Mapping::insert() fails in E2E tests

[SkymanOne]

Summary

When trying to use a message contains Mapping::insert() in E2E test, the test errors at dry-run stage for this specific function.

Reproducer

1. Checkout gn/mapping-take-tests
2. Navigate to examples/mappings folder
3. Run substrate-contracts-node V0.21.0 - substrate-contracts-node --dev -linfo,runtime::contracts=debug
4. Run RUST_LOG=debug cargo test
5. You should get the following error message

***
2022-11-09T16:02:34Z INFO  ink_e2e] [test: insert_and_get_works] call: [E8, 8D, 16, 46, 01, 00, 00, 00, 00, 00, 00, 00, 00, 00, 00, 00, 00, 00, 00, 00]
[2022-11-09T16:02:34Z INFO  ink_e2e] [test: insert_and_get_works] call dry run: Err(Module(ModuleError { index: 7, error: 25, message: None }))
[2022-11-09T16:02:34Z INFO  ink_e2e] [test: insert_and_get_works] call dry run debug message: 
test mappings::e2e_tests::insert_and_get_works ... FAILED

failures:

---- mappings::e2e_tests::insert_and_get_works stdout ----
thread 'mappings::e2e_tests::insert_and_get_works' panicked at 'call failed: CallDryRun', lib.rs:117:18
note: run with `RUST_BACKTRACE=1` environment variable to display a backtrace


failures:
    mappings::e2e_tests::insert_and_get_works

test result: FAILED. 0 passed; 1 failed; 0 ignored; 0 measured; 0 filtered out; finished in 0.23s

error: test failed, to rerun pass '--lib'
***

Module error that the message refers to is StorageDepositLimitExhausted

Edge cases

• Commenting out the code block containing Mapping::take() does not cause the problem, so it is definitely to do something with the function.
• Specifying storage_limit has no effect
• Executing normal & dry-run call via cargo-contract and PolkadotJS app do not result in error
• The implementations of dry-run call both in cargo-contract and E2E crate seem to be the same.

[athei]

Which node do you run against (which version)? Just asking for proper reproduction. Not that I think that it actually broke in the node.

[SkymanOne]

Which node do you run against (which version)? Just asking for proper reproduction. Not that I think that it actually broke in the node.

Updated the description to indicate the node version