deps(dig-wallet): four semver-incompatible steps outstanding — chia-query 0.6.2→0.16 may force the chia umbrella migration (#308)

[MichaelTaylor3d]

Task

dig-wallet has four semver-incompatible dep steps outstanding, split out of PR #354 rather than
shimmed:

dep	declared	latest
chia-query	0.6.2	0.16.0
dig-node-control-interface	0.19	0.21
dig-offers	0.2	0.3
dig-clvm	0.2	0.3

All four are semver-incompatible, so nothing picks them up incidentally and no cargo update resolves
them.

The one that decides the shape

chia-query 0.16.0 almost certainly carries the chia 0.36 line, and dig-wallet is stranded on
chia 0.26 because the chia umbrella crate has no 0.36 release at all — its versions step from
the 0.26 line straight to 0.44+. That is tracked as dig-node#308, and it means every chia::* import
must be rewritten onto facade crates (measured previously at 68 call sites across 20 files, with
chia::bls::SecretKey at 11 sites the exact place a partial rewrite splits the crate internally).

So determine first whether chia-query 0.16 forces the umbrella migration. If it does, this ticket is
blocked on dig-node#308 and should say so rather than half-migrating. If it does not — if 0.16's
chia-* deps can coexist with what dig-wallet declares — the other three may be independently movable.

Do not bridge two chia lines with a conversion shim. That is the §4.1 byte-drift class and it is never
the fix. If the delta does not line up, report the shape and stop — three lanes have done exactly that
this week and each time it was the right call.

Ordering note

dig-offers 0.3 is already published on chia-wallet-sdk ^0.36, along with the whole builder family
(dig-options 0.4, dig-tips 0.3, dig-cat 0.3, dig-merkle 0.9, dig-did 0.8, dig-nft 0.3) — that
cascade completed under eco#3158. So the upstream side is ready; what is unresolved is whether dig-wallet
can receive it.

Evidence

• Verify from the RESOLVED Cargo.lock, not the caret declarations. One line per chia-* family
  among crates this ecosystem declares.
• A caret-compatible transitive is invisible to cargo fetch — a sibling patch bump elsewhere rewrote
  190 lock lines and left two dig-nat and two dig-tls lines while printing success. Name every
  crate explicitly with cargo update -p <crate>.
• cargo tree -d is NOT a usable gate here — chia-bls alone resolves to four-plus lines whose roots
  are upstream (clvmr, chialisp, rue-lir), none reaching a public DIG signature. Trace a duplicate to
  its root and report the trace.
• The real check: does any chia type reach a PUBLIC signature, and from how many lines? Do not check
  an enumerated list of crate names — two crates shipped internally split on 2026-08-22 precisely because
  the invariant was written as a list and chia-bls was not on it.
• Read passed-counts, never ok; mutations assert their patch applied; parse the test result: line.

Repo hygiene

dig-node has live lanes. Work in your own worktree off origin/main, one cargo process at a time,
never the shared checkout — it is several minors behind and carries unrelated uncommitted modifications.

Release hazard: dig-node's cron cuts a stable tag at midnight UTC unattended, so main reaches
real users without a human. The merge gate is the last line of defence.

Provenance

Split out of DIG-Network/dig-node PR #354 (eco#2790, NC-12 registry), where §2.4b was measured and found
unsatisfiable in dig-wallet without this cascade. Measured and reported rather than shimmed.

Blocked-on candidate: #308