Teng Zhang

It seems that currently dev_setup.sh does not support Windows directly. Maybe you could use docker instead?

#830 **partially** solves the issue by allowing updating global spec variables in the function body. However, local ghost variables are not supported yet.

cc @vineethk

> Note that currently the test in aptos-transactional-test-harness seems > to be succeeding when run. So perhaps I'm doing something wrong > to enable paranoid checks here? Have you checked...

For the following code (from the test `writeref_borrow_invalid.move`): ``` module 0x8675309::M { struct G has copy, drop { v: u64 } struct S has copy, drop { g: G }...

Hi @flex-alcove, thanks for submitting the issue. Would you mind also submitting this to the revela repo: https://github.com/verichains/revela where we keep track all issues related to the decompiler? Thanks!

Hi, @tnowacki @tzakian, I am wondering whether the panic described in this issue is intentional or not? If not, could you give me some direction on how to fix it?...

There are some special comments such as `//#run ` or `//#publish`, could you add test cases for them? Here is an example: [bug_9717.move](https://github.com/aptos-labs/aptos-core/blob/353d5657d2165f65741fdeb85195cf73b9c12058/third_party/move/move-compiler-v2/transactional-tests/tests/inlining/bug_9717.move)

Confirmed that V1 behaves the same except that it generates the warning `unnecessary trailing semicolon`. cc @fEst1ck @wrwg

With PR #11313, users can set the environment variable`ENABLE_V2` to run e2e-move test using compiler V2.