Cardano Updates

The SDK sent MediationKeysUpdateList to the mediator but did not process
the response, so per-recipient failures reported by the mediator went
unnoticed.

Register a handler for the keylist-update-response protocol message.
Each entry in body.updated is inspected; results other than success or
no_change are logged as warnings with the recipient DID and action.
Confirmed against the Identus mediator (MediatorCoordinationExecuter.scala)
which returns this response with tuples of (recipient_did, action, result).

Closes #391

Signed-off-by: Seydi Charyyev <[email protected]>

JWT.verify did not validate the nbf (not before) claim, so JWTs with nbf in the future were incorrectly considered valid. This is a security issue per RFC 7519 Section 4.1.5.

Added an explicit nbf check after JWT decode: if nbf is present and the current time is before it, verify() returns false. JWTs without an nbf claim keep the previous behavior (no nbf enforced).

This is a sister fix to #489/#550 (which addressed the exp claim).

Closes #551

Signed-off-by: Seydi Charyyev <[email protected]>

This is consistent with the naiive N^2 diffusion of votes early sketch
that is implemented.

This is now possible because of de-duplication of the vote notification mechanism

JWT.verify did not validate the exp (expiration) claim, so expired JWTs were incorrectly considered valid. This is a security issue per RFC 7519 Section 4.1.4.

Added an explicit exp check after JWT decode: if exp is present and the current time is at or past it, verify() returns false. JWTs without an exp claim keep the previous behavior (no expiration enforced).

Closes #489

Signed-off-by: Seydi Charyyev <[email protected]>

Remove Buffer setup code for testing environment.

Signed-off-by: Abhayraj Jaiswal <[email protected]>

This results in votes being forwarded already. However, this is
obviously not safe to do this way in adversarial settings.

This is obviously a dirty hack, but works to drive the development right
now and side-steps key registration, while already exploring how to pass
the voting key into the right consensus components.

Pure additions for consumers that need to look up packages by
plan-id rather than canonical name.

* `package.identifier.unit-id` (modules/package.nix): a new
  optional field carrying plan.json's per-entry `id` (defaults to
  null when not derived from a plan entry).
* `modules/install-plan/planned.nix`: when constructing the
  per-plan-id package entries, populate `package.identifier`
  with `name` / `version` / `unit-id`.  Without this,
  reading `package.identifier.{name,version}` on a plan-id-keyed
  entry threw — the canonical-name-keyed entry already gets a
  real identifier via `load-cabal-plan.nix`, but the per-plan-id
  ones (used when one canonical name has multiple plan instances)
  did not.
* `package-ids-by-name` (modules/plan.nix +
  install-plan/override-package-by-name.nix): a name → [id]
  index over `plan-json.install-plan`.  A canonical pkg-name can
  appear under several plan ids (per-instance UnitIDs that differ
  only in their dep UnitIDs, or a multi-component package split
  across per-component entries); consumers can now iterate the
  ids unambiguously instead of name-matching against
  `config.packages`.

Replaces console.log with silent error handling to prevent leaking error details
to stdout in production. Matches the error handling pattern used in JWT.verify().

Signed-off-by: A-Chronicle <[email protected]>

* dummy GHC: mirror real GHC's `--info` capabilities and ghc-pkg `id:` field

Bring plan-to-nix's dummy GHC closer to the real cross GHC it
stands in for, so cabal-install computes the same UnitId hashes
either way.

`--info` capability fields cabal-install reads to decide
configure-args (`--enable-shared` / `--disable-shared`,
`--enable-library-for-ghci` / `--disable-library-for-ghci`,
etc.) — those land in `pkgHashConfigInputs` and feed the UnitId
hash.  The current dummy stops at platform fields, so cabal
silently picks the conservative `--disable-*` defaults.  Mirror
the real cross GHC's reported capabilities, conditioned on
target platform:

  * x86_64-w64-mingw32 (Windows mingw): no `Support shared
    libraries` field, `Support dynamic-too: NO`, `GHC Dynamic:
    NO`, RTS ways without any `_dyn` entries, Stage 1.
  * ghcjs / wasm: stage-1 with only `v debug` RTS ways and no
    dynamic linking.
  * native Linux / Darwin: dynamic everything, Stage 2.

ghc-pkg dump entries:

  * Add the missing `id:` line.  Real `ghc-pkg dump` always
    emits one, and consumers (including cabal-install's solver)
    rely on it.
  * Mirror the `-inplace` convention real GHC uses for in-tree
    libraries (base, ghc-prim, text, ...) by appending `-inplace`
    to both the package's own id and to dep ids referenced from
    other packages.  Skip the suffix for `rts` and
    `system-cxx-std-lib`, which are registered with bare
    `<name>-<version>` ids in the real package conf and would
    otherwise fork dep-id strings from what cabal computes.

* nix-tools: move cabal-install patch back inside nix-tools/ tree

#2501 placed the shared patch at `builder/cabal-install-patches/`
and referenced it from `nix-tools/cabal-install-patches.nix` as
`../builder/cabal-install-patches/...`.  That breaks the static
nix-tools build (e.g. job
`aarch64-darwin.nix-tools.static.zipped.nix-tools-static-no-ifd`,
build 1046198 on ci.zw3rk.com) with

  error: access to absolute path '/nix/store/builder' is forbidden
         in pure evaluation mode (use '--impure' to override)

`cabalProject'` copies `src = ./.` (the nix-tools dir) to the
nix store and re-evaluates module paths from there.  A path that
escapes the source tree resolves to `/nix/store/<...>/../builder`
→ `/nix/store/builder`, which doesn't exist and is rejected
under pure-eval.

Move the patch into `nix-tools/cabal-install-patches/` so the
relative path (`./cabal-install-patches/...`) stays inside the
copied source tree.  Both the regular and static nix-tools
builds still import the same shared module, so they continue to
patch cabal-install identically.

* dummy GHC: convert in-script bash comments to nix comments

The previous commit added two large explanatory `#` comment
blocks *inside* the dummy GHC's shell script (one before the
`${...}` capability-fields interpolation, one before the
`suffix()` bash function in the ghc-pkg-dump script).  Bash
comments inside the `''...''` strings are part of the resulting
script — every word change re-hashes the derivation and busts
caches downstream.

Move both blocks into nix-level `${ # ... "" }` interpolations
that evaluate to the empty string.  The comments stay adjacent
to the code they describe but no longer land in the script,
so re-wording them only churns the .nix file.

The previous commit added two large explanatory `#` comment
blocks *inside* the dummy GHC's shell script (one before the
`${...}` capability-fields interpolation, one before the
`suffix()` bash function in the ghc-pkg-dump script).  Bash
comments inside the `''...''` strings are part of the resulting
script — every word change re-hashes the derivation and busts
caches downstream.

Move both blocks into nix-level `${ # ... "" }` interpolations
that evaluate to the empty string.  The comments stay adjacent
to the code they describe but no longer land in the script,
so re-wording them only churns the .nix file.