You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Right now when the user adds an annex to their Elements transaction environment, it's silently dropped, causing sighash computations and other jets to provide incorrect results.
There is a FIXME in src/jet/elementsc/c_env.rs in new_raw_input about this, as of current master 06e656a. This issue is to track the FIXME.
See also BlockstreamResearch/simplicity#311 which is a libsimplicity issue which causes some minor complication (basically, we've gotta build a vector of annex pointers, which is fine, just a little inefficient).
+1 on this. I'm working on reimplementing chia-gaming state-channels in Simplicity, and my slash-path programs require annex padding to broadcast.
Programs within budget broadcast fine on regtest, but programs that need annex padding fail consensus with "Assertion failed inside jet" which can be traced to the sighash mismatch between simplicity-lang's new_raw_input passing annex: null while C-side build_txEnv includes the annex via annexHash.
Not asking for a timeline on this - just adding signal that this blocks the class of high-cost programs that the annex mechanism was designed for. Happy to test against a branch in the future if useful!
Activity
EvanWinget commentedon May 1, 2026
+1 on this. I'm working on reimplementing chia-gaming state-channels in Simplicity, and my slash-path programs require annex padding to broadcast.
Programs within budget broadcast fine on regtest, but programs that need annex padding fail consensus with "Assertion failed inside jet" which can be traced to the sighash mismatch between simplicity-lang's new_raw_input passing annex: null while C-side build_txEnv includes the annex via annexHash.
Not asking for a timeline on this - just adding signal that this blocks the class of high-cost programs that the annex mechanism was designed for. Happy to test against a branch in the future if useful!
delta1 commentedon May 1, 2026
#345
EvanWinget commentedon May 1, 2026
Thank you for the PR link - I missed this earlier!