Disable EOF↔Legacy cross-creation #825
Conversation
Codecov ReportAttention: Patch coverage is
Additional details and impacted files@@ Coverage Diff @@ ## master #825 +/- ## ======================================= Coverage 98.11% 98.12% ======================================= Files 119 118 -1 Lines 11752 11842 +90 ======================================= + Hits 11531 11620 +89 - Misses 221 222 +1
Flags with carried forward coverage won't be shown. Click here to find out more.
|
gumb0 force-pushed the eof/undefine-legacy-create branch from 815094d to 7bad33d Compare 
circle.yml Outdated
| - download_execution_tests: | ||
| repo: ipsilon/tests | ||
| rev: eof-relaxed-stack-validation-20240221 | ||
| rev: create3-not-supported |
There was a problem hiding this comment.
I took the branch form ethereum/tests#1337 and rolled back some changes there for supporting truncated data in containers. In this PR truncated data is not supported yet.
test/state/host.cpp Outdated
| { | ||
| // Only CREATE3/4 is allowed to deploy code starting with EF. | ||
| // It must be valid EOF, which was validated before execution. | ||
| if (msg.kind == EVMC_CREATE || msg.kind == EVMC_CREATE2) |
There was a problem hiding this comment.
It will be always true on current master, but anticipates support for CREATE3/4
gumb0 force-pushed the eof/undefine-legacy-create branch from 7bad33d to 509dfea Compare 
chfast changed the title 
test/state/host.cpp Outdated
| { | ||
| // Only CREATE3/4 is allowed to deploy code starting with EF. | ||
| // It must be valid EOF, which was validated before execution. | ||
| if (msg.kind == EVMC_CREATE || msg.kind == EVMC_CREATE2) |
| msg.input_size = init_code_size; | ||
| | ||
| // EOF initcode is not allowed for legacy creation | ||
| if (is_eof_container({msg.input_data, msg.input_size})) |
There was a problem hiding this comment.
I think this unnecessarily modifies current behavior. If the initcode starts with EF it fails during execution.
There was a problem hiding this comment.
Do you mean to change the spec? I think motivation for light failure was better UX and being aligned with legacy creation transaction being invalid with EOF.
ipsilon/eof#53
cc @pdobacz
There was a problem hiding this comment.
Yup, if we skip this, this is a change to the spec we produced (in both aspects as Andrei noted: CREATE and create tx, as they should remain aligned). I don't have a preference except that we settle the spec soon.
gumb0 force-pushed the eof/undefine-legacy-create branch from 509dfea to aff04d3 Compare 
gumb0 force-pushed the eof/undefine-legacy-create branch 2 times, most recently from 2e00a7a to 68e7291 Compare gumb0 force-pushed the eof/undefine-legacy-create branch from 68e7291 to ba4e37a Compare gumb0 force-pushed the eof/undefine-legacy-create branch from ba4e37a to f3133ba Compare 
gumb0 force-pushed the eof/undefine-legacy-create branch from f3133ba to 75365fb Compare gumb0 force-pushed the eof/undefine-legacy-create branch from 75365fb to 3bf96fe Compare gumb0 force-pushed the eof/undefine-legacy-create branch from 3bf96fe to 34f7b2b Compare 
Pulled out of #553