Cardano Updates

Adapt to the cli version automatically

Adapt to the cli version automatically

## Problem

The CI build host (`zur1-s-d-029`) runs ~28 GHA self-hosted runners
as a single user against **one shared tmpfs `/tmp`**. Any job running
`rm -rf /tmp/e2e-* /tmp/test-cluster*` on `if: always()` will wipe
sibling runners' **live** VolatileDB files.

Under UTxO-HD (cardano-node >= 10.7.0) the consensus layer re-opens
VolatileDB files by path through `fs-api`, so an unlinked
`blocks-*.dat` now crashes the node with
`ApiMisuse (ClosedDBError (UnexpectedFailure (FileSystemError
FsResourceDoesNotExist …)))`
instead of tolerating the unlink via open fd's.

This reproduced on master (`253d290bfd`) — Conway Integration Tests
crashed at 15:46:43 UTC on 2026-04-20 with two pool nodes failing
simultaneously on
`/tmp/test-cluster436150/pool-*/db/volatile/blocks-0.dat`.
See also upstream ouroboros-consensus#1991.

## Fix

- Set `TMPDIR: ${{ runner.temp }}` on every job that launches a local
  test cluster or E2E run. `$RUNNER_TEMP` is per-job, per-runner, and
  auto-cleaned by the runner service between jobs, so clusters live
  in a private directory that sibling runners cannot touch.
- Drop the pre-existing dangerous cleanup in
  `.github/workflows/linux-e2e.yml` and `.github/workflows/release.yml`
  — `$RUNNER_TEMP` makes them redundant.
- Replace the shared fixed path `TMPDIR: /tmp/gha-bench` in
  `linux-benchmarks.yml` and `restoration-benchmarks.yml` for the
  same reason.

No new cleanup steps are introduced.

No longer allow leftover bytes when decoding IPv4/IPv6 addresses regardless
of decoder version.

* fix: correct CBOR encoding and decoding for negative big ints

Signed-off-by: Jonathan Lim <[email protected]>

* fix: use alloc_integer instead of alloc

Signed-off-by: Jonathan Lim <[email protected]>

* chore: formatting

Signed-off-by: Jonathan Lim <[email protected]>

* chore: added tests showing the negative bigint handling

Signed-off-by: Jonathan Lim <[email protected]>

* style: formatting

Signed-off-by: Jonathan Lim <[email protected]>

---------

Signed-off-by: Jonathan Lim <[email protected]>

Apply these to Redeemers.

The build host runs ~28 GHA runners as a single user against one tmpfs
/tmp, so any job running `rm -rf /tmp/e2e-* /tmp/test-cluster*` on
`if: always()` wipes sibling runners' live VolatileDB files. Under
UTxO-HD (node >= 10.7.0) the consensus layer re-opens these files by
path and crashes with `FsResourceDoesNotExist` instead of tolerating
the unlink via open fd's, which manifested as the Conway Integration
Tests failure on master at 253d290bfd.

Point TMPDIR at \$RUNNER_TEMP — per-job, per-runner, auto-cleaned by
the runner service — so test clusters live in a private directory that
sibling runners cannot touch. Drop the now-redundant (and dangerous)
`rm -rf /tmp/...` cleanup steps from linux-e2e and release. Replace
the shared `/tmp/gha-bench` fixed path in benchmarks with RUNNER_TEMP
for the same reason.

No longer allow leftover bytes when decoding IPv4/IPv6 addresses regardless
of decoder version.

Signed-off-by: Chris Guiney <[email protected]>

Switches log-tailer to ghcr.io/cardano-foundation/cardano-node-antithesis/log-tailer:a7d1d7d
(built from the previous commit). Antithesis cannot pull docker.io/alpine
via its pull-through mirror, so the service needs a pre-published image.

Moves log-tailer.sh to components/log-tailer/ and adds an alpine-based
Dockerfile so the image can be built and published via publish-images
CI. Antithesis registry mirror only resolves pre-mirrored images, so
a dedicated ghcr.io image is required.

This should fix the following error:

    error: hash mismatch in fixed-output derivation '/nix/store/nzqbf3y4kqanwi2rq5h1n27qglhlmvmk-head.hackage.ghc.haskell.org.drv':
             specified: sha256-7jhEysaHgczdNolOCuE9K6ZpLD3KJzsGSlgPHlZOvlg=
                got:    sha256-yFGpWo9ulwBEUyzTLKJBtbCWi7Ly4NEtocBRogdew9g=

No longer allow leftover bytes when decoding IPv4/IPv6 addresses regardless
of decoder version.

The build host runs ~28 GHA runners as a single user against one tmpfs
/tmp, so any job running `rm -rf /tmp/e2e-* /tmp/test-cluster*` on
`if: always()` wipes sibling runners' live VolatileDB files. Under
UTxO-HD (node >= 10.7.0) the consensus layer re-opens these files by
path and crashes with `FsResourceDoesNotExist` instead of tolerating
the unlink via open fd's, which manifested as the Conway Integration
Tests failure on master at 253d290bfd.

Point TMPDIR at \$RUNNER_TEMP — per-job, per-runner, auto-cleaned by
the runner service — so test clusters live in a private directory that
sibling runners cannot touch. Drop the now-redundant (and dangerous)
`rm -rf /tmp/...` cleanup steps from linux-e2e and release. Replace
the shared `/tmp/gha-bench` fixed path in benchmarks with RUNNER_TEMP
for the same reason.