

GitOrigin-RevId: 995258e833089940c596f8bb60724065c10c16c1
GitOrigin-RevId: 33a82b2642b9f6fd7d1629a5c1dee06f8d330a2e
…0037)

GitOrigin-RevId: e950a33073fe92cb0c2f55067a3e5bcdf0f55a2a
Bumps [dunce](https://gitlab.com/kornelski/dunce) from 1.0.3 to 1.0.4.
- [Commits](https://gitlab.com/kornelski/dunce/compare/v1.0.3...v1.0.4)

---
updated-dependencies:
- dependency-name: dunce
  dependency-type: direct:production
  update-type: version-update:semver-patch
...

Signed-off-by: dependabot[bot] <support@github.com>
Co-authored-by: dependabot[bot] <49699333+dependabot[bot]@users.noreply.github.com>
GitOrigin-RevId: 9d167fec55cbc01dbb1aa49bc07dffecd48f93be
GitOrigin-RevId: c88dae53c7dfee59fadd08ce04becd9106c509f4
GitOrigin-RevId: 6a4ef359f9722e3f51e240e3e1df0184744ad7dc
…e functions (#9065) (#10173)

Co-authored-by: Aalok Thakkar <aalok@Aaloks-MacBook-Pro.local>
GitOrigin-RevId: dee8ca77240e7a8164199673f41f2bcb5e43e08b
Bumps [evm](https://github.com/sorpaas/rust-evm) from 0.33.1 to 0.36.0.
- [Commits](https://github.com/sorpaas/rust-evm/commits/v0.36.0)

---
updated-dependencies:
- dependency-name: evm
  dependency-type: direct:production
...

Signed-off-by: dependabot[bot] <support@github.com>
Co-authored-by: dependabot[bot] <49699333+dependabot[bot]@users.noreply.github.com>
GitOrigin-RevId: 69b517c7e751d4f634c5f80599f78932f23f187e
…ibute warnings on aptos_stdlib (#10175)

PR 9229 introduced warnings on unknown attributes, but suppressed them for aptos_stdlib to avoid user confusion about some legacy occurences. It also removed those attributes from the codebase, so now that it is released it becomes safe to Fix issue 9410 and remove the special-case for the aptos_stdlib.

GitOrigin-RevId: 82af7acd8239bae4f580b9f6ce65aeb7cf59c411
GitOrigin-RevId: 833e6b320c02c8f992a5eeb3d96a38c6c9f6454b
GitOrigin-RevId: 3b99bb75aa68c48814559aafbc3a351c64b45b37
…) (#10064)

Fix issue #9717 by breaking critical edges in the CFG in move-compiler (v1).

While issue 9717 mentions "inline", this is a red herring as inlining a vector loop just makes it more likely to have a reference parameter on the stack, which leads to a failure to correctly drop dead references in move-compiler/src/cfgir/liveness/mod.rs. I've left the initial test case bug_9717.move and added bug_9717_looponly.move that just has a loop illustrating the problem, along with several variants on the loop (one of them, break2, also exhibiting the problem).

Anyway, the failure to properly place the reference drop happens when the constructed CFG has critical edges (edges from a node with multiple outgoing edges to a node with multiple ingoing edges). Such a situation is well-known in compiler literature to make it difficult to place instructions precisely based on certain analyses.

Fortunately, this seems to be easily fixed by adding a small pass to add a node on each such critical edge, so that the drop can be placed properly even in the case of a break.

Unfortunately, later passes can't deal with the resulting deep expression trees and chains of direct jumps that result in some cases, so I had to fix hlir/translate.rs to reduce stack depth for a given expression, and cfgir/optimize/inline_blocks.rs to properly remove the unneeded jumps.

New tests also reveal misplaced warnings/errors about unused variables in the presence of inlining, so that was also fixed.

GitOrigin-RevId: a2374a550d9fb0cc58179528dcc35e617fec9ace
* Rename variant

* [move-vm] Store ability in runtime type

* fixup! [move-vm] Store ability in runtime type

* fixup! fixup! [move-vm] Store ability in runtime type

* fixup! fixup! fixup! [move-vm] Store ability in runtime type

* [move_vm] Split loader into multiple files

* [move_vm] Cache struct type for modules

* [move_vm] Move depth cache to type cache

* [move-vm] Use name to replace type index

* [move-vm] Remove function index

* [move-vm] Inline functions

* [move-vm] Remove cached index from definition

* [move-vm] Check cross module linking before loading

* [move-vm] Remove global struct cache

* [move-vm] Inline struct name to avoid excessive memory allocation

* [move-vm] Split function in loader

* [move-vm] Split out module cache

* [e2e-test] Add randomized test for loader

* Fix Lint

* [move-vm] Cache signature resolution

* [move-vm] Removed unneeded signature token

* [move-vm] Arc-ed type argument

* Fix Zekun's comments

* fixup! [e2e-test] Add randomized test for loader

* add comments for the test strategy

* Rename struct name

* More renaming

* Addressing more comments

GitOrigin-RevId: 13eaf34e7aa002f113710087ea39a895ab779213
GitOrigin-RevId: edcae567b88d706b848e89885005efc772259715
…(#10292)

Co-authored-by: Aalok Thakkar <aalok@Aaloks-MacBook-Pro.local>
GitOrigin-RevId: 64491838ad6e996621fa52756c58f583fc7e07f3
* [move unit tests] Run extended checker as part of unit tests

Closes #9251

This runs the extended checker as part of Aptos unit tests (either our own Rust integrated tests or from the CLI). It uses the same technique as we already used for native extensions specific to Aptos: a hook is defined where additional, move-model based validations can be run. This is hook is then connected to the extended checker when running Aptos tests.

The implementation also optimizes the construction of the move model: if that one is already needed by abi generation (which is the default), it is not constructed a 2nd time for the extended checker -- both  for the existing build step and the new test step. This should avoid one full additional compilation (source -> bytecode -> model run).

* Extended checks until now excluded test code, leading to wrong usage of entry functions and attributes marked as test-only. Because fixing this is a breaking change, this commit adds the behavior to check test code via a new CLI option `--check-test-code`. This flag should eventually become default behavior.

Also fixes some reviewer comments.

GitOrigin-RevId: 79a18db5c410501d309962979b4e33f01cc7d651
…(#10470)

* remove assert

* add a test case

GitOrigin-RevId: 1369e2edf6042c5dcbe9d20bae68e0f8748cf29c
Add compiler warnings about use of #[deprecated] functions, modules, etc., in Move code, guarded by an environment variable MOVE_COMPILER_WARN_OF_DEPRECATION_USE which must be set (e.g. =1) to see the warnings.

By default, warnings within Aptos stdlib and framework files are suppressed, but can be seen by setting the environment variable WARN_OF_DEPRECATION_USE_IN_APTOS_LIBS=1 to facilitate cleanups there.

Checks happen in expansion pass of Compiler V1, so checks should also apply to Compiler V2.

Fixes
#8720: Support deprecated attribute.

GitOrigin-RevId: 5556a3ca8691365c4dc91549dded00f272229643
* [move_vm] Fix loader dependency check

* fixup! [move_vm] Fix loader dependency check

GitOrigin-RevId: d3ddfa397154d9dcbeb4a7bd88bcad4660551afe
* [move-vm] Fix broken compatibility rule

* Add test

GitOrigin-RevId: 40cb01701e8a6ed85d3aea392bbf91b14dbb9589
GitOrigin-RevId: a956f0a034bbfc8190f6a270ba20066b1d64bf8f
* [compiler-v2] prefer specific ld

* add test

* update tests

* refactor

---------

Co-authored-by: Zekun Wang <zekun.wang@aptoslabs.com>
GitOrigin-RevId: 84608ce680e02eaab6c079eea89036c70772c98b
… (#9198)

* [aggregator] New resolver and ID for aggregator  (#8941)

This PR adds the following:
    Aggregator ID can be either a table item or a u64 (unused);
    Aggregator does not use TableResolver anymore, and uses a custom one to enable value embedding;
    Aggregator test infrastructure has been moved around to allow for new IDs and enable cross-crate testing.

* [aggregator] New aggregator resolver

* [aggregator] Moving test infra around

* Addressed comments: Result<Option<u128>> --> Result<u128>

* Addressed comments

- ephemeral id is using u64 instead of u128
- history validation is factored out
- better error messages
- into_state_key --> as_state_key to operate on reference

* Require permission check before running run-indexer-tests

* [move][aggregator] Identifiers for aggregators & snapshots (#9075)

This PR allows exchanging values with identifiers in Move values.
Currently, the change happens at (de-)serialization time.

1. Resolvers have two methods now:
      - One to get bytes (the old flow)
      - One to get values with layouts (new flow)
    Both return Vec<u8> right now because we do not have
    Move values, but in general the second option can be
    used instead.

2. More tests.

3. Got rid of not necessary trait implementations and
    fixed some variable names.

4. ValueExchange is more generic now, and we can use a
    OnTagTransformation trait instead.

* Multi-versioned data structure specifically for aggregators / aggregator snapshots (#9472)

* Add native functions for Aggregator V2 (#9301)

* Add aggregator v2 native functions

* rust lint

* Addressing PR comments

* Addressing PR comments

* Addressing PR comments

* Aggregator v2 and Snapshots Implementation (#9368)

* Aggregators v2 and snapshots implementation

* Addressing a PR comment

* Minor changes

* Address all comments.

extract delta_math.rs
update AggregatorChange and AggregatorSnapshot

* additional comments

---------

Co-authored-by: Igor <igor-aptos@users.noreply.github.com>

* Fixing aggregator tests (#10059)

* [aggregator] Minor improvements (#10078)

Refactoring of aggregators V1 and V2. No new logic added.

1. We now how a dedicated `test` folder. So far only
    test utilities are there, e.g. fake aggregator view.
2. Aggregators use `AggregatorVersionedID` which is either
     a state key (V1) or an int (V2). Previously this separation
    was a bit vague.
3. Some minor refactoring included: names of functions,
    removed unused types / APIs.

Combined with:

consolidate aggregator ids (#10115)

* agg v2 natives flag (#10069)

review comments

* AggregatorSnapshot native functions implementation (#10124)

* snapshot natives

* addressed comments

* addressed review comments

* addressed comments

---------

* Propagate MoveTypeLayout to BlockSTM (#10127)

* Implementing VersionedAggregators and AggregatorChange for aggregator V2 (#10128)

* Implementing read for aggregator V2

adding AggregatorApplyChange utilities

unit tests and PR comments

* addressing comments

* connecting MVHashMap with AggregatorChange for read and write (#10152)

addressing comments

* Value exchange for resources on read (#10132)

addressed comments

connect ID generation

* [aggregators_v2] Refactor and improve aggregator tests (#10075)

with api enabled and execution disabled, all tests now pass on top of #10069

make e2e unit tests more comprehensive

* [aggregator_v2] Fixes found through testing (#10276)

* [aggregator_v2] Fixes found through testing

with these sequential execution works (excluding exchange/materialization)

* remove log

* addressing comments

---------

* [Draft] Replacing Aggregator V2 IDs with values when committing the transaction  (#10373)

* Replace aggregator v2 ids with values

* Fixing errors

* compiling fixes

* Minor changes

* Addressing PR comments

---------

* Refactor aggregator v2 exceptions, and stop (ab)using STORAGE_ERROR (#10334)

* exceptions

* addressing review comments

---------

* Rename common classes to DelayedField preffix (#10386)

* Replacing aggregator v2 identifiers with values before transaction commitment (#10390)

* Replace delayed field ids with values before commitment

* Removing a comment

* Minor fixes

* Modularize aggregator v1 materialization

* fixes (#10359)

* followup comments from refactoring agg v2 exceptions (#10407)

* Fix to check overflow/underflow in V1 aggregators, and throw appropriate exceptions (#10426)

* Replace aggregator identifiers with values in sequential execution (#10448)

* Replace identifiers with values in sequential execution

* Remove incorporate_delta_writes

* Minor change

* Fixing an error

* Addressing PR comments

* Addressing PR comments

* lint

* Addressing PR comments

* Minor funciton name change

* Remove delayed_field_keys_in_read_set

* rebase artifacts

* Add re-execution during try_commit if validation fails (#10422)

* [trivial][agg_v2] Generate IDs starting with random large number (#10525)

to easily detect issues with exchange

* rebase artifacts

* Pass layout to initiate the exchange (#10485)

* Pass layout to initiate the exchange

- why are methods with layout named with "value" and methods without named with "bytes"? that is very confusing
- where should the gating - i.e. only do exchange if AGGREGATOR_EXECUTION flag is set - be?

* removing duplicated methods

* [small][agg_v2] Fix identifier mapping - layout caching, and MVHashMap storing (#10503)

* [small][agg_v2] Fix agg v2 error handling - Success can also have failure error codes (#10428)

* Move DeltaHistory to CapturedReads (#10461)

* [agg_v2] Fallback To Sequential on PanicError::CodeInvariantError (#10531)

* rebase artifact

* [agg_v2] do not override on set base value (#10543)

* Aggregator v2 unit tests (#10593)

* Unit test to check aggregator v2 Id value exchange

* Removing print statements

* Add another unit test for aggregatorsnapshot

* Adding unit test for string snapshot

* Updating string snapshots

* Aggregator string snapshot test

* Fix unit test

* Add MoveTaylout to resource group write set (#10537)

* rebase artifacts

* [agg_v2] Fix resource groups rebase bug and try_commit deadlock (#10606)

* [agg_v2] Fix resource groups rebase bug

* commit deadlock in versioned delayed fields

* rebase artifact

---------

* Improve agg v2 e2e tests (#10506)

Previously we were only doing transactions from single account, so if they were executed speculatively, they would fail on SEQUENCE_NUMBER_TOO_NEW. modifying to use different account for each transaction, so all can happen in parallel.

adding all modes that currently work

* [agg_v2] Enable sequential exchange testing (#10524)

* [agg_v2] Use aggregator v2 fallback when resolver is not "delayed field capable" (#10592)

* [agg_v2] Use aggregator v2 fallback when resolver is not "delayed field capable"

* Adding fallback with disabled if block has WriteSetPayload::Direct transaction

* addressing review comments

---------

* [agg_v2] categorize all TODOs, and fix a few inline (#10603)

fix v1 failure error (back to what it was before)

* [agg_v2] Fixup of agg v2 tests, so they all pass (#10624)

---------

Co-authored-by: George Mitenkov <georgemitenk0v@gmail.com>
Co-authored-by: Rati Gelashvili <gelash@users.noreply.github.com>
Co-authored-by: Satya Vusirikala <satyasatya123456@gmail.com>
Co-authored-by: Igor <igor-aptos@users.noreply.github.com>
GitOrigin-RevId: 08f0aaa2523d8d9da5e539e7a642f99b7d726088
brmataptos and others added 21 commits April 30, 2024 14:05
Add an implementation of construct_test_plan in move-compiler-v2 so that it need not depend on the V1 compiler to build test plans for it.
Fixes #12789
Fixes #11589
Fixes #11706

GitOrigin-RevId: d6483636c325f76c5207bb20f6160f0f395cba39
* check metadata

* add tests

* add metadata test

* handle review comments

* add feature flag

* update release.yaml

* minor

GitOrigin-RevId: afcf9fd10000791c05d576d87542490eba9eb5dd
* add tests

* fix grammar

GitOrigin-RevId: 808327685787c260c55d3b4a5e5621a6f7e53c85
…ack for randomness txns (#12695)

* initial

* clean up

* generalize to get_required_deposit

* gated by randomness feature flags

* merge required deposit into txn metadata

* new vm error code for insufficient balance; update prologue_common inline; bypass prover tests

* polish required gas calculation

* add ut

* update

* update move

* more uts

* rename move functions

* clean up

* rename

* clean up

* update

* update

GitOrigin-RevId: f5855072772464728bd0c685beb52e5198f1cbae
For error messages like "f calls g at loc", the loc is previously the location of f, the caller, but it better be the location of g, the callee.

GitOrigin-RevId: a72ef8a716ecf3ab207c8377cb94c9c5aedaf5b4
* chore: remove repetitive words

* chore: remove repetitive words

* chore: remove repetitive words

GitOrigin-RevId: 6060f9b2669d583c6a439b1aee6fd51ce674c087
GitOrigin-RevId: 35c0b1c2eb13e82e5321d60a77b42b7a9b20f1d1
Test moved or were removed but outputs were left behind.

GitOrigin-RevId: 303583996a85cb2fe80555e27b61cfe9a74090aa
… with the same name (#12961)

* fix script conflict

* Update third_party/move/move-compiler-v2/src/file_format_generator/mod.rs

Co-authored-by: Vineeth Kashyap <vineeth.rv@gmail.com>

---------

Co-authored-by: Vineeth Kashyap <vineeth.rv@gmail.com>
GitOrigin-RevId: a1ec67862f8a339e1acd5ad4b92a143e84552b98
…ap (#12944)

* Fix disassembler use of move-compiler-v2 source map by fixing the way parameter types are added to the source map. Re-enable the source map in disassembler.

* distinguish output of MOVE_COMPILER_V2=true tests by using .v2_exp for move-cli tests

GitOrigin-RevId: ded8a8d816c5a7aec0c7db4fb93f75c7d6d44f72
GitOrigin-RevId: ba83d27f7af42939260fbccc0651a0894dcc4c0b
* Extended the Move native function set for `transaction_context`

- defined the struct `TransactionContext` to populate and pass it to the function `new_session`
- added new native functions to `transaction_context` such as sender, secondary_signers, fee_payer_opt, and max_gas_amount, gas_unit_price
- added `types/src/error.rs` for the Move error code convention
- added e2e test cases

* Add support for retrieving entry function payload from the transaction context

- Add `entry_function_payload` and proper getter methods for the payload struct
- Add e2e tests

* Add the support for multisig payload

- Add `multisig_payload` and proper getter methods for the payload struct
- Add e2e tests

GitOrigin-RevId: e2c8a6726a9bc4be464a755b47a113771a89e72c
* [FA] Coin to FA conversion

* migrate optional

* Add FA test helpers for APT, dependent, tests (#12028)

* migrate optional

* Add FA test helpers for APT, dependent, tests

* Rebuild DocGen

* Fix test broken by base branch force push

---------

Co-authored-by: Aaron Gao <lightmark@gmail.com>

* add test and fix bugs

* add feature gate

* fix issues

* fix ref getters and prover

* add the migration flag

* change map value type to Object<Metadata>

* [coin <-> FA] Add is_balance_at_least() to not lookup FS if CoinStore has enough for gas

* fix prover and lint

* removing paired_metadata call from withdraw, if CoinStore is sufficient

* fix issues

* add PRIMARY_APT_FUNGIBLE_STORE_AT_USER_ADDRESS flag

* add an event whenever a new pair is created

* add apt_migration feature

* fixup! add apt_migration feature

* address kevin most comments

* use entry functions to replace feature gates

* fixup! use entry functions to replace feature gates

* fixup! use entry functions to replace feature gates

* fixup! fix issues

* fixup! use entry functions to replace feature gates

---------

Co-authored-by: alnoki <43892045+alnoki@users.noreply.github.com>
Co-authored-by: Igor <igor-aptos@users.noreply.github.com>
GitOrigin-RevId: 541ff04bc5ce904fc9a9f9b81a138aa15433f2ae
GitOrigin-RevId: 482dcd07a7fc4e7d98fd91c713edaeb770e153ee
* [aptos-framework] Dispatchable Token Standard

* Add api to pfs

GitOrigin-RevId: bbf569abd260d94bc30fe96da297d2aecb193644
* fix

* try experiment

GitOrigin-RevId: 4ba801d51c8f3a1efd1067648aa536978a5d53de
This switches Rust dependencies in `third_party/move` to workspace dependencies, and is motivated by preparing for the new `move-language/move-on-aptos` repo.

- Those was done via IDE search&replace and henceforth everything in the tree should have been capture
- There is a conflict with inconsistent use of `rand` crate in aptos-core which was already there before. #13031 has been created to track the issue.
- There are a lot of new crates added to `./Cargo.toml`. On those deps where there before via third_party/move, so nothing really new here.
- One obsolete crate in `third_party/move/benchmarks` was deleted as it did not even compile any longer before this change.
- The crate `third_party/move/move-analyzer` has been deleted since its superceded by external code from MoveBit and is not used from anywhere in the Aptos code base

GitOrigin-RevId: f6e928190116f44f7005479bfbe63cf6bd19b360
… previous hidden uses of V1 compiler in V2 transactional tests to build precompiled libs. (#12943)

Added/changed environment flags to help debugging situations which are useful for isolating and testing compiler v2 support:

Prevent accidental use of V1: Adds an env var MOVE_COMPILER_BLOCK_V1 (=1 or =true) which causes a panic if V1 compiler is used beyond expansion phase.
- Show backtraces for compiler errors for debugging, without necessarily having verbose compiler output: MOVE_COMPILER_BACKTRACE (=1 or =true). Note that you also need to set RUST_BACKTRACE to actually get anything.
- Aliases for these and various other MOVE_COMPILER_XYZ flags as MVC_XYZ to please slow typists. (MVC_V2, MVC_BLOCK_V1, MVC_BACKTRACE, MVC_DEBUG).
- Loosens requirement on move-prover to allow a value of 1 (in addition to "true") to enable MOVE_COMPILER_V2.
- Fix transactional tests (for third_party and also aptos) to be able to avoid V1 compiler:

Remove uses of (v1 compiled) precompiled libs in V2 transactional tests.
If MOVE_COMPILER_V2=true, avoid comparison testing and exclusively use V2 compiler.

GitOrigin-RevId: e4ffe1ee4209653c7b8cf9fda9bbd379ba1b19d4
lint

comments

amount as on chain config

fix test failure

require gas amount instead of octa amount

lint

update

update default amount

fix test

GitOrigin-RevId: 8d6485a6a0505f0d22c764a814552d808bc42f8a
GitOrigin-RevId: 87a7f1d10d295a965634fe9abe2067535ab24aa8
* fix

* release

GitOrigin-RevId: e5d6d257eefdf9530ce6eb5129e2f6cbbbea8b88
@wrwg wrwg marked this pull request as ready for review May 1, 2024 03:35
@wrwg wrwg force-pushed the to_move branch 4 times, most recently from 124400b to b64472d Compare May 1, 2024 04:22
This PR contains commits from aptos-core over the last 9 months and brings the aptos-main branch up-to-date to get ready for forking.

The Cargo.toml has been updated to contain workspace dependencies as now used by the Move code.
@wrwg wrwg merged commit bee3bf9 into aptos-main May 1, 2024
7 checks passed
Sign up for free to subscribe to this conversation on GitHub. Already have an account? Sign in.
Labels
None yet
Projects
None yet
Development

Successfully merging this pull request may close these issues.

None yet