Summary
tinyminiscript rejects valid sorted multisig descriptor fragments in both the classic and tapscript families:
sortedmulti(...)sortedmulti_a(...)
The current parser has branches for multi and multi_a, but no matching branches for the sorted variants.
Repro 1: classic sortedmulti(...)
wsh(sortedmulti(2,0279be667ef9dcbbac55a06295ce870b07029bfcdb2dce28d959f2815b16f81798,02c6047f9441ed7d6d3045406e95c07cd85c778e4b8cef3ca7abac09b95c709ee5,02f9308a019258c31049344f85f89d5229b531c845836f99b08601f113bce036f9))
Observed rejection:
UnexpectedToken { expected: "0 or 1", found: ("sortedmulti", 5) }
Repro 2: tapscript sortedmulti_a(...)
tr(02a34b99f22c790c4e36b2b3c2c35a36db06226e41c692fc82b8b56ac1c540c5bd,sortedmulti_a(2,50929b74c1a04954b78b4b6035e97a5e078a5a0f28ec96d547bfee9ace803ac0,6db06226e41c692fc82b8b56ac1c540c5bd50929b74c1a04954b78b4b6035e97,442e8300a13cfc307ffaad450d79fd434b05ee7f7a2ea03c0cb5dd7ad4a82e89))
Observed rejection:
UnexpectedToken { expected: "0 or 1", found: ("sortedmulti_a", 71) }
Control
The closely related multi_a(...) form is accepted by the same parser:
tr(02a34b99f22c790c4e36b2b3c2c35a36db06226e41c692fc82b8b56ac1c540c5bd,multi_a(2,50929b74c1a04954b78b4b6035e97a5e078a5a0f28ec96d547bfee9ace803ac0,6db06226e41c692fc82b8b56ac1c540c5bd50929b74c1a04954b78b4b6035e97,442e8300a13cfc307ffaad450d79fd434b05ee7f7a2ea03c0cb5dd7ad4a82e89))
Cross-check
The repros above are parse-accepted and compiled by maintained peers including:
rust-miniscriptbitcoin/bitcoinembit
and for the classic sortedmulti(...) form also:
bitcoinerlab/descriptorsNBitcoinelements-miniscript
Why this looks like one bug family
In src/parser/mod.rs, the parser implements explicit branches for multi and multi_a, but there is no corresponding branch for sortedmulti or sortedmulti_a.
Summary
tinyminiscriptrejects valid sorted multisig descriptor fragments in both the classic and tapscript families:sortedmulti(...)sortedmulti_a(...)The current parser has branches for
multiandmulti_a, but no matching branches for the sorted variants.Repro 1: classic sortedmulti(...)
Observed rejection:
Repro 2: tapscript sortedmulti_a(...)
Observed rejection:
Control
The closely related
multi_a(...)form is accepted by the same parser:Cross-check
The repros above are parse-accepted and compiled by maintained peers including:
rust-miniscriptbitcoin/bitcoinembitand for the classic
sortedmulti(...)form also:bitcoinerlab/descriptorsNBitcoinelements-miniscriptWhy this looks like one bug family
In
src/parser/mod.rs, the parser implements explicit branches formultiandmulti_a, but there is no corresponding branch forsortedmultiorsortedmulti_a.