Runs passed LD_PRELOAD=libfaketime.so.1 by bare name on a host without libfaketime. The loader skipped it silently, so those runs recorded a fixed clock they never had. A preload must now be an absolute path to a file, which is bound into the sandbox and recorded by sha256.
campaign.py review applies or rejects an open finding with the reason read in the code. Applied fields go through the profile editor with the finding's records as evidence. The rule table reads on_mismatch: warn and counts a signature as a content check.
Records that back the same field or issue claim share one line. Each
line says what the runs showed: the BIOS the core ran, the file strace
saw opened, or what a maintainer watched in an issue's run
(tests/campaign/observations.yml). A run with nothing to state stays
off the page.
Emulator pages show a run only when a profile's evidence names it or it
answers an issue, as one line naming the BIOS the core ran. The Tested
blocks, region outcomes, coverage, campaign and pack tables and the
benchmark figures are gone from the site; the records stay.
The region outcome per core and per game says the BIOS each boot ran in fewer words, so the PlayStation system page stays under its visible word ceiling.
The site serves install.sh at /retrobios/i and install.ps1 at
/retrobios/i.ps1. generate_site.py writes both copies with the
installer address set to install.py at the commit the site is built
from (GITHUB_SHA in CI, HEAD locally), so the pinned SHA-256 and the
downloaded installer always come from one commit, including while a
deploy is still running.
Every command a user copies names the short address. The raw form
stays on the installer page as the fallback. Deploy Site now runs on
changes to install.sh and install.ps1. The install box weight keeps
the address length its ratio was fitted on, so no console page moves.
region_check and the site name a boot on the core's built-in BIOS as such: the other regions' files were refused. A boot is called another region's BIOS only when the core's log names that file, and a boot whose log names no BIOS says so.
core_probe rejudge appends, for each live record the oracle now judges otherwise, a record with the same runs judged again: verdicts, bios_used and inputs.oracle recomputed, supersedes the earlier one, judged_at and judged_by beside recorded_at. Records of another probe are left to it. The campaign reads a superseded record as its successor, and a file the log shows refused or loaded settles validation with no BIOS screen.
Record version 5: every repetition and variant carries bios_used, staged:<dest> when a load line of the core names the file, builtin when a fallback line says it ran its own BIOS, unknown otherwise. An opened file is no proof. The oracle gains fallback_patterns and load_patterns per core, and a record judged again carries judged_at and judged_by.
run_cli entry point, --platform through platform_arg with -p and the registry's suggestion, help on every option, typed probe_lock, options documented, the module in the testing guide. The platform Tested rows fold under one line, run output joined without semicolons, and the measurements page section fits its start-level budget.
Four cores and three games put 430 words on the PlayStation page, past its word ceiling. A system page now shows one line per core, and each game's outcome with its four records waits in a closed fold.
On a console's own page the region block came after the title with no section above it, h1 then h3, and validate_site refused the PlayStation and Saturn pages. It is a section there and stays under the console heading on a maker's page.
PCSX-ReARMed opens the three regional BIOS files whatever the disc, and the rule called that a confirmation because one of them held the region. A run that opened its own region's file and the others too now settles nothing.
The details box takes a plain border instead of a side stripe, dots are rounded squares instead of pills, the strip's label joins variants with full stops, and one file tried reads as 1 file.
Twelve region records made the Beetle Saturn Tested block 266 words where the page allows 120. A region run reads as one outcome per game on the system page, so the emulator page now names them in one line.
Without a BIOS, PCSX2 logs and shows 'PCSX2 requires a PlayStation 2 BIOS in order to run'. That line is now one of PCSX2's error patterns, for the next run's verdict: this record was judged without it.
Run output on a platform's Tested block is escaped for HTML, MDX and Markdown. inline_html links only http(s) and relative targets, quotes escaped. Pack probe runs stay off the emulator pages, shown on the platform's.
Every value a run produced (claim, method, statement, host, flags, variant labels, core, content and file names) is escaped with quotes and its braces made inert for MDX, never read as Markdown. A record link is written only for an http(s) or relative target. Reference runs fold with the tests from before the oracle.
The first runs showed it: beetle_psx reaches the game of the US disc on the Japanese and European BIOS, through a different licence screen, so its frames differ from the full run while the oracle says boots. The interim rule read those frames as a failure. A variant now boots on the same frames as the full run or by its oracle verdict, and otherwise has no answer.
The interpreter bind took each link's grandparent, so ~/.local/bin/python3 exposed all of ~/.local inside the sandbox. A tree is now bound only when it is a Python prefix (lib/pythonX.Y) or a venv (pyvenv.cfg), never / or the home folder or above it. Every other link is bound as a single file, and a path inside a bound tree is not bound again, since bwrap cannot mount over a link in a read-only tree.
The sandbox bound only /usr, /lib and /bin. On the CI runner setup-python puts Python under /opt/hostedtoolcache, so the worker could not start and every core_probe and campaign run crashed (7 tests, plus their replay without jsonschema). A venv, pyenv or uv do the same. The interpreter's prefix and every link on the way to it are now bound read-only, each at its own path. Reproduced with uv's Python 3.13 outside /usr, 92 tests pass under it and under the system Python.
Merging the card and platform branches left two SHORT_HASH definitions, 8 and 12. The later one silently won, so the platform rows showed 12 characters while their tests expected 8. One constant now, 12, the card anchor's length. The copy-button selector skips buttons that carry their own data-copy, and the tests follow the merged markup.
Each file is a details whose summary is the row: name, folder when it is not the bare name, status chip. The short hash sits on one copy button beside it, since a summary is itself a control. Opened, the evidence runs across the list. A list with no folder drops the column. Code spans in the evidence are plain text.
The lists split by the share of a page they render to (elements against 10,000, bytes against 600,000) instead of Markdown length. With its parts, Batocera goes from 12 pages, 18,854 elements and 1.21 MB to 10 pages, 13,671 and 1.11 MB. A part that folds back into its platform page redirects there, from the record in website/src/data/platform-parts.json.
A finding line is written again when the records behind it change, so an open finding cites the oracle runs rather than a frame comparison. A contradicted validation shows the checks the runs prove, content and size, beside the profile's list.
Abdessamad Derraz