clay-borg/tools/trials.py

356 lines
16 KiB
Python
Raw Normal View History

CB-WP-0027 T01-T04: the commentary track The meta view beside the table, and a note channel that provably cannot carry a move. T01 (ADR-0014). ADR-0007 D5 is SCOPED, NOT AMENDED, and the reason it was easy is that PointerFact::parse already refuses any unrecognised field -- a comment could not reach the command path even by accident. So /command carries pointer facts, /note carries text, and Note has no code path to GroundCommand. Comments live in trials/<date>-<slug>.md, not in ScenarioFile: a scenario is executed, replayed and hashed, and prose in it is data the runner must ignore, which is how a format rots. The state hash binds; round and step are for reading. And the retention question, decided before any comment was written: RAW NOTES NEVER LEAVE clay-borg. A note reaches ground-game only by being promoted to a register finding, by a human, with the wording chosen then -- "the DARVO sequence is infuriating" is useful signal and a bad way to open a message to the game's designer. T02. CSS grid, minmax(0,1fr) on both tracks -- load-bearing, because a grid child defaults to min-content width and without it the SVG table refuses to shrink and pushes the meta column off-screen, looking correct on the developer's monitor and broken everywhere else. Single-column fallback under 64rem. The running tally moved into the panel so it is visible WHILE PLAYING; it only appeared on the ending page before, and a score you see once the game is over informs nothing. T03. A plain <form method="post">, so the box works with the script disabled; the command channel needs JavaScript because a drag is not a form submission, a comment is one. 303 See Other so a reload does not re-post. esc()'s first hostile input: <script>alert(1)</script> renders escaped AND STILL READABLE -- escaping that eats the player's words is its own defect. Verified over real HTTP: note posted 303, hostile note stored as text, empty note refused 400, game did not advance. T04. tools/trials.py and make trials. THE REPORT'S DESIGN CHANGED BECAUSE I RAN IT: the first version called any note without a recording an orphan, so a live session reported every note as broken -- the recording is only written at game end. A metric that cries wolf is one nobody reads, which is the exact failure this pass exists to prevent. Now ok / pending / orphan, and only orphan is a target-0 number. The self-test exercises the REPORTING path, not just the parser, because design-baseline.py had a green self-test and an unexercised reporting path and that is where it rotted. And a latent Makefile defect surfaced: make trials did nothing, because trials is also a directory and Make saw an up-to-date file. design, difficulty and trials -- added by CB-WP-0022, CB-WP-0025 and this pass -- were ALL missing from .PHONY; only the one that collided revealed it. make all: exit 0. 49 render tests, 26 cb-play, loop-lint clean. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
2026-08-06 10:37:46 +02:00
#!/usr/bin/env python3
"""trials — what the players said, and where (CB-WP-0027 T04).
**Surfacing is the deliverable, not storage.** ADR-0014 D6: a commentary
feature that stores comments and shows them nowhere would be this
project's signature failure in a new medium — after the message that sat
unread for four days and the ten rulings that arrived and were never
collected.
So this reports every note from every trial log, with its position and its
age, and flags the ones whose position can no longer be reached.
Reuses `design.py`'s parsing shape because the trial log deliberately
reuses `FindingRegister.md`'s idiom: a table between HTML-comment markers
inside a readable Markdown file.
"""
import os, re, sys, glob, datetime
ROOT = os.path.dirname(os.path.dirname(os.path.abspath(__file__)))
TRIALS = os.path.join(ROOT, "trials")
BEGIN = "<!-- trial-log:begin -->"
END = "<!-- trial-log:end -->"
# GameDesign §3.1's figure, shared rather than reinvented: a note that
# cannot expire ages into an apparent finding.
NOTE_EXPIRY_DAYS = 30
# The trial-log table's columns, named once. `write_trial_log` in
# `hotseat.rs` writes them; this reads them; a row that is not this shape
# is an error rather than a row to skip.
CB-WP-0033: a game is the unit make trials reported "positions unreachable: 4, target 0 — the recording exists but the position moved". All four were false. A recording holds one hash, the final state, and reachability asked whether the note's hash was in that file — so a mid-game note could never match, and a post-game note from any but the last game could not either. Instance 8 of the ADR-0018 family: vary only WHEN a note was written and the answer flips, with nothing having moved. The root cause was not the metric. play again reused state belonging to a game: it overwrote the previous game's recording (data loss), never cleared the journal (game 2's log opened with game 1's commands), and so a note's command index pointed into a recording without those commands. Fixing reachability alone would have gone green while a session still destroyed its own evidence. A note now binds by (game, after) — an index into the recording's own commands list, which a reader can replay to. The hash keeps a job as the integrity check at the end of a game, where it can actually fail. Game 1 keeps the path it was given, so GameDesign §5's documented invocation is unchanged; later games get -2, -3 and nothing is overwritten. Legacy 5-column logs stay readable and are reported as legacy, never as orphans — an unsubstantiated orphan claim is the defect being fixed. All three fixes mutation-proven, including at the call site via a real two-game session. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
2026-08-07 17:41:11 +02:00
COLUMNS = ("n", "game", "after", "round", "step", "state_hash", "comment")
# ADR-0019 D4. Logs written before the binding was fixed. They are parsed
# and reported, and they are NOT called orphans: their positions were
# never checkable by anything this repo can run, and an unsubstantiated
# orphan claim is the exact defect ADR-0019 exists to delete.
LEGACY_COLUMNS = ("n", "round", "step", "state_hash", "comment")
CB-WP-0027 T01-T04: the commentary track The meta view beside the table, and a note channel that provably cannot carry a move. T01 (ADR-0014). ADR-0007 D5 is SCOPED, NOT AMENDED, and the reason it was easy is that PointerFact::parse already refuses any unrecognised field -- a comment could not reach the command path even by accident. So /command carries pointer facts, /note carries text, and Note has no code path to GroundCommand. Comments live in trials/<date>-<slug>.md, not in ScenarioFile: a scenario is executed, replayed and hashed, and prose in it is data the runner must ignore, which is how a format rots. The state hash binds; round and step are for reading. And the retention question, decided before any comment was written: RAW NOTES NEVER LEAVE clay-borg. A note reaches ground-game only by being promoted to a register finding, by a human, with the wording chosen then -- "the DARVO sequence is infuriating" is useful signal and a bad way to open a message to the game's designer. T02. CSS grid, minmax(0,1fr) on both tracks -- load-bearing, because a grid child defaults to min-content width and without it the SVG table refuses to shrink and pushes the meta column off-screen, looking correct on the developer's monitor and broken everywhere else. Single-column fallback under 64rem. The running tally moved into the panel so it is visible WHILE PLAYING; it only appeared on the ending page before, and a score you see once the game is over informs nothing. T03. A plain <form method="post">, so the box works with the script disabled; the command channel needs JavaScript because a drag is not a form submission, a comment is one. 303 See Other so a reload does not re-post. esc()'s first hostile input: <script>alert(1)</script> renders escaped AND STILL READABLE -- escaping that eats the player's words is its own defect. Verified over real HTTP: note posted 303, hostile note stored as text, empty note refused 400, game did not advance. T04. tools/trials.py and make trials. THE REPORT'S DESIGN CHANGED BECAUSE I RAN IT: the first version called any note without a recording an orphan, so a live session reported every note as broken -- the recording is only written at game end. A metric that cries wolf is one nobody reads, which is the exact failure this pass exists to prevent. Now ok / pending / orphan, and only orphan is a target-0 number. The self-test exercises the REPORTING path, not just the parser, because design-baseline.py had a green self-test and an unexercised reporting path and that is where it rotted. And a latent Makefile defect surfaced: make trials did nothing, because trials is also a directory and Make saw an up-to-date file. design, difficulty and trials -- added by CB-WP-0022, CB-WP-0025 and this pass -- were ALL missing from .PHONY; only the one that collided revealed it. make all: exit 0. 49 render tests, 26 cb-play, loop-lint clean. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
2026-08-06 10:37:46 +02:00
def parse(text):
"""Rows between the markers. A log without the block is an error, not
an empty log silently reporting zero notes for a file full of them
is precisely the failure this tool exists to prevent."""
if BEGIN not in text or END not in text:
raise ValueError("no trial-log block")
block = text.split(BEGIN)[1].split(END)[0]
rows = []
for line in block.splitlines():
line = line.strip()
if not line.startswith("|") or line.startswith("|---"):
continue
cells = [c.strip() for c in line.strip("|").split("|")]
if cells[0] == "n":
CB-WP-0027 T01-T04: the commentary track The meta view beside the table, and a note channel that provably cannot carry a move. T01 (ADR-0014). ADR-0007 D5 is SCOPED, NOT AMENDED, and the reason it was easy is that PointerFact::parse already refuses any unrecognised field -- a comment could not reach the command path even by accident. So /command carries pointer facts, /note carries text, and Note has no code path to GroundCommand. Comments live in trials/<date>-<slug>.md, not in ScenarioFile: a scenario is executed, replayed and hashed, and prose in it is data the runner must ignore, which is how a format rots. The state hash binds; round and step are for reading. And the retention question, decided before any comment was written: RAW NOTES NEVER LEAVE clay-borg. A note reaches ground-game only by being promoted to a register finding, by a human, with the wording chosen then -- "the DARVO sequence is infuriating" is useful signal and a bad way to open a message to the game's designer. T02. CSS grid, minmax(0,1fr) on both tracks -- load-bearing, because a grid child defaults to min-content width and without it the SVG table refuses to shrink and pushes the meta column off-screen, looking correct on the developer's monitor and broken everywhere else. Single-column fallback under 64rem. The running tally moved into the panel so it is visible WHILE PLAYING; it only appeared on the ending page before, and a score you see once the game is over informs nothing. T03. A plain <form method="post">, so the box works with the script disabled; the command channel needs JavaScript because a drag is not a form submission, a comment is one. 303 See Other so a reload does not re-post. esc()'s first hostile input: <script>alert(1)</script> renders escaped AND STILL READABLE -- escaping that eats the player's words is its own defect. Verified over real HTTP: note posted 303, hostile note stored as text, empty note refused 400, game did not advance. T04. tools/trials.py and make trials. THE REPORT'S DESIGN CHANGED BECAUSE I RAN IT: the first version called any note without a recording an orphan, so a live session reported every note as broken -- the recording is only written at game end. A metric that cries wolf is one nobody reads, which is the exact failure this pass exists to prevent. Now ok / pending / orphan, and only orphan is a target-0 number. The self-test exercises the REPORTING path, not just the parser, because design-baseline.py had a green self-test and an unexercised reporting path and that is where it rotted. And a latent Makefile defect surfaced: make trials did nothing, because trials is also a directory and Make saw an up-to-date file. design, difficulty and trials -- added by CB-WP-0022, CB-WP-0025 and this pass -- were ALL missing from .PHONY; only the one that collided revealed it. make all: exit 0. 49 render tests, 26 cb-play, loop-lint clean. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
2026-08-06 10:37:46 +02:00
continue
# CB-WP-0032. A wrong column count RAISES; it used to `continue`.
#
# The docstring above already says silently reporting zero notes
# for a file full of them is the failure this tool exists to
# prevent -- and a `continue` here did exactly that for every row
# at once. Adding a sixth column to the trial log (which CB-WP-0032
# considered, and did not do) would have emptied every existing
# log without a word.
CB-WP-0033: a game is the unit make trials reported "positions unreachable: 4, target 0 — the recording exists but the position moved". All four were false. A recording holds one hash, the final state, and reachability asked whether the note's hash was in that file — so a mid-game note could never match, and a post-game note from any but the last game could not either. Instance 8 of the ADR-0018 family: vary only WHEN a note was written and the answer flips, with nothing having moved. The root cause was not the metric. play again reused state belonging to a game: it overwrote the previous game's recording (data loss), never cleared the journal (game 2's log opened with game 1's commands), and so a note's command index pointed into a recording without those commands. Fixing reachability alone would have gone green while a session still destroyed its own evidence. A note now binds by (game, after) — an index into the recording's own commands list, which a reader can replay to. The hash keeps a job as the integrity check at the end of a game, where it can actually fail. Game 1 keeps the path it was given, so GameDesign §5's documented invocation is unchanged; later games get -2, -3 and nothing is overwritten. Legacy 5-column logs stay readable and are reported as legacy, never as orphans — an unsubstantiated orphan claim is the defect being fixed. All three fixes mutation-proven, including at the call site via a real two-game session. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
2026-08-07 17:41:11 +02:00
if len(cells) == len(COLUMNS):
rows.append(dict(zip(COLUMNS, cells)))
elif len(cells) == len(LEGACY_COLUMNS):
row = dict(zip(LEGACY_COLUMNS, cells))
row["legacy"] = True
rows.append(row)
else:
raise ValueError(
CB-WP-0033: a game is the unit make trials reported "positions unreachable: 4, target 0 — the recording exists but the position moved". All four were false. A recording holds one hash, the final state, and reachability asked whether the note's hash was in that file — so a mid-game note could never match, and a post-game note from any but the last game could not either. Instance 8 of the ADR-0018 family: vary only WHEN a note was written and the answer flips, with nothing having moved. The root cause was not the metric. play again reused state belonging to a game: it overwrote the previous game's recording (data loss), never cleared the journal (game 2's log opened with game 1's commands), and so a note's command index pointed into a recording without those commands. Fixing reachability alone would have gone green while a session still destroyed its own evidence. A note now binds by (game, after) — an index into the recording's own commands list, which a reader can replay to. The hash keeps a job as the integrity check at the end of a game, where it can actually fail. Game 1 keeps the path it was given, so GameDesign §5's documented invocation is unchanged; later games get -2, -3 and nothing is overwritten. Legacy 5-column logs stay readable and are reported as legacy, never as orphans — an unsubstantiated orphan claim is the defect being fixed. All three fixes mutation-proven, including at the call site via a real two-game session. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
2026-08-07 17:41:11 +02:00
f"trial-log row has {len(cells)} columns, expected "
f"{len(COLUMNS)} or {len(LEGACY_COLUMNS)}: {line!r}"
)
CB-WP-0027 T01-T04: the commentary track The meta view beside the table, and a note channel that provably cannot carry a move. T01 (ADR-0014). ADR-0007 D5 is SCOPED, NOT AMENDED, and the reason it was easy is that PointerFact::parse already refuses any unrecognised field -- a comment could not reach the command path even by accident. So /command carries pointer facts, /note carries text, and Note has no code path to GroundCommand. Comments live in trials/<date>-<slug>.md, not in ScenarioFile: a scenario is executed, replayed and hashed, and prose in it is data the runner must ignore, which is how a format rots. The state hash binds; round and step are for reading. And the retention question, decided before any comment was written: RAW NOTES NEVER LEAVE clay-borg. A note reaches ground-game only by being promoted to a register finding, by a human, with the wording chosen then -- "the DARVO sequence is infuriating" is useful signal and a bad way to open a message to the game's designer. T02. CSS grid, minmax(0,1fr) on both tracks -- load-bearing, because a grid child defaults to min-content width and without it the SVG table refuses to shrink and pushes the meta column off-screen, looking correct on the developer's monitor and broken everywhere else. Single-column fallback under 64rem. The running tally moved into the panel so it is visible WHILE PLAYING; it only appeared on the ending page before, and a score you see once the game is over informs nothing. T03. A plain <form method="post">, so the box works with the script disabled; the command channel needs JavaScript because a drag is not a form submission, a comment is one. 303 See Other so a reload does not re-post. esc()'s first hostile input: <script>alert(1)</script> renders escaped AND STILL READABLE -- escaping that eats the player's words is its own defect. Verified over real HTTP: note posted 303, hostile note stored as text, empty note refused 400, game did not advance. T04. tools/trials.py and make trials. THE REPORT'S DESIGN CHANGED BECAUSE I RAN IT: the first version called any note without a recording an orphan, so a live session reported every note as broken -- the recording is only written at game end. A metric that cries wolf is one nobody reads, which is the exact failure this pass exists to prevent. Now ok / pending / orphan, and only orphan is a target-0 number. The self-test exercises the REPORTING path, not just the parser, because design-baseline.py had a green self-test and an unexercised reporting path and that is where it rotted. And a latent Makefile defect surfaced: make trials did nothing, because trials is also a directory and Make saw an up-to-date file. design, difficulty and trials -- added by CB-WP-0022, CB-WP-0025 and this pass -- were ALL missing from .PHONY; only the one that collided revealed it. make all: exit 0. 49 render tests, 26 cb-play, loop-lint clean. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
2026-08-06 10:37:46 +02:00
return rows
CB-WP-0033: a game is the unit make trials reported "positions unreachable: 4, target 0 — the recording exists but the position moved". All four were false. A recording holds one hash, the final state, and reachability asked whether the note's hash was in that file — so a mid-game note could never match, and a post-game note from any but the last game could not either. Instance 8 of the ADR-0018 family: vary only WHEN a note was written and the answer flips, with nothing having moved. The root cause was not the metric. play again reused state belonging to a game: it overwrote the previous game's recording (data loss), never cleared the journal (game 2's log opened with game 1's commands), and so a note's command index pointed into a recording without those commands. Fixing reachability alone would have gone green while a session still destroyed its own evidence. A note now binds by (game, after) — an index into the recording's own commands list, which a reader can replay to. The hash keeps a job as the integrity check at the end of a game, where it can actually fail. Game 1 keeps the path it was given, so GameDesign §5's documented invocation is unchanged; later games get -2, -3 and nothing is overwritten. Legacy 5-column logs stay readable and are reported as legacy, never as orphans — an unsubstantiated orphan claim is the defect being fixed. All three fixes mutation-proven, including at the call site via a real two-game session. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
2026-08-07 17:41:11 +02:00
def recording_for(md_path, row):
"""The recording holding this note's game (ADR-0019 D2).
Game 1 keeps the base path; later games carry `-2`, `-3`. A legacy row
names no game, so it gets the base path and is judged as legacy.
"""
stem = os.path.splitext(md_path)[0]
n = int(row.get("game", 1) or 1)
path = f"{stem}.yaml" if n <= 1 else f"{stem}-{n}.yaml"
return path if os.path.exists(path) else None
def command_count(recording):
"""How many commands the recording holds — the range a note's `after`
must fall inside. Counted from the `commands:` block, because that is
the list `after` indexes into."""
text = open(recording).read()
if "commands:" not in text:
return 0
block = text.split("commands:", 1)[1].split("\nexpect:", 1)[0]
return sum(1 for line in block.splitlines() if line.startswith("- actor:"))
CB-WP-0027 T01-T04: the commentary track The meta view beside the table, and a note channel that provably cannot carry a move. T01 (ADR-0014). ADR-0007 D5 is SCOPED, NOT AMENDED, and the reason it was easy is that PointerFact::parse already refuses any unrecognised field -- a comment could not reach the command path even by accident. So /command carries pointer facts, /note carries text, and Note has no code path to GroundCommand. Comments live in trials/<date>-<slug>.md, not in ScenarioFile: a scenario is executed, replayed and hashed, and prose in it is data the runner must ignore, which is how a format rots. The state hash binds; round and step are for reading. And the retention question, decided before any comment was written: RAW NOTES NEVER LEAVE clay-borg. A note reaches ground-game only by being promoted to a register finding, by a human, with the wording chosen then -- "the DARVO sequence is infuriating" is useful signal and a bad way to open a message to the game's designer. T02. CSS grid, minmax(0,1fr) on both tracks -- load-bearing, because a grid child defaults to min-content width and without it the SVG table refuses to shrink and pushes the meta column off-screen, looking correct on the developer's monitor and broken everywhere else. Single-column fallback under 64rem. The running tally moved into the panel so it is visible WHILE PLAYING; it only appeared on the ending page before, and a score you see once the game is over informs nothing. T03. A plain <form method="post">, so the box works with the script disabled; the command channel needs JavaScript because a drag is not a form submission, a comment is one. 303 See Other so a reload does not re-post. esc()'s first hostile input: <script>alert(1)</script> renders escaped AND STILL READABLE -- escaping that eats the player's words is its own defect. Verified over real HTTP: note posted 303, hostile note stored as text, empty note refused 400, game did not advance. T04. tools/trials.py and make trials. THE REPORT'S DESIGN CHANGED BECAUSE I RAN IT: the first version called any note without a recording an orphan, so a live session reported every note as broken -- the recording is only written at game end. A metric that cries wolf is one nobody reads, which is the exact failure this pass exists to prevent. Now ok / pending / orphan, and only orphan is a target-0 number. The self-test exercises the REPORTING path, not just the parser, because design-baseline.py had a green self-test and an unexercised reporting path and that is where it rotted. And a latent Makefile defect surfaced: make trials did nothing, because trials is also a directory and Make saw an up-to-date file. design, difficulty and trials -- added by CB-WP-0022, CB-WP-0025 and this pass -- were ALL missing from .PHONY; only the one that collided revealed it. make all: exit 0. 49 render tests, 26 cb-play, loop-lint clean. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
2026-08-06 10:37:46 +02:00
def logs(root=TRIALS):
"""Every trial log, with its recording beside it if there is one."""
out = []
for path in sorted(glob.glob(os.path.join(root, "*.md"))):
try:
rows = parse(open(path).read())
except ValueError as e:
# The REASON, not a guess at it. This said "no trial-log block
# -- not a trial log?" for every failure, which is now wrong
# for two of them: a malformed row is not a missing block, and
# telling a maintainer to check for the wrong thing costs more
# than saying nothing.
print(f" WARN {os.path.basename(path)}: {e}")
CB-WP-0027 T01-T04: the commentary track The meta view beside the table, and a note channel that provably cannot carry a move. T01 (ADR-0014). ADR-0007 D5 is SCOPED, NOT AMENDED, and the reason it was easy is that PointerFact::parse already refuses any unrecognised field -- a comment could not reach the command path even by accident. So /command carries pointer facts, /note carries text, and Note has no code path to GroundCommand. Comments live in trials/<date>-<slug>.md, not in ScenarioFile: a scenario is executed, replayed and hashed, and prose in it is data the runner must ignore, which is how a format rots. The state hash binds; round and step are for reading. And the retention question, decided before any comment was written: RAW NOTES NEVER LEAVE clay-borg. A note reaches ground-game only by being promoted to a register finding, by a human, with the wording chosen then -- "the DARVO sequence is infuriating" is useful signal and a bad way to open a message to the game's designer. T02. CSS grid, minmax(0,1fr) on both tracks -- load-bearing, because a grid child defaults to min-content width and without it the SVG table refuses to shrink and pushes the meta column off-screen, looking correct on the developer's monitor and broken everywhere else. Single-column fallback under 64rem. The running tally moved into the panel so it is visible WHILE PLAYING; it only appeared on the ending page before, and a score you see once the game is over informs nothing. T03. A plain <form method="post">, so the box works with the script disabled; the command channel needs JavaScript because a drag is not a form submission, a comment is one. 303 See Other so a reload does not re-post. esc()'s first hostile input: <script>alert(1)</script> renders escaped AND STILL READABLE -- escaping that eats the player's words is its own defect. Verified over real HTTP: note posted 303, hostile note stored as text, empty note refused 400, game did not advance. T04. tools/trials.py and make trials. THE REPORT'S DESIGN CHANGED BECAUSE I RAN IT: the first version called any note without a recording an orphan, so a live session reported every note as broken -- the recording is only written at game end. A metric that cries wolf is one nobody reads, which is the exact failure this pass exists to prevent. Now ok / pending / orphan, and only orphan is a target-0 number. The self-test exercises the REPORTING path, not just the parser, because design-baseline.py had a green self-test and an unexercised reporting path and that is where it rotted. And a latent Makefile defect surfaced: make trials did nothing, because trials is also a directory and Make saw an up-to-date file. design, difficulty and trials -- added by CB-WP-0022, CB-WP-0025 and this pass -- were ALL missing from .PHONY; only the one that collided revealed it. make all: exit 0. 49 render tests, 26 cb-play, loop-lint clean. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
2026-08-06 10:37:46 +02:00
continue
recording = os.path.splitext(path)[0] + ".yaml"
out.append((path, rows, recording if os.path.exists(recording) else None))
return out
def reachability(row, recording):
"""Can this note's position be reached? One of three answers.
**"No recording" and "the hash is not in the recording" are different
states and must not be one number.** A note written mid-game is
pending the recording is only written when the game ends while a
note whose hash is absent from a recording that *exists* means the
position moved. Collapsing them makes every live session report
orphans, and a metric that cries wolf is one nobody reads, which is
the failure this whole pass is about.
Orphans are reported, never deleted (ADR-0014 D3): a moved position is
worth knowing, by the same reasoning that rewrote `gr-e01` rather than
retiring it.
"""
CB-WP-0033: a game is the unit make trials reported "positions unreachable: 4, target 0 — the recording exists but the position moved". All four were false. A recording holds one hash, the final state, and reachability asked whether the note's hash was in that file — so a mid-game note could never match, and a post-game note from any but the last game could not either. Instance 8 of the ADR-0018 family: vary only WHEN a note was written and the answer flips, with nothing having moved. The root cause was not the metric. play again reused state belonging to a game: it overwrote the previous game's recording (data loss), never cleared the journal (game 2's log opened with game 1's commands), and so a note's command index pointed into a recording without those commands. Fixing reachability alone would have gone green while a session still destroyed its own evidence. A note now binds by (game, after) — an index into the recording's own commands list, which a reader can replay to. The hash keeps a job as the integrity check at the end of a game, where it can actually fail. Game 1 keeps the path it was given, so GameDesign §5's documented invocation is unchanged; later games get -2, -3 and nothing is overwritten. Legacy 5-column logs stay readable and are reported as legacy, never as orphans — an unsubstantiated orphan claim is the defect being fixed. All three fixes mutation-proven, including at the call site via a real two-game session. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
2026-08-07 17:41:11 +02:00
if row.get("legacy"):
return "legacy"
CB-WP-0027 T01-T04: the commentary track The meta view beside the table, and a note channel that provably cannot carry a move. T01 (ADR-0014). ADR-0007 D5 is SCOPED, NOT AMENDED, and the reason it was easy is that PointerFact::parse already refuses any unrecognised field -- a comment could not reach the command path even by accident. So /command carries pointer facts, /note carries text, and Note has no code path to GroundCommand. Comments live in trials/<date>-<slug>.md, not in ScenarioFile: a scenario is executed, replayed and hashed, and prose in it is data the runner must ignore, which is how a format rots. The state hash binds; round and step are for reading. And the retention question, decided before any comment was written: RAW NOTES NEVER LEAVE clay-borg. A note reaches ground-game only by being promoted to a register finding, by a human, with the wording chosen then -- "the DARVO sequence is infuriating" is useful signal and a bad way to open a message to the game's designer. T02. CSS grid, minmax(0,1fr) on both tracks -- load-bearing, because a grid child defaults to min-content width and without it the SVG table refuses to shrink and pushes the meta column off-screen, looking correct on the developer's monitor and broken everywhere else. Single-column fallback under 64rem. The running tally moved into the panel so it is visible WHILE PLAYING; it only appeared on the ending page before, and a score you see once the game is over informs nothing. T03. A plain <form method="post">, so the box works with the script disabled; the command channel needs JavaScript because a drag is not a form submission, a comment is one. 303 See Other so a reload does not re-post. esc()'s first hostile input: <script>alert(1)</script> renders escaped AND STILL READABLE -- escaping that eats the player's words is its own defect. Verified over real HTTP: note posted 303, hostile note stored as text, empty note refused 400, game did not advance. T04. tools/trials.py and make trials. THE REPORT'S DESIGN CHANGED BECAUSE I RAN IT: the first version called any note without a recording an orphan, so a live session reported every note as broken -- the recording is only written at game end. A metric that cries wolf is one nobody reads, which is the exact failure this pass exists to prevent. Now ok / pending / orphan, and only orphan is a target-0 number. The self-test exercises the REPORTING path, not just the parser, because design-baseline.py had a green self-test and an unexercised reporting path and that is where it rotted. And a latent Makefile defect surfaced: make trials did nothing, because trials is also a directory and Make saw an up-to-date file. design, difficulty and trials -- added by CB-WP-0022, CB-WP-0025 and this pass -- were ALL missing from .PHONY; only the one that collided revealed it. make all: exit 0. 49 render tests, 26 cb-play, loop-lint clean. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
2026-08-06 10:37:46 +02:00
if not recording:
return "pending"
CB-WP-0033: a game is the unit make trials reported "positions unreachable: 4, target 0 — the recording exists but the position moved". All four were false. A recording holds one hash, the final state, and reachability asked whether the note's hash was in that file — so a mid-game note could never match, and a post-game note from any but the last game could not either. Instance 8 of the ADR-0018 family: vary only WHEN a note was written and the answer flips, with nothing having moved. The root cause was not the metric. play again reused state belonging to a game: it overwrote the previous game's recording (data loss), never cleared the journal (game 2's log opened with game 1's commands), and so a note's command index pointed into a recording without those commands. Fixing reachability alone would have gone green while a session still destroyed its own evidence. A note now binds by (game, after) — an index into the recording's own commands list, which a reader can replay to. The hash keeps a job as the integrity check at the end of a game, where it can actually fail. Game 1 keeps the path it was given, so GameDesign §5's documented invocation is unchanged; later games get -2, -3 and nothing is overwritten. Legacy 5-column logs stay readable and are reported as legacy, never as orphans — an unsubstantiated orphan claim is the defect being fixed. All three fixes mutation-proven, including at the call site via a real two-game session. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
2026-08-07 17:41:11 +02:00
after = int(row["after"])
n = command_count(recording)
if after > n:
# The note points past the end of its own game.
return "orphan"
if after == n:
# The one position the recording states outright. A mismatch here
# is real: the note claims the end of a game the recording does
# not agree with.
return "ok" if row["state_hash"] in open(recording).read() else "orphan"
# Reachable by replaying `after` commands of this recording.
return "ok"
CB-WP-0027 T01-T04: the commentary track The meta view beside the table, and a note channel that provably cannot carry a move. T01 (ADR-0014). ADR-0007 D5 is SCOPED, NOT AMENDED, and the reason it was easy is that PointerFact::parse already refuses any unrecognised field -- a comment could not reach the command path even by accident. So /command carries pointer facts, /note carries text, and Note has no code path to GroundCommand. Comments live in trials/<date>-<slug>.md, not in ScenarioFile: a scenario is executed, replayed and hashed, and prose in it is data the runner must ignore, which is how a format rots. The state hash binds; round and step are for reading. And the retention question, decided before any comment was written: RAW NOTES NEVER LEAVE clay-borg. A note reaches ground-game only by being promoted to a register finding, by a human, with the wording chosen then -- "the DARVO sequence is infuriating" is useful signal and a bad way to open a message to the game's designer. T02. CSS grid, minmax(0,1fr) on both tracks -- load-bearing, because a grid child defaults to min-content width and without it the SVG table refuses to shrink and pushes the meta column off-screen, looking correct on the developer's monitor and broken everywhere else. Single-column fallback under 64rem. The running tally moved into the panel so it is visible WHILE PLAYING; it only appeared on the ending page before, and a score you see once the game is over informs nothing. T03. A plain <form method="post">, so the box works with the script disabled; the command channel needs JavaScript because a drag is not a form submission, a comment is one. 303 See Other so a reload does not re-post. esc()'s first hostile input: <script>alert(1)</script> renders escaped AND STILL READABLE -- escaping that eats the player's words is its own defect. Verified over real HTTP: note posted 303, hostile note stored as text, empty note refused 400, game did not advance. T04. tools/trials.py and make trials. THE REPORT'S DESIGN CHANGED BECAUSE I RAN IT: the first version called any note without a recording an orphan, so a live session reported every note as broken -- the recording is only written at game end. A metric that cries wolf is one nobody reads, which is the exact failure this pass exists to prevent. Now ok / pending / orphan, and only orphan is a target-0 number. The self-test exercises the REPORTING path, not just the parser, because design-baseline.py had a green self-test and an unexercised reporting path and that is where it rotted. And a latent Makefile defect surfaced: make trials did nothing, because trials is also a directory and Make saw an up-to-date file. design, difficulty and trials -- added by CB-WP-0022, CB-WP-0025 and this pass -- were ALL missing from .PHONY; only the one that collided revealed it. make all: exit 0. 49 render tests, 26 cb-play, loop-lint clean. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
2026-08-06 10:37:46 +02:00
def age_days(path, today):
"""Dated from the log's filename (`YYYY-MM-DD-slug.md`), falling back
to mtime the name is the intent, the mtime is what happened."""
m = re.match(r"(\d{4}-\d{2}-\d{2})", os.path.basename(path))
if m:
try:
return (today - datetime.date.fromisoformat(m.group(1))).days
except ValueError:
pass
return (today - datetime.date.fromtimestamp(os.path.getmtime(path))).days
def report(root=TRIALS, today=None):
today = today or datetime.date.today()
if not os.path.isdir(root):
print("trials — no trials/ directory yet\n")
print(" Play one: cb-play --serve 0 --record trials/<date>-<slug>.yaml \\")
print(" --trial trials/<date>-<slug>.md")
return 0
found = logs(root)
print("trials — what the players said, and where\n")
CB-WP-0033: a game is the unit make trials reported "positions unreachable: 4, target 0 — the recording exists but the position moved". All four were false. A recording holds one hash, the final state, and reachability asked whether the note's hash was in that file — so a mid-game note could never match, and a post-game note from any but the last game could not either. Instance 8 of the ADR-0018 family: vary only WHEN a note was written and the answer flips, with nothing having moved. The root cause was not the metric. play again reused state belonging to a game: it overwrote the previous game's recording (data loss), never cleared the journal (game 2's log opened with game 1's commands), and so a note's command index pointed into a recording without those commands. Fixing reachability alone would have gone green while a session still destroyed its own evidence. A note now binds by (game, after) — an index into the recording's own commands list, which a reader can replay to. The hash keeps a job as the integrity check at the end of a game, where it can actually fail. Game 1 keeps the path it was given, so GameDesign §5's documented invocation is unchanged; later games get -2, -3 and nothing is overwritten. Legacy 5-column logs stay readable and are reported as legacy, never as orphans — an unsubstantiated orphan claim is the defect being fixed. All three fixes mutation-proven, including at the call site via a real two-game session. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
2026-08-07 17:41:11 +02:00
total, orphans, pending, expired, legacy = 0, 0, 0, 0, 0
for path, rows, _ in found:
CB-WP-0027 T01-T04: the commentary track The meta view beside the table, and a note channel that provably cannot carry a move. T01 (ADR-0014). ADR-0007 D5 is SCOPED, NOT AMENDED, and the reason it was easy is that PointerFact::parse already refuses any unrecognised field -- a comment could not reach the command path even by accident. So /command carries pointer facts, /note carries text, and Note has no code path to GroundCommand. Comments live in trials/<date>-<slug>.md, not in ScenarioFile: a scenario is executed, replayed and hashed, and prose in it is data the runner must ignore, which is how a format rots. The state hash binds; round and step are for reading. And the retention question, decided before any comment was written: RAW NOTES NEVER LEAVE clay-borg. A note reaches ground-game only by being promoted to a register finding, by a human, with the wording chosen then -- "the DARVO sequence is infuriating" is useful signal and a bad way to open a message to the game's designer. T02. CSS grid, minmax(0,1fr) on both tracks -- load-bearing, because a grid child defaults to min-content width and without it the SVG table refuses to shrink and pushes the meta column off-screen, looking correct on the developer's monitor and broken everywhere else. Single-column fallback under 64rem. The running tally moved into the panel so it is visible WHILE PLAYING; it only appeared on the ending page before, and a score you see once the game is over informs nothing. T03. A plain <form method="post">, so the box works with the script disabled; the command channel needs JavaScript because a drag is not a form submission, a comment is one. 303 See Other so a reload does not re-post. esc()'s first hostile input: <script>alert(1)</script> renders escaped AND STILL READABLE -- escaping that eats the player's words is its own defect. Verified over real HTTP: note posted 303, hostile note stored as text, empty note refused 400, game did not advance. T04. tools/trials.py and make trials. THE REPORT'S DESIGN CHANGED BECAUSE I RAN IT: the first version called any note without a recording an orphan, so a live session reported every note as broken -- the recording is only written at game end. A metric that cries wolf is one nobody reads, which is the exact failure this pass exists to prevent. Now ok / pending / orphan, and only orphan is a target-0 number. The self-test exercises the REPORTING path, not just the parser, because design-baseline.py had a green self-test and an unexercised reporting path and that is where it rotted. And a latent Makefile defect surfaced: make trials did nothing, because trials is also a directory and Make saw an up-to-date file. design, difficulty and trials -- added by CB-WP-0022, CB-WP-0025 and this pass -- were ALL missing from .PHONY; only the one that collided revealed it. make all: exit 0. 49 render tests, 26 cb-play, loop-lint clean. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
2026-08-06 10:37:46 +02:00
name = os.path.basename(path)
age = age_days(path, today)
CB-WP-0033: a game is the unit make trials reported "positions unreachable: 4, target 0 — the recording exists but the position moved". All four were false. A recording holds one hash, the final state, and reachability asked whether the note's hash was in that file — so a mid-game note could never match, and a post-game note from any but the last game could not either. Instance 8 of the ADR-0018 family: vary only WHEN a note was written and the answer flips, with nothing having moved. The root cause was not the metric. play again reused state belonging to a game: it overwrote the previous game's recording (data loss), never cleared the journal (game 2's log opened with game 1's commands), and so a note's command index pointed into a recording without those commands. Fixing reachability alone would have gone green while a session still destroyed its own evidence. A note now binds by (game, after) — an index into the recording's own commands list, which a reader can replay to. The hash keeps a job as the integrity check at the end of a game, where it can actually fail. Game 1 keeps the path it was given, so GameDesign §5's documented invocation is unchanged; later games get -2, -3 and nothing is overwritten. Legacy 5-column logs stay readable and are reported as legacy, never as orphans — an unsubstantiated orphan claim is the defect being fixed. All three fixes mutation-proven, including at the call site via a real two-game session. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
2026-08-07 17:41:11 +02:00
print(f" {name} ({len(rows)} note(s), {age}d)")
CB-WP-0027 T01-T04: the commentary track The meta view beside the table, and a note channel that provably cannot carry a move. T01 (ADR-0014). ADR-0007 D5 is SCOPED, NOT AMENDED, and the reason it was easy is that PointerFact::parse already refuses any unrecognised field -- a comment could not reach the command path even by accident. So /command carries pointer facts, /note carries text, and Note has no code path to GroundCommand. Comments live in trials/<date>-<slug>.md, not in ScenarioFile: a scenario is executed, replayed and hashed, and prose in it is data the runner must ignore, which is how a format rots. The state hash binds; round and step are for reading. And the retention question, decided before any comment was written: RAW NOTES NEVER LEAVE clay-borg. A note reaches ground-game only by being promoted to a register finding, by a human, with the wording chosen then -- "the DARVO sequence is infuriating" is useful signal and a bad way to open a message to the game's designer. T02. CSS grid, minmax(0,1fr) on both tracks -- load-bearing, because a grid child defaults to min-content width and without it the SVG table refuses to shrink and pushes the meta column off-screen, looking correct on the developer's monitor and broken everywhere else. Single-column fallback under 64rem. The running tally moved into the panel so it is visible WHILE PLAYING; it only appeared on the ending page before, and a score you see once the game is over informs nothing. T03. A plain <form method="post">, so the box works with the script disabled; the command channel needs JavaScript because a drag is not a form submission, a comment is one. 303 See Other so a reload does not re-post. esc()'s first hostile input: <script>alert(1)</script> renders escaped AND STILL READABLE -- escaping that eats the player's words is its own defect. Verified over real HTTP: note posted 303, hostile note stored as text, empty note refused 400, game did not advance. T04. tools/trials.py and make trials. THE REPORT'S DESIGN CHANGED BECAUSE I RAN IT: the first version called any note without a recording an orphan, so a live session reported every note as broken -- the recording is only written at game end. A metric that cries wolf is one nobody reads, which is the exact failure this pass exists to prevent. Now ok / pending / orphan, and only orphan is a target-0 number. The self-test exercises the REPORTING path, not just the parser, because design-baseline.py had a green self-test and an unexercised reporting path and that is where it rotted. And a latent Makefile defect surfaced: make trials did nothing, because trials is also a directory and Make saw an up-to-date file. design, difficulty and trials -- added by CB-WP-0022, CB-WP-0025 and this pass -- were ALL missing from .PHONY; only the one that collided revealed it. make all: exit 0. 49 render tests, 26 cb-play, loop-lint clean. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
2026-08-06 10:37:46 +02:00
for r in rows:
total += 1
CB-WP-0033: a game is the unit make trials reported "positions unreachable: 4, target 0 — the recording exists but the position moved". All four were false. A recording holds one hash, the final state, and reachability asked whether the note's hash was in that file — so a mid-game note could never match, and a post-game note from any but the last game could not either. Instance 8 of the ADR-0018 family: vary only WHEN a note was written and the answer flips, with nothing having moved. The root cause was not the metric. play again reused state belonging to a game: it overwrote the previous game's recording (data loss), never cleared the journal (game 2's log opened with game 1's commands), and so a note's command index pointed into a recording without those commands. Fixing reachability alone would have gone green while a session still destroyed its own evidence. A note now binds by (game, after) — an index into the recording's own commands list, which a reader can replay to. The hash keeps a job as the integrity check at the end of a game, where it can actually fail. Game 1 keeps the path it was given, so GameDesign §5's documented invocation is unchanged; later games get -2, -3 and nothing is overwritten. Legacy 5-column logs stay readable and are reported as legacy, never as orphans — an unsubstantiated orphan claim is the defect being fixed. All three fixes mutation-proven, including at the call site via a real two-game session. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
2026-08-07 17:41:11 +02:00
# ADR-0019 D2: each note names its own game, and each game has
# its own recording. One recording per FILE was the assumption
# that made `play again` overwrite the evidence.
rec = recording_for(path, r)
state = reachability(r, rec)
CB-WP-0027 T01-T04: the commentary track The meta view beside the table, and a note channel that provably cannot carry a move. T01 (ADR-0014). ADR-0007 D5 is SCOPED, NOT AMENDED, and the reason it was easy is that PointerFact::parse already refuses any unrecognised field -- a comment could not reach the command path even by accident. So /command carries pointer facts, /note carries text, and Note has no code path to GroundCommand. Comments live in trials/<date>-<slug>.md, not in ScenarioFile: a scenario is executed, replayed and hashed, and prose in it is data the runner must ignore, which is how a format rots. The state hash binds; round and step are for reading. And the retention question, decided before any comment was written: RAW NOTES NEVER LEAVE clay-borg. A note reaches ground-game only by being promoted to a register finding, by a human, with the wording chosen then -- "the DARVO sequence is infuriating" is useful signal and a bad way to open a message to the game's designer. T02. CSS grid, minmax(0,1fr) on both tracks -- load-bearing, because a grid child defaults to min-content width and without it the SVG table refuses to shrink and pushes the meta column off-screen, looking correct on the developer's monitor and broken everywhere else. Single-column fallback under 64rem. The running tally moved into the panel so it is visible WHILE PLAYING; it only appeared on the ending page before, and a score you see once the game is over informs nothing. T03. A plain <form method="post">, so the box works with the script disabled; the command channel needs JavaScript because a drag is not a form submission, a comment is one. 303 See Other so a reload does not re-post. esc()'s first hostile input: <script>alert(1)</script> renders escaped AND STILL READABLE -- escaping that eats the player's words is its own defect. Verified over real HTTP: note posted 303, hostile note stored as text, empty note refused 400, game did not advance. T04. tools/trials.py and make trials. THE REPORT'S DESIGN CHANGED BECAUSE I RAN IT: the first version called any note without a recording an orphan, so a live session reported every note as broken -- the recording is only written at game end. A metric that cries wolf is one nobody reads, which is the exact failure this pass exists to prevent. Now ok / pending / orphan, and only orphan is a target-0 number. The self-test exercises the REPORTING path, not just the parser, because design-baseline.py had a green self-test and an unexercised reporting path and that is where it rotted. And a latent Makefile defect surfaced: make trials did nothing, because trials is also a directory and Make saw an up-to-date file. design, difficulty and trials -- added by CB-WP-0022, CB-WP-0025 and this pass -- were ALL missing from .PHONY; only the one that collided revealed it. make all: exit 0. 49 render tests, 26 cb-play, loop-lint clean. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
2026-08-06 10:37:46 +02:00
orphans += state == "orphan"
pending += state == "pending"
CB-WP-0033: a game is the unit make trials reported "positions unreachable: 4, target 0 — the recording exists but the position moved". All four were false. A recording holds one hash, the final state, and reachability asked whether the note's hash was in that file — so a mid-game note could never match, and a post-game note from any but the last game could not either. Instance 8 of the ADR-0018 family: vary only WHEN a note was written and the answer flips, with nothing having moved. The root cause was not the metric. play again reused state belonging to a game: it overwrote the previous game's recording (data loss), never cleared the journal (game 2's log opened with game 1's commands), and so a note's command index pointed into a recording without those commands. Fixing reachability alone would have gone green while a session still destroyed its own evidence. A note now binds by (game, after) — an index into the recording's own commands list, which a reader can replay to. The hash keeps a job as the integrity check at the end of a game, where it can actually fail. Game 1 keeps the path it was given, so GameDesign §5's documented invocation is unchanged; later games get -2, -3 and nothing is overwritten. Legacy 5-column logs stay readable and are reported as legacy, never as orphans — an unsubstantiated orphan claim is the defect being fixed. All three fixes mutation-proven, including at the call site via a real two-game session. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
2026-08-07 17:41:11 +02:00
legacy += state == "legacy"
CB-WP-0027 T01-T04: the commentary track The meta view beside the table, and a note channel that provably cannot carry a move. T01 (ADR-0014). ADR-0007 D5 is SCOPED, NOT AMENDED, and the reason it was easy is that PointerFact::parse already refuses any unrecognised field -- a comment could not reach the command path even by accident. So /command carries pointer facts, /note carries text, and Note has no code path to GroundCommand. Comments live in trials/<date>-<slug>.md, not in ScenarioFile: a scenario is executed, replayed and hashed, and prose in it is data the runner must ignore, which is how a format rots. The state hash binds; round and step are for reading. And the retention question, decided before any comment was written: RAW NOTES NEVER LEAVE clay-borg. A note reaches ground-game only by being promoted to a register finding, by a human, with the wording chosen then -- "the DARVO sequence is infuriating" is useful signal and a bad way to open a message to the game's designer. T02. CSS grid, minmax(0,1fr) on both tracks -- load-bearing, because a grid child defaults to min-content width and without it the SVG table refuses to shrink and pushes the meta column off-screen, looking correct on the developer's monitor and broken everywhere else. Single-column fallback under 64rem. The running tally moved into the panel so it is visible WHILE PLAYING; it only appeared on the ending page before, and a score you see once the game is over informs nothing. T03. A plain <form method="post">, so the box works with the script disabled; the command channel needs JavaScript because a drag is not a form submission, a comment is one. 303 See Other so a reload does not re-post. esc()'s first hostile input: <script>alert(1)</script> renders escaped AND STILL READABLE -- escaping that eats the player's words is its own defect. Verified over real HTTP: note posted 303, hostile note stored as text, empty note refused 400, game did not advance. T04. tools/trials.py and make trials. THE REPORT'S DESIGN CHANGED BECAUSE I RAN IT: the first version called any note without a recording an orphan, so a live session reported every note as broken -- the recording is only written at game end. A metric that cries wolf is one nobody reads, which is the exact failure this pass exists to prevent. Now ok / pending / orphan, and only orphan is a target-0 number. The self-test exercises the REPORTING path, not just the parser, because design-baseline.py had a green self-test and an unexercised reporting path and that is where it rotted. And a latent Makefile defect surfaced: make trials did nothing, because trials is also a directory and Make saw an up-to-date file. design, difficulty and trials -- added by CB-WP-0022, CB-WP-0025 and this pass -- were ALL missing from .PHONY; only the one that collided revealed it. make all: exit 0. 49 render tests, 26 cb-play, loop-lint clean. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
2026-08-06 10:37:46 +02:00
expired += age > NOTE_EXPIRY_DAYS
CB-WP-0033: a game is the unit make trials reported "positions unreachable: 4, target 0 — the recording exists but the position moved". All four were false. A recording holds one hash, the final state, and reachability asked whether the note's hash was in that file — so a mid-game note could never match, and a post-game note from any but the last game could not either. Instance 8 of the ADR-0018 family: vary only WHEN a note was written and the answer flips, with nothing having moved. The root cause was not the metric. play again reused state belonging to a game: it overwrote the previous game's recording (data loss), never cleared the journal (game 2's log opened with game 1's commands), and so a note's command index pointed into a recording without those commands. Fixing reachability alone would have gone green while a session still destroyed its own evidence. A note now binds by (game, after) — an index into the recording's own commands list, which a reader can replay to. The hash keeps a job as the integrity check at the end of a game, where it can actually fail. Game 1 keeps the path it was given, so GameDesign §5's documented invocation is unchanged; later games get -2, -3 and nothing is overwritten. Legacy 5-column logs stay readable and are reported as legacy, never as orphans — an unsubstantiated orphan claim is the defect being fixed. All three fixes mutation-proven, including at the call site via a real two-game session. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
2026-08-07 17:41:11 +02:00
mark = {"orphan": "orphan", "pending": " .. ",
"legacy": "legacy", "ok": " "}[state]
where = f"g{r.get('game', '?')}+{r.get('after', '?')}"
print(f" {mark} {where:<7} r{r['round']:<2} {r['step']:<14} "
f"{r['comment'][:60]}")
CB-WP-0027 T01-T04: the commentary track The meta view beside the table, and a note channel that provably cannot carry a move. T01 (ADR-0014). ADR-0007 D5 is SCOPED, NOT AMENDED, and the reason it was easy is that PointerFact::parse already refuses any unrecognised field -- a comment could not reach the command path even by accident. So /command carries pointer facts, /note carries text, and Note has no code path to GroundCommand. Comments live in trials/<date>-<slug>.md, not in ScenarioFile: a scenario is executed, replayed and hashed, and prose in it is data the runner must ignore, which is how a format rots. The state hash binds; round and step are for reading. And the retention question, decided before any comment was written: RAW NOTES NEVER LEAVE clay-borg. A note reaches ground-game only by being promoted to a register finding, by a human, with the wording chosen then -- "the DARVO sequence is infuriating" is useful signal and a bad way to open a message to the game's designer. T02. CSS grid, minmax(0,1fr) on both tracks -- load-bearing, because a grid child defaults to min-content width and without it the SVG table refuses to shrink and pushes the meta column off-screen, looking correct on the developer's monitor and broken everywhere else. Single-column fallback under 64rem. The running tally moved into the panel so it is visible WHILE PLAYING; it only appeared on the ending page before, and a score you see once the game is over informs nothing. T03. A plain <form method="post">, so the box works with the script disabled; the command channel needs JavaScript because a drag is not a form submission, a comment is one. 303 See Other so a reload does not re-post. esc()'s first hostile input: <script>alert(1)</script> renders escaped AND STILL READABLE -- escaping that eats the player's words is its own defect. Verified over real HTTP: note posted 303, hostile note stored as text, empty note refused 400, game did not advance. T04. tools/trials.py and make trials. THE REPORT'S DESIGN CHANGED BECAUSE I RAN IT: the first version called any note without a recording an orphan, so a live session reported every note as broken -- the recording is only written at game end. A metric that cries wolf is one nobody reads, which is the exact failure this pass exists to prevent. Now ok / pending / orphan, and only orphan is a target-0 number. The self-test exercises the REPORTING path, not just the parser, because design-baseline.py had a green self-test and an unexercised reporting path and that is where it rotted. And a latent Makefile defect surfaced: make trials did nothing, because trials is also a directory and Make saw an up-to-date file. design, difficulty and trials -- added by CB-WP-0022, CB-WP-0025 and this pass -- were ALL missing from .PHONY; only the one that collided revealed it. make all: exit 0. 49 render tests, 26 cb-play, loop-lint clean. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
2026-08-06 10:37:46 +02:00
print()
print(f" trial logs {len(found)}")
print(f" notes {total}")
print(f" positions unreachable {orphans} target 0"
CB-WP-0033: a game is the unit make trials reported "positions unreachable: 4, target 0 — the recording exists but the position moved". All four were false. A recording holds one hash, the final state, and reachability asked whether the note's hash was in that file — so a mid-game note could never match, and a post-game note from any but the last game could not either. Instance 8 of the ADR-0018 family: vary only WHEN a note was written and the answer flips, with nothing having moved. The root cause was not the metric. play again reused state belonging to a game: it overwrote the previous game's recording (data loss), never cleared the journal (game 2's log opened with game 1's commands), and so a note's command index pointed into a recording without those commands. Fixing reachability alone would have gone green while a session still destroyed its own evidence. A note now binds by (game, after) — an index into the recording's own commands list, which a reader can replay to. The hash keeps a job as the integrity check at the end of a game, where it can actually fail. Game 1 keeps the path it was given, so GameDesign §5's documented invocation is unchanged; later games get -2, -3 and nothing is overwritten. Legacy 5-column logs stay readable and are reported as legacy, never as orphans — an unsubstantiated orphan claim is the defect being fixed. All three fixes mutation-proven, including at the call site via a real two-game session. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
2026-08-07 17:41:11 +02:00
+ (" <-- a note points past the end of its own game" if orphans else ""))
CB-WP-0027 T01-T04: the commentary track The meta view beside the table, and a note channel that provably cannot carry a move. T01 (ADR-0014). ADR-0007 D5 is SCOPED, NOT AMENDED, and the reason it was easy is that PointerFact::parse already refuses any unrecognised field -- a comment could not reach the command path even by accident. So /command carries pointer facts, /note carries text, and Note has no code path to GroundCommand. Comments live in trials/<date>-<slug>.md, not in ScenarioFile: a scenario is executed, replayed and hashed, and prose in it is data the runner must ignore, which is how a format rots. The state hash binds; round and step are for reading. And the retention question, decided before any comment was written: RAW NOTES NEVER LEAVE clay-borg. A note reaches ground-game only by being promoted to a register finding, by a human, with the wording chosen then -- "the DARVO sequence is infuriating" is useful signal and a bad way to open a message to the game's designer. T02. CSS grid, minmax(0,1fr) on both tracks -- load-bearing, because a grid child defaults to min-content width and without it the SVG table refuses to shrink and pushes the meta column off-screen, looking correct on the developer's monitor and broken everywhere else. Single-column fallback under 64rem. The running tally moved into the panel so it is visible WHILE PLAYING; it only appeared on the ending page before, and a score you see once the game is over informs nothing. T03. A plain <form method="post">, so the box works with the script disabled; the command channel needs JavaScript because a drag is not a form submission, a comment is one. 303 See Other so a reload does not re-post. esc()'s first hostile input: <script>alert(1)</script> renders escaped AND STILL READABLE -- escaping that eats the player's words is its own defect. Verified over real HTTP: note posted 303, hostile note stored as text, empty note refused 400, game did not advance. T04. tools/trials.py and make trials. THE REPORT'S DESIGN CHANGED BECAUSE I RAN IT: the first version called any note without a recording an orphan, so a live session reported every note as broken -- the recording is only written at game end. A metric that cries wolf is one nobody reads, which is the exact failure this pass exists to prevent. Now ok / pending / orphan, and only orphan is a target-0 number. The self-test exercises the REPORTING path, not just the parser, because design-baseline.py had a green self-test and an unexercised reporting path and that is where it rotted. And a latent Makefile defect surfaced: make trials did nothing, because trials is also a directory and Make saw an up-to-date file. design, difficulty and trials -- added by CB-WP-0022, CB-WP-0025 and this pass -- were ALL missing from .PHONY; only the one that collided revealed it. make all: exit 0. 49 render tests, 26 cb-play, loop-lint clean. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
2026-08-06 10:37:46 +02:00
if pending:
print(f" awaiting a recording {pending}"
" (normal during a live session; the recording lands at game end)")
CB-WP-0033: a game is the unit make trials reported "positions unreachable: 4, target 0 — the recording exists but the position moved". All four were false. A recording holds one hash, the final state, and reachability asked whether the note's hash was in that file — so a mid-game note could never match, and a post-game note from any but the last game could not either. Instance 8 of the ADR-0018 family: vary only WHEN a note was written and the answer flips, with nothing having moved. The root cause was not the metric. play again reused state belonging to a game: it overwrote the previous game's recording (data loss), never cleared the journal (game 2's log opened with game 1's commands), and so a note's command index pointed into a recording without those commands. Fixing reachability alone would have gone green while a session still destroyed its own evidence. A note now binds by (game, after) — an index into the recording's own commands list, which a reader can replay to. The hash keeps a job as the integrity check at the end of a game, where it can actually fail. Game 1 keeps the path it was given, so GameDesign §5's documented invocation is unchanged; later games get -2, -3 and nothing is overwritten. Legacy 5-column logs stay readable and are reported as legacy, never as orphans — an unsubstantiated orphan claim is the defect being fixed. All three fixes mutation-proven, including at the call site via a real two-game session. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
2026-08-07 17:41:11 +02:00
if legacy:
print(f" legacy binding {legacy}"
" (written before ADR-0019; position not checkable, NOT orphaned)")
CB-WP-0027 T01-T04: the commentary track The meta view beside the table, and a note channel that provably cannot carry a move. T01 (ADR-0014). ADR-0007 D5 is SCOPED, NOT AMENDED, and the reason it was easy is that PointerFact::parse already refuses any unrecognised field -- a comment could not reach the command path even by accident. So /command carries pointer facts, /note carries text, and Note has no code path to GroundCommand. Comments live in trials/<date>-<slug>.md, not in ScenarioFile: a scenario is executed, replayed and hashed, and prose in it is data the runner must ignore, which is how a format rots. The state hash binds; round and step are for reading. And the retention question, decided before any comment was written: RAW NOTES NEVER LEAVE clay-borg. A note reaches ground-game only by being promoted to a register finding, by a human, with the wording chosen then -- "the DARVO sequence is infuriating" is useful signal and a bad way to open a message to the game's designer. T02. CSS grid, minmax(0,1fr) on both tracks -- load-bearing, because a grid child defaults to min-content width and without it the SVG table refuses to shrink and pushes the meta column off-screen, looking correct on the developer's monitor and broken everywhere else. Single-column fallback under 64rem. The running tally moved into the panel so it is visible WHILE PLAYING; it only appeared on the ending page before, and a score you see once the game is over informs nothing. T03. A plain <form method="post">, so the box works with the script disabled; the command channel needs JavaScript because a drag is not a form submission, a comment is one. 303 See Other so a reload does not re-post. esc()'s first hostile input: <script>alert(1)</script> renders escaped AND STILL READABLE -- escaping that eats the player's words is its own defect. Verified over real HTTP: note posted 303, hostile note stored as text, empty note refused 400, game did not advance. T04. tools/trials.py and make trials. THE REPORT'S DESIGN CHANGED BECAUSE I RAN IT: the first version called any note without a recording an orphan, so a live session reported every note as broken -- the recording is only written at game end. A metric that cries wolf is one nobody reads, which is the exact failure this pass exists to prevent. Now ok / pending / orphan, and only orphan is a target-0 number. The self-test exercises the REPORTING path, not just the parser, because design-baseline.py had a green self-test and an unexercised reporting path and that is where it rotted. And a latent Makefile defect surfaced: make trials did nothing, because trials is also a directory and Make saw an up-to-date file. design, difficulty and trials -- added by CB-WP-0022, CB-WP-0025 and this pass -- were ALL missing from .PHONY; only the one that collided revealed it. make all: exit 0. 49 render tests, 26 cb-play, loop-lint clean. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
2026-08-06 10:37:46 +02:00
print(f" notes past {NOTE_EXPIRY_DAYS} days {expired} target 0")
CB-WP-0033: a game is the unit make trials reported "positions unreachable: 4, target 0 — the recording exists but the position moved". All four were false. A recording holds one hash, the final state, and reachability asked whether the note's hash was in that file — so a mid-game note could never match, and a post-game note from any but the last game could not either. Instance 8 of the ADR-0018 family: vary only WHEN a note was written and the answer flips, with nothing having moved. The root cause was not the metric. play again reused state belonging to a game: it overwrote the previous game's recording (data loss), never cleared the journal (game 2's log opened with game 1's commands), and so a note's command index pointed into a recording without those commands. Fixing reachability alone would have gone green while a session still destroyed its own evidence. A note now binds by (game, after) — an index into the recording's own commands list, which a reader can replay to. The hash keeps a job as the integrity check at the end of a game, where it can actually fail. Game 1 keeps the path it was given, so GameDesign §5's documented invocation is unchanged; later games get -2, -3 and nothing is overwritten. Legacy 5-column logs stay readable and are reported as legacy, never as orphans — an unsubstantiated orphan claim is the defect being fixed. All three fixes mutation-proven, including at the call site via a real two-game session. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
2026-08-07 17:41:11 +02:00
if total and not orphans and not legacy:
CB-WP-0027 T01-T04: the commentary track The meta view beside the table, and a note channel that provably cannot carry a move. T01 (ADR-0014). ADR-0007 D5 is SCOPED, NOT AMENDED, and the reason it was easy is that PointerFact::parse already refuses any unrecognised field -- a comment could not reach the command path even by accident. So /command carries pointer facts, /note carries text, and Note has no code path to GroundCommand. Comments live in trials/<date>-<slug>.md, not in ScenarioFile: a scenario is executed, replayed and hashed, and prose in it is data the runner must ignore, which is how a format rots. The state hash binds; round and step are for reading. And the retention question, decided before any comment was written: RAW NOTES NEVER LEAVE clay-borg. A note reaches ground-game only by being promoted to a register finding, by a human, with the wording chosen then -- "the DARVO sequence is infuriating" is useful signal and a bad way to open a message to the game's designer. T02. CSS grid, minmax(0,1fr) on both tracks -- load-bearing, because a grid child defaults to min-content width and without it the SVG table refuses to shrink and pushes the meta column off-screen, looking correct on the developer's monitor and broken everywhere else. Single-column fallback under 64rem. The running tally moved into the panel so it is visible WHILE PLAYING; it only appeared on the ending page before, and a score you see once the game is over informs nothing. T03. A plain <form method="post">, so the box works with the script disabled; the command channel needs JavaScript because a drag is not a form submission, a comment is one. 303 See Other so a reload does not re-post. esc()'s first hostile input: <script>alert(1)</script> renders escaped AND STILL READABLE -- escaping that eats the player's words is its own defect. Verified over real HTTP: note posted 303, hostile note stored as text, empty note refused 400, game did not advance. T04. tools/trials.py and make trials. THE REPORT'S DESIGN CHANGED BECAUSE I RAN IT: the first version called any note without a recording an orphan, so a live session reported every note as broken -- the recording is only written at game end. A metric that cries wolf is one nobody reads, which is the exact failure this pass exists to prevent. Now ok / pending / orphan, and only orphan is a target-0 number. The self-test exercises the REPORTING path, not just the parser, because design-baseline.py had a green self-test and an unexercised reporting path and that is where it rotted. And a latent Makefile defect surfaced: make trials did nothing, because trials is also a directory and Make saw an up-to-date file. design, difficulty and trials -- added by CB-WP-0022, CB-WP-0025 and this pass -- were ALL missing from .PHONY; only the one that collided revealed it. make all: exit 0. 49 render tests, 26 cb-play, loop-lint clean. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
2026-08-06 10:37:46 +02:00
print("\n Every note points at a position a reader can reach.")
print("\n Raw notes stay in this repo (ADR-0014 D4). A note reaches"
"\n ground-game only by being promoted to a register finding, by a"
"\n human, with the wording chosen then.")
return 0
def self_test():
"""Positive controls, including for the REPORTING path.
`design-baseline.py` had a green self-test and an unexercised
reporting path, and the reporting path is where it rotted (ADR-0012
D8). So this runs `report` against a fixture and checks what it says.
"""
import tempfile, io, contextlib
ok = True
def check(name, cond, detail=""):
nonlocal ok
ok = ok and bool(cond)
print(f" [{'ok ' if cond else 'FAIL'}] {name}" + (f"{detail}" if detail else ""))
CB-WP-0033: a game is the unit make trials reported "positions unreachable: 4, target 0 — the recording exists but the position moved". All four were false. A recording holds one hash, the final state, and reachability asked whether the note's hash was in that file — so a mid-game note could never match, and a post-game note from any but the last game could not either. Instance 8 of the ADR-0018 family: vary only WHEN a note was written and the answer flips, with nothing having moved. The root cause was not the metric. play again reused state belonging to a game: it overwrote the previous game's recording (data loss), never cleared the journal (game 2's log opened with game 1's commands), and so a note's command index pointed into a recording without those commands. Fixing reachability alone would have gone green while a session still destroyed its own evidence. A note now binds by (game, after) — an index into the recording's own commands list, which a reader can replay to. The hash keeps a job as the integrity check at the end of a game, where it can actually fail. Game 1 keeps the path it was given, so GameDesign §5's documented invocation is unchanged; later games get -2, -3 and nothing is overwritten. Legacy 5-column logs stay readable and are reported as legacy, never as orphans — an unsubstantiated orphan claim is the defect being fixed. All three fixes mutation-proven, including at the call site via a real two-game session. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
2026-08-07 17:41:11 +02:00
def log_of(*rows):
head = ("| n | game | after | round | step | state_hash | comment |\n"
"|---|---|---|---|---|---|---|\n")
return f"# Trial log\n\n{BEGIN}\n\n{head}" + "".join(rows) + f"\n{END}\n"
good = log_of("| 1 | 1 | 4 | 3 | Select | abc123def456 | why is SOLVE doing nothing |\n")
legacy = (f"# Trial log\n\n{BEGIN}\n\n"
"| n | round | step | state_hash | comment |\n|---|---|---|---|---|\n"
"| 1 | 3 | Select | abc123def456 | why is SOLVE doing nothing |\n"
f"\n{END}\n")
def recording(commands, final_hash):
"""A recording with `commands` commands and one final hash — the
shape ADR-0019 D3 reads: a list to index into, plus the one
position the file states outright."""
body = "commands:\n"
for _ in range(commands):
body += "- actor: P1\n cmd: select_action\n args: {}\n"
return body + f"expect:\n state_hash: {final_hash}\n"
CB-WP-0027 T01-T04: the commentary track The meta view beside the table, and a note channel that provably cannot carry a move. T01 (ADR-0014). ADR-0007 D5 is SCOPED, NOT AMENDED, and the reason it was easy is that PointerFact::parse already refuses any unrecognised field -- a comment could not reach the command path even by accident. So /command carries pointer facts, /note carries text, and Note has no code path to GroundCommand. Comments live in trials/<date>-<slug>.md, not in ScenarioFile: a scenario is executed, replayed and hashed, and prose in it is data the runner must ignore, which is how a format rots. The state hash binds; round and step are for reading. And the retention question, decided before any comment was written: RAW NOTES NEVER LEAVE clay-borg. A note reaches ground-game only by being promoted to a register finding, by a human, with the wording chosen then -- "the DARVO sequence is infuriating" is useful signal and a bad way to open a message to the game's designer. T02. CSS grid, minmax(0,1fr) on both tracks -- load-bearing, because a grid child defaults to min-content width and without it the SVG table refuses to shrink and pushes the meta column off-screen, looking correct on the developer's monitor and broken everywhere else. Single-column fallback under 64rem. The running tally moved into the panel so it is visible WHILE PLAYING; it only appeared on the ending page before, and a score you see once the game is over informs nothing. T03. A plain <form method="post">, so the box works with the script disabled; the command channel needs JavaScript because a drag is not a form submission, a comment is one. 303 See Other so a reload does not re-post. esc()'s first hostile input: <script>alert(1)</script> renders escaped AND STILL READABLE -- escaping that eats the player's words is its own defect. Verified over real HTTP: note posted 303, hostile note stored as text, empty note refused 400, game did not advance. T04. tools/trials.py and make trials. THE REPORT'S DESIGN CHANGED BECAUSE I RAN IT: the first version called any note without a recording an orphan, so a live session reported every note as broken -- the recording is only written at game end. A metric that cries wolf is one nobody reads, which is the exact failure this pass exists to prevent. Now ok / pending / orphan, and only orphan is a target-0 number. The self-test exercises the REPORTING path, not just the parser, because design-baseline.py had a green self-test and an unexercised reporting path and that is where it rotted. And a latent Makefile defect surfaced: make trials did nothing, because trials is also a directory and Make saw an up-to-date file. design, difficulty and trials -- added by CB-WP-0022, CB-WP-0025 and this pass -- were ALL missing from .PHONY; only the one that collided revealed it. make all: exit 0. 49 render tests, 26 cb-play, loop-lint clean. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
2026-08-06 10:37:46 +02:00
check("a trial log parses", len(parse(good)) == 1)
CB-WP-0033: a game is the unit make trials reported "positions unreachable: 4, target 0 — the recording exists but the position moved". All four were false. A recording holds one hash, the final state, and reachability asked whether the note's hash was in that file — so a mid-game note could never match, and a post-game note from any but the last game could not either. Instance 8 of the ADR-0018 family: vary only WHEN a note was written and the answer flips, with nothing having moved. The root cause was not the metric. play again reused state belonging to a game: it overwrote the previous game's recording (data loss), never cleared the journal (game 2's log opened with game 1's commands), and so a note's command index pointed into a recording without those commands. Fixing reachability alone would have gone green while a session still destroyed its own evidence. A note now binds by (game, after) — an index into the recording's own commands list, which a reader can replay to. The hash keeps a job as the integrity check at the end of a game, where it can actually fail. Game 1 keeps the path it was given, so GameDesign §5's documented invocation is unchanged; later games get -2, -3 and nothing is overwritten. Legacy 5-column logs stay readable and are reported as legacy, never as orphans — an unsubstantiated orphan claim is the defect being fixed. All three fixes mutation-proven, including at the call site via a real two-game session. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
2026-08-07 17:41:11 +02:00
check("a legacy 5-column log still parses", len(parse(legacy)) == 1,
"ADR-0019 D4 \u2014 old logs stay readable")
check("a legacy row is MARKED legacy", parse(legacy)[0].get("legacy") is True)
CB-WP-0027 T01-T04: the commentary track The meta view beside the table, and a note channel that provably cannot carry a move. T01 (ADR-0014). ADR-0007 D5 is SCOPED, NOT AMENDED, and the reason it was easy is that PointerFact::parse already refuses any unrecognised field -- a comment could not reach the command path even by accident. So /command carries pointer facts, /note carries text, and Note has no code path to GroundCommand. Comments live in trials/<date>-<slug>.md, not in ScenarioFile: a scenario is executed, replayed and hashed, and prose in it is data the runner must ignore, which is how a format rots. The state hash binds; round and step are for reading. And the retention question, decided before any comment was written: RAW NOTES NEVER LEAVE clay-borg. A note reaches ground-game only by being promoted to a register finding, by a human, with the wording chosen then -- "the DARVO sequence is infuriating" is useful signal and a bad way to open a message to the game's designer. T02. CSS grid, minmax(0,1fr) on both tracks -- load-bearing, because a grid child defaults to min-content width and without it the SVG table refuses to shrink and pushes the meta column off-screen, looking correct on the developer's monitor and broken everywhere else. Single-column fallback under 64rem. The running tally moved into the panel so it is visible WHILE PLAYING; it only appeared on the ending page before, and a score you see once the game is over informs nothing. T03. A plain <form method="post">, so the box works with the script disabled; the command channel needs JavaScript because a drag is not a form submission, a comment is one. 303 See Other so a reload does not re-post. esc()'s first hostile input: <script>alert(1)</script> renders escaped AND STILL READABLE -- escaping that eats the player's words is its own defect. Verified over real HTTP: note posted 303, hostile note stored as text, empty note refused 400, game did not advance. T04. tools/trials.py and make trials. THE REPORT'S DESIGN CHANGED BECAUSE I RAN IT: the first version called any note without a recording an orphan, so a live session reported every note as broken -- the recording is only written at game end. A metric that cries wolf is one nobody reads, which is the exact failure this pass exists to prevent. Now ok / pending / orphan, and only orphan is a target-0 number. The self-test exercises the REPORTING path, not just the parser, because design-baseline.py had a green self-test and an unexercised reporting path and that is where it rotted. And a latent Makefile defect surfaced: make trials did nothing, because trials is also a directory and Make saw an up-to-date file. design, difficulty and trials -- added by CB-WP-0022, CB-WP-0025 and this pass -- were ALL missing from .PHONY; only the one that collided revealed it. make all: exit 0. 49 render tests, 26 cb-play, loop-lint clean. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
2026-08-06 10:37:46 +02:00
check("the header row is not a note", all(r["n"] != "n" for r in parse(good)))
# CB-WP-0032. A row of the wrong shape must be LOUD. This used to
# `continue`, so a format change would have reported zero notes for
# every log at once -- the failure the parser's own docstring names.
CB-WP-0033: a game is the unit make trials reported "positions unreachable: 4, target 0 — the recording exists but the position moved". All four were false. A recording holds one hash, the final state, and reachability asked whether the note's hash was in that file — so a mid-game note could never match, and a post-game note from any but the last game could not either. Instance 8 of the ADR-0018 family: vary only WHEN a note was written and the answer flips, with nothing having moved. The root cause was not the metric. play again reused state belonging to a game: it overwrote the previous game's recording (data loss), never cleared the journal (game 2's log opened with game 1's commands), and so a note's command index pointed into a recording without those commands. Fixing reachability alone would have gone green while a session still destroyed its own evidence. A note now binds by (game, after) — an index into the recording's own commands list, which a reader can replay to. The hash keeps a job as the integrity check at the end of a game, where it can actually fail. Game 1 keeps the path it was given, so GameDesign §5's documented invocation is unchanged; later games get -2, -3 and nothing is overwritten. Legacy 5-column logs stay readable and are reported as legacy, never as orphans — an unsubstantiated orphan claim is the defect being fixed. All three fixes mutation-proven, including at the call site via a real two-game session. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
2026-08-07 17:41:11 +02:00
# Six columns is neither the current shape (7) nor the legacy one (5),
# so it must RAISE rather than be skipped.
six = log_of("| 1 | 1 | 4 | 3 | Select | abc123def456 |\n")
try:
parse(six)
check("a row with the wrong column count is an error", False)
except ValueError as e:
check("a row with the wrong column count is an error", "6 columns" in str(e),
"a silent skip would empty every log at once")
# And the surrounding walk must report the REASON it failed.
with tempfile.TemporaryDirectory() as d:
with open(os.path.join(d, "2026-08-07-bad.md"), "w") as fh:
fh.write(six)
buf = io.StringIO()
with contextlib.redirect_stdout(buf):
logs(d)
check("the walk names the real reason", "6 columns" in buf.getvalue(),
"it used to blame a missing block for every failure")
CB-WP-0027 T01-T04: the commentary track The meta view beside the table, and a note channel that provably cannot carry a move. T01 (ADR-0014). ADR-0007 D5 is SCOPED, NOT AMENDED, and the reason it was easy is that PointerFact::parse already refuses any unrecognised field -- a comment could not reach the command path even by accident. So /command carries pointer facts, /note carries text, and Note has no code path to GroundCommand. Comments live in trials/<date>-<slug>.md, not in ScenarioFile: a scenario is executed, replayed and hashed, and prose in it is data the runner must ignore, which is how a format rots. The state hash binds; round and step are for reading. And the retention question, decided before any comment was written: RAW NOTES NEVER LEAVE clay-borg. A note reaches ground-game only by being promoted to a register finding, by a human, with the wording chosen then -- "the DARVO sequence is infuriating" is useful signal and a bad way to open a message to the game's designer. T02. CSS grid, minmax(0,1fr) on both tracks -- load-bearing, because a grid child defaults to min-content width and without it the SVG table refuses to shrink and pushes the meta column off-screen, looking correct on the developer's monitor and broken everywhere else. Single-column fallback under 64rem. The running tally moved into the panel so it is visible WHILE PLAYING; it only appeared on the ending page before, and a score you see once the game is over informs nothing. T03. A plain <form method="post">, so the box works with the script disabled; the command channel needs JavaScript because a drag is not a form submission, a comment is one. 303 See Other so a reload does not re-post. esc()'s first hostile input: <script>alert(1)</script> renders escaped AND STILL READABLE -- escaping that eats the player's words is its own defect. Verified over real HTTP: note posted 303, hostile note stored as text, empty note refused 400, game did not advance. T04. tools/trials.py and make trials. THE REPORT'S DESIGN CHANGED BECAUSE I RAN IT: the first version called any note without a recording an orphan, so a live session reported every note as broken -- the recording is only written at game end. A metric that cries wolf is one nobody reads, which is the exact failure this pass exists to prevent. Now ok / pending / orphan, and only orphan is a target-0 number. The self-test exercises the REPORTING path, not just the parser, because design-baseline.py had a green self-test and an unexercised reporting path and that is where it rotted. And a latent Makefile defect surfaced: make trials did nothing, because trials is also a directory and Make saw an up-to-date file. design, difficulty and trials -- added by CB-WP-0022, CB-WP-0025 and this pass -- were ALL missing from .PHONY; only the one that collided revealed it. make all: exit 0. 49 render tests, 26 cb-play, loop-lint clean. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
2026-08-06 10:37:46 +02:00
try:
parse("# Trial log\n\nno block here\n")
check("a file with no block is an error, not an empty log", False)
except ValueError:
check("a file with no block is an error, not an empty log", True,
"silently reporting zero is the failure this tool prevents")
with tempfile.TemporaryDirectory() as d:
open(os.path.join(d, "2026-08-06-x.md"), "w").write(good)
CB-WP-0033: a game is the unit make trials reported "positions unreachable: 4, target 0 — the recording exists but the position moved". All four were false. A recording holds one hash, the final state, and reachability asked whether the note's hash was in that file — so a mid-game note could never match, and a post-game note from any but the last game could not either. Instance 8 of the ADR-0018 family: vary only WHEN a note was written and the answer flips, with nothing having moved. The root cause was not the metric. play again reused state belonging to a game: it overwrote the previous game's recording (data loss), never cleared the journal (game 2's log opened with game 1's commands), and so a note's command index pointed into a recording without those commands. Fixing reachability alone would have gone green while a session still destroyed its own evidence. A note now binds by (game, after) — an index into the recording's own commands list, which a reader can replay to. The hash keeps a job as the integrity check at the end of a game, where it can actually fail. Game 1 keeps the path it was given, so GameDesign §5's documented invocation is unchanged; later games get -2, -3 and nothing is overwritten. Legacy 5-column logs stay readable and are reported as legacy, never as orphans — an unsubstantiated orphan claim is the defect being fixed. All three fixes mutation-proven, including at the call site via a real two-game session. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
2026-08-07 17:41:11 +02:00
row = parse(good)[0] # game 1, after 4, hash abc123def456
CB-WP-0027 T01-T04: the commentary track The meta view beside the table, and a note channel that provably cannot carry a move. T01 (ADR-0014). ADR-0007 D5 is SCOPED, NOT AMENDED, and the reason it was easy is that PointerFact::parse already refuses any unrecognised field -- a comment could not reach the command path even by accident. So /command carries pointer facts, /note carries text, and Note has no code path to GroundCommand. Comments live in trials/<date>-<slug>.md, not in ScenarioFile: a scenario is executed, replayed and hashed, and prose in it is data the runner must ignore, which is how a format rots. The state hash binds; round and step are for reading. And the retention question, decided before any comment was written: RAW NOTES NEVER LEAVE clay-borg. A note reaches ground-game only by being promoted to a register finding, by a human, with the wording chosen then -- "the DARVO sequence is infuriating" is useful signal and a bad way to open a message to the game's designer. T02. CSS grid, minmax(0,1fr) on both tracks -- load-bearing, because a grid child defaults to min-content width and without it the SVG table refuses to shrink and pushes the meta column off-screen, looking correct on the developer's monitor and broken everywhere else. Single-column fallback under 64rem. The running tally moved into the panel so it is visible WHILE PLAYING; it only appeared on the ending page before, and a score you see once the game is over informs nothing. T03. A plain <form method="post">, so the box works with the script disabled; the command channel needs JavaScript because a drag is not a form submission, a comment is one. 303 See Other so a reload does not re-post. esc()'s first hostile input: <script>alert(1)</script> renders escaped AND STILL READABLE -- escaping that eats the player's words is its own defect. Verified over real HTTP: note posted 303, hostile note stored as text, empty note refused 400, game did not advance. T04. tools/trials.py and make trials. THE REPORT'S DESIGN CHANGED BECAUSE I RAN IT: the first version called any note without a recording an orphan, so a live session reported every note as broken -- the recording is only written at game end. A metric that cries wolf is one nobody reads, which is the exact failure this pass exists to prevent. Now ok / pending / orphan, and only orphan is a target-0 number. The self-test exercises the REPORTING path, not just the parser, because design-baseline.py had a green self-test and an unexercised reporting path and that is where it rotted. And a latent Makefile defect surfaced: make trials did nothing, because trials is also a directory and Make saw an up-to-date file. design, difficulty and trials -- added by CB-WP-0022, CB-WP-0025 and this pass -- were ALL missing from .PHONY; only the one that collided revealed it. make all: exit 0. 49 render tests, 26 cb-play, loop-lint clean. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
2026-08-06 10:37:46 +02:00
check("a note with no recording is PENDING, not orphaned",
CB-WP-0033: a game is the unit make trials reported "positions unreachable: 4, target 0 — the recording exists but the position moved". All four were false. A recording holds one hash, the final state, and reachability asked whether the note's hash was in that file — so a mid-game note could never match, and a post-game note from any but the last game could not either. Instance 8 of the ADR-0018 family: vary only WHEN a note was written and the answer flips, with nothing having moved. The root cause was not the metric. play again reused state belonging to a game: it overwrote the previous game's recording (data loss), never cleared the journal (game 2's log opened with game 1's commands), and so a note's command index pointed into a recording without those commands. Fixing reachability alone would have gone green while a session still destroyed its own evidence. A note now binds by (game, after) — an index into the recording's own commands list, which a reader can replay to. The hash keeps a job as the integrity check at the end of a game, where it can actually fail. Game 1 keeps the path it was given, so GameDesign §5's documented invocation is unchanged; later games get -2, -3 and nothing is overwritten. Legacy 5-column logs stay readable and are reported as legacy, never as orphans — an unsubstantiated orphan claim is the defect being fixed. All three fixes mutation-proven, including at the call site via a real two-game session. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
2026-08-07 17:41:11 +02:00
reachability(row, None) == "pending",
CB-WP-0027 T01-T04: the commentary track The meta view beside the table, and a note channel that provably cannot carry a move. T01 (ADR-0014). ADR-0007 D5 is SCOPED, NOT AMENDED, and the reason it was easy is that PointerFact::parse already refuses any unrecognised field -- a comment could not reach the command path even by accident. So /command carries pointer facts, /note carries text, and Note has no code path to GroundCommand. Comments live in trials/<date>-<slug>.md, not in ScenarioFile: a scenario is executed, replayed and hashed, and prose in it is data the runner must ignore, which is how a format rots. The state hash binds; round and step are for reading. And the retention question, decided before any comment was written: RAW NOTES NEVER LEAVE clay-borg. A note reaches ground-game only by being promoted to a register finding, by a human, with the wording chosen then -- "the DARVO sequence is infuriating" is useful signal and a bad way to open a message to the game's designer. T02. CSS grid, minmax(0,1fr) on both tracks -- load-bearing, because a grid child defaults to min-content width and without it the SVG table refuses to shrink and pushes the meta column off-screen, looking correct on the developer's monitor and broken everywhere else. Single-column fallback under 64rem. The running tally moved into the panel so it is visible WHILE PLAYING; it only appeared on the ending page before, and a score you see once the game is over informs nothing. T03. A plain <form method="post">, so the box works with the script disabled; the command channel needs JavaScript because a drag is not a form submission, a comment is one. 303 See Other so a reload does not re-post. esc()'s first hostile input: <script>alert(1)</script> renders escaped AND STILL READABLE -- escaping that eats the player's words is its own defect. Verified over real HTTP: note posted 303, hostile note stored as text, empty note refused 400, game did not advance. T04. tools/trials.py and make trials. THE REPORT'S DESIGN CHANGED BECAUSE I RAN IT: the first version called any note without a recording an orphan, so a live session reported every note as broken -- the recording is only written at game end. A metric that cries wolf is one nobody reads, which is the exact failure this pass exists to prevent. Now ok / pending / orphan, and only orphan is a target-0 number. The self-test exercises the REPORTING path, not just the parser, because design-baseline.py had a green self-test and an unexercised reporting path and that is where it rotted. And a latent Makefile defect surfaced: make trials did nothing, because trials is also a directory and Make saw an up-to-date file. design, difficulty and trials -- added by CB-WP-0022, CB-WP-0025 and this pass -- were ALL missing from .PHONY; only the one that collided revealed it. make all: exit 0. 49 render tests, 26 cb-play, loop-lint clean. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
2026-08-06 10:37:46 +02:00
"a live session must not report orphans")
CB-WP-0033: a game is the unit make trials reported "positions unreachable: 4, target 0 — the recording exists but the position moved". All four were false. A recording holds one hash, the final state, and reachability asked whether the note's hash was in that file — so a mid-game note could never match, and a post-game note from any but the last game could not either. Instance 8 of the ADR-0018 family: vary only WHEN a note was written and the answer flips, with nothing having moved. The root cause was not the metric. play again reused state belonging to a game: it overwrote the previous game's recording (data loss), never cleared the journal (game 2's log opened with game 1's commands), and so a note's command index pointed into a recording without those commands. Fixing reachability alone would have gone green while a session still destroyed its own evidence. A note now binds by (game, after) — an index into the recording's own commands list, which a reader can replay to. The hash keeps a job as the integrity check at the end of a game, where it can actually fail. Game 1 keeps the path it was given, so GameDesign §5's documented invocation is unchanged; later games get -2, -3 and nothing is overwritten. Legacy 5-column logs stay readable and are reported as legacy, never as orphans — an unsubstantiated orphan claim is the defect being fixed. All three fixes mutation-proven, including at the call site via a real two-game session. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
2026-08-07 17:41:11 +02:00
CB-WP-0027 T01-T04: the commentary track The meta view beside the table, and a note channel that provably cannot carry a move. T01 (ADR-0014). ADR-0007 D5 is SCOPED, NOT AMENDED, and the reason it was easy is that PointerFact::parse already refuses any unrecognised field -- a comment could not reach the command path even by accident. So /command carries pointer facts, /note carries text, and Note has no code path to GroundCommand. Comments live in trials/<date>-<slug>.md, not in ScenarioFile: a scenario is executed, replayed and hashed, and prose in it is data the runner must ignore, which is how a format rots. The state hash binds; round and step are for reading. And the retention question, decided before any comment was written: RAW NOTES NEVER LEAVE clay-borg. A note reaches ground-game only by being promoted to a register finding, by a human, with the wording chosen then -- "the DARVO sequence is infuriating" is useful signal and a bad way to open a message to the game's designer. T02. CSS grid, minmax(0,1fr) on both tracks -- load-bearing, because a grid child defaults to min-content width and without it the SVG table refuses to shrink and pushes the meta column off-screen, looking correct on the developer's monitor and broken everywhere else. Single-column fallback under 64rem. The running tally moved into the panel so it is visible WHILE PLAYING; it only appeared on the ending page before, and a score you see once the game is over informs nothing. T03. A plain <form method="post">, so the box works with the script disabled; the command channel needs JavaScript because a drag is not a form submission, a comment is one. 303 See Other so a reload does not re-post. esc()'s first hostile input: <script>alert(1)</script> renders escaped AND STILL READABLE -- escaping that eats the player's words is its own defect. Verified over real HTTP: note posted 303, hostile note stored as text, empty note refused 400, game did not advance. T04. tools/trials.py and make trials. THE REPORT'S DESIGN CHANGED BECAUSE I RAN IT: the first version called any note without a recording an orphan, so a live session reported every note as broken -- the recording is only written at game end. A metric that cries wolf is one nobody reads, which is the exact failure this pass exists to prevent. Now ok / pending / orphan, and only orphan is a target-0 number. The self-test exercises the REPORTING path, not just the parser, because design-baseline.py had a green self-test and an unexercised reporting path and that is where it rotted. And a latent Makefile defect surfaced: make trials did nothing, because trials is also a directory and Make saw an up-to-date file. design, difficulty and trials -- added by CB-WP-0022, CB-WP-0025 and this pass -- were ALL missing from .PHONY; only the one that collided revealed it. make all: exit 0. 49 render tests, 26 cb-play, loop-lint clean. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
2026-08-06 10:37:46 +02:00
rec = os.path.join(d, "2026-08-06-x.yaml")
CB-WP-0033: a game is the unit make trials reported "positions unreachable: 4, target 0 — the recording exists but the position moved". All four were false. A recording holds one hash, the final state, and reachability asked whether the note's hash was in that file — so a mid-game note could never match, and a post-game note from any but the last game could not either. Instance 8 of the ADR-0018 family: vary only WHEN a note was written and the answer flips, with nothing having moved. The root cause was not the metric. play again reused state belonging to a game: it overwrote the previous game's recording (data loss), never cleared the journal (game 2's log opened with game 1's commands), and so a note's command index pointed into a recording without those commands. Fixing reachability alone would have gone green while a session still destroyed its own evidence. A note now binds by (game, after) — an index into the recording's own commands list, which a reader can replay to. The hash keeps a job as the integrity check at the end of a game, where it can actually fail. Game 1 keeps the path it was given, so GameDesign §5's documented invocation is unchanged; later games get -2, -3 and nothing is overwritten. Legacy 5-column logs stay readable and are reported as legacy, never as orphans — an unsubstantiated orphan claim is the defect being fixed. All three fixes mutation-proven, including at the call site via a real two-game session. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
2026-08-07 17:41:11 +02:00
# ADR-0019 D3. A MID-GAME note is reachable by replaying `after`
# commands. Under the old check this case was IMPOSSIBLE to pass:
# the recording holds one hash, so every mid-game note was an
# orphan whatever had happened.
open(rec, "w").write(recording(10, "zzz"))
check("a mid-game note inside its game is reachable",
reachability(row, rec) == "ok",
"the old check called every one of these an orphan")
# And it can still say NO: past the end of its own game.
open(rec, "w").write(recording(2, "zzz"))
check("a note past the end of its game is an orphan",
reachability(row, rec) == "orphan",
"without this the metric cannot go red")
# At the end, the hash is the integrity check — and it can fail.
end_row = parse(log_of(
"| 1 | 1 | 4 | 3 | after the end | abc123def456 | done |\n"))[0]
open(rec, "w").write(recording(4, "abc123def456"))
check("an end-of-game note agreeing with the recording is reachable",
reachability(end_row, rec) == "ok")
open(rec, "w").write(recording(4, "something-else"))
check("an end-of-game note whose hash disagrees is an orphan",
reachability(end_row, rec) == "orphan",
CB-WP-0027 T01-T04: the commentary track The meta view beside the table, and a note channel that provably cannot carry a move. T01 (ADR-0014). ADR-0007 D5 is SCOPED, NOT AMENDED, and the reason it was easy is that PointerFact::parse already refuses any unrecognised field -- a comment could not reach the command path even by accident. So /command carries pointer facts, /note carries text, and Note has no code path to GroundCommand. Comments live in trials/<date>-<slug>.md, not in ScenarioFile: a scenario is executed, replayed and hashed, and prose in it is data the runner must ignore, which is how a format rots. The state hash binds; round and step are for reading. And the retention question, decided before any comment was written: RAW NOTES NEVER LEAVE clay-borg. A note reaches ground-game only by being promoted to a register finding, by a human, with the wording chosen then -- "the DARVO sequence is infuriating" is useful signal and a bad way to open a message to the game's designer. T02. CSS grid, minmax(0,1fr) on both tracks -- load-bearing, because a grid child defaults to min-content width and without it the SVG table refuses to shrink and pushes the meta column off-screen, looking correct on the developer's monitor and broken everywhere else. Single-column fallback under 64rem. The running tally moved into the panel so it is visible WHILE PLAYING; it only appeared on the ending page before, and a score you see once the game is over informs nothing. T03. A plain <form method="post">, so the box works with the script disabled; the command channel needs JavaScript because a drag is not a form submission, a comment is one. 303 See Other so a reload does not re-post. esc()'s first hostile input: <script>alert(1)</script> renders escaped AND STILL READABLE -- escaping that eats the player's words is its own defect. Verified over real HTTP: note posted 303, hostile note stored as text, empty note refused 400, game did not advance. T04. tools/trials.py and make trials. THE REPORT'S DESIGN CHANGED BECAUSE I RAN IT: the first version called any note without a recording an orphan, so a live session reported every note as broken -- the recording is only written at game end. A metric that cries wolf is one nobody reads, which is the exact failure this pass exists to prevent. Now ok / pending / orphan, and only orphan is a target-0 number. The self-test exercises the REPORTING path, not just the parser, because design-baseline.py had a green self-test and an unexercised reporting path and that is where it rotted. And a latent Makefile defect surfaced: make trials did nothing, because trials is also a directory and Make saw an up-to-date file. design, difficulty and trials -- added by CB-WP-0022, CB-WP-0025 and this pass -- were ALL missing from .PHONY; only the one that collided revealed it. make all: exit 0. 49 render tests, 26 cb-play, loop-lint clean. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
2026-08-06 10:37:46 +02:00
"the position moved \u2014 the case the flag exists for")
CB-WP-0033: a game is the unit make trials reported "positions unreachable: 4, target 0 — the recording exists but the position moved". All four were false. A recording holds one hash, the final state, and reachability asked whether the note's hash was in that file — so a mid-game note could never match, and a post-game note from any but the last game could not either. Instance 8 of the ADR-0018 family: vary only WHEN a note was written and the answer flips, with nothing having moved. The root cause was not the metric. play again reused state belonging to a game: it overwrote the previous game's recording (data loss), never cleared the journal (game 2's log opened with game 1's commands), and so a note's command index pointed into a recording without those commands. Fixing reachability alone would have gone green while a session still destroyed its own evidence. A note now binds by (game, after) — an index into the recording's own commands list, which a reader can replay to. The hash keeps a job as the integrity check at the end of a game, where it can actually fail. Game 1 keeps the path it was given, so GameDesign §5's documented invocation is unchanged; later games get -2, -3 and nothing is overwritten. Legacy 5-column logs stay readable and are reported as legacy, never as orphans — an unsubstantiated orphan claim is the defect being fixed. All three fixes mutation-proven, including at the call site via a real two-game session. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
2026-08-07 17:41:11 +02:00
# A legacy row is never called an orphan (ADR-0019 D4).
open(rec, "w").write(recording(2, "no-match"))
check("a legacy row is LEGACY, never orphan",
reachability(parse(legacy)[0], rec) == "legacy",
"an unsubstantiated orphan claim is the defect being fixed")
# ADR-0019 D2: game 2 reads its OWN recording.
open(os.path.join(d, "g.md"), "w").write("x")
open(os.path.join(d, "g.yaml"), "w").write(recording(1, "one"))
open(os.path.join(d, "g-2.yaml"), "w").write(recording(9, "two"))
g2 = parse(log_of("| 1 | 2 | 5 | 1 | Select | two | second game |\n"))[0]
check("a note from game 2 is judged against game 2's recording",
recording_for(os.path.join(d, "g.md"), g2).endswith("g-2.yaml")
and reachability(g2, recording_for(os.path.join(d, "g.md"), g2)) == "ok",
"one recording per FILE is what let play again overwrite it")
CB-WP-0027 T01-T04: the commentary track The meta view beside the table, and a note channel that provably cannot carry a move. T01 (ADR-0014). ADR-0007 D5 is SCOPED, NOT AMENDED, and the reason it was easy is that PointerFact::parse already refuses any unrecognised field -- a comment could not reach the command path even by accident. So /command carries pointer facts, /note carries text, and Note has no code path to GroundCommand. Comments live in trials/<date>-<slug>.md, not in ScenarioFile: a scenario is executed, replayed and hashed, and prose in it is data the runner must ignore, which is how a format rots. The state hash binds; round and step are for reading. And the retention question, decided before any comment was written: RAW NOTES NEVER LEAVE clay-borg. A note reaches ground-game only by being promoted to a register finding, by a human, with the wording chosen then -- "the DARVO sequence is infuriating" is useful signal and a bad way to open a message to the game's designer. T02. CSS grid, minmax(0,1fr) on both tracks -- load-bearing, because a grid child defaults to min-content width and without it the SVG table refuses to shrink and pushes the meta column off-screen, looking correct on the developer's monitor and broken everywhere else. Single-column fallback under 64rem. The running tally moved into the panel so it is visible WHILE PLAYING; it only appeared on the ending page before, and a score you see once the game is over informs nothing. T03. A plain <form method="post">, so the box works with the script disabled; the command channel needs JavaScript because a drag is not a form submission, a comment is one. 303 See Other so a reload does not re-post. esc()'s first hostile input: <script>alert(1)</script> renders escaped AND STILL READABLE -- escaping that eats the player's words is its own defect. Verified over real HTTP: note posted 303, hostile note stored as text, empty note refused 400, game did not advance. T04. tools/trials.py and make trials. THE REPORT'S DESIGN CHANGED BECAUSE I RAN IT: the first version called any note without a recording an orphan, so a live session reported every note as broken -- the recording is only written at game end. A metric that cries wolf is one nobody reads, which is the exact failure this pass exists to prevent. Now ok / pending / orphan, and only orphan is a target-0 number. The self-test exercises the REPORTING path, not just the parser, because design-baseline.py had a green self-test and an unexercised reporting path and that is where it rotted. And a latent Makefile defect surfaced: make trials did nothing, because trials is also a directory and Make saw an up-to-date file. design, difficulty and trials -- added by CB-WP-0022, CB-WP-0025 and this pass -- were ALL missing from .PHONY; only the one that collided revealed it. make all: exit 0. 49 render tests, 26 cb-play, loop-lint clean. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
2026-08-06 10:37:46 +02:00
# THE control design-baseline.py lacked: exercise the reporting
# path and assert on what it printed.
buf = io.StringIO()
with contextlib.redirect_stdout(buf):
report(d, datetime.date(2026, 8, 6))
out = buf.getvalue()
check("the reporting path runs and names the note",
"why is SOLVE doing nothing" in out, "not just the parser")
check("the reporting path counts", "notes 1" in out)
print("trials self-test (positive control)")
return 0 if ok else 1
if __name__ == "__main__":
sys.exit(self_test() if "--self-test" in sys.argv else report())