Skip to content

HSCL1 0.1.7-alpha.0

Pre-release
Pre-release

Choose a tag to compare

@timfromhcs timfromhcs released this 12 Nov 10:34
· 35 commits to main since this release

HSCL1 0.1.7-alpha.0

A pre-alpha live/installer image. Not a v1 release: physical hardware, USB boot, Secure Boot
acceptance, TPM unlocking and release signing are not verified, and the installer still has an
unidentified startup crash that the wrapper retries (see Not verified).

  • ISO: HSCL1-0.1.7-alpha.0-amd64.iso
  • SHA-256: 400fb9d7aca53ef4f2b4926fb26732e0cd987b6363fc0f9800e2591d8e951df0
  • Size: 5 027 201 024 bytes (split into 3 parts, each below 1900 MiB)
  • Built from: commit 88f0fbec1ed1c0e254475c2adeab72568731f924 on main, clean tree, xz
  • Base: Kubuntu 26.04.1 desktop amd64 (Ubuntu resolute), KDE Plasma 6.6 LTS
  • Languages: English, German, French

Verified on this artifact

  • UEFI boot at 4 GB (QEMU 9.2.0, WHPX): desktop reached in 192 s, scripts/qa_report.py exits 0
    (9 checks ok, one info line: this firmware has Secure Boot off). Evidence: artifacts/boot-0.1.7-uefi.
  • Image gates: distro/live/verify-image.sh passed every check on the built tree, including the two
    new locale checks and the ESP guard's wiring.
  • Workshop stays inert: workshop_default: disabled in the ISO manifest.

What changed since 0.1.6-alpha.0

  • The installer's SIGKILL has a named sender. An audit rule on kill(2) during 12 sequential starts
    recorded sender drkonqi (pid 5919, ppid 5841) kill(pid=5841, SIGKILL) success=yes -> victim calamares:
    Calamares crashes, KCrash forks drkonqi from inside it, and drkonqi kills its own parent before
    collecting a backtrace. That explains the exit status 137 and the missing core dump. The crash itself is
    still unidentified (queue row Q-68).
  • A mitigation claim was corrected. KCRASH_DUMP_ONLY=1 does not stop the fault: 6 starts with it
    produced 2 exit-137 deaths against 1 in 6 without it, and kernel.core_pattern plus ulimit -c unlimited
    produced zero cores either way. The restart loop is the mitigation, not the variable.
  • The live locale fix never worked and now does. LANG="en_US.UTF-8" was only appended when no LANG=
    line existed, but the base image already sets LANG=C.UTF-8 - so nothing changed. The hook now replaces
    the value and generates the locale, and the image gate proves the locale exists, not just the variable.
  • The ESP guard explains itself and no longer breaks a legacy install. On failure it prints the mount
    state, /boot/grub contents and the installer log tail; a target that carries BIOS GRUB is passed
    instead of being failed for having no ESP by design.
  • docs/CLAIMS.md is machine-checked. Four rows had no artifact or a broken table (one had four cells);
    tests/unit/test_claims_table.py now fails the build on a lost cell or a claim without evidence.
  • The installer logo is not stretched (measured: rendered ink ratio 5.381 against artwork 5.333,
    +0.9 %). The audit entry that said otherwise is corrected.

Not verified in this release

  • physical hardware, USB boot and TPM unlocking are not tested
  • the ISO is unsigned: no release signature exists yet
  • Secure Boot acceptance is not demonstrated - the test firmware runs with Secure Boot off, although
    the image carries the signed Canonical/Microsoft boot components
  • the installer still crashes at startup in roughly one start in six and is restarted by the wrapper (Q-68)
  • the empty-ESP cause is unknown; the install is guarded and now fails visibly instead (Q-66)
  • no encrypted install was re-run on this exact artifact; the LUKS2 evidence is from the 0.1.4/0.1.5 line
  • the SBOM is CLI-scoped (hscl-cli, 99 packages); there is no ISO package-list SBOM

Verify your download

$ gh release download v0.1.7-alpha.0 --repo timfromhcs/HSCL1
$ git clone https://github.com/timfromhcs/HSCL1 ../HSCL1-src
$ bash ../HSCL1-src/scripts/release-join.sh . --verify-only
$ bash ../HSCL1-src/scripts/release-join.sh .

SHA256SUMS is LF-only, so sha256sum -c SHA256SUMS works on Linux; the join script checks the three parts
and the rejoined image.

See docs/KNOWN_LIMITATIONS.md, docs/CLAIMS.md and docs/BUILD_LOG.md for the full, dated evidence.