Square patches invited camping and left the space between them empty. Rules 8 run on a
format-2 habitat of deposits:
- pools and single-nutrient bubbles at free micrometre centres, each with a stable id
and a contact radius; the cellular-v7 opening scatters bubbles reproducibly from the
genesis commit (three in ten in loose clumps) and lists pools with --pool;
- every cell drains what it touches, up to its intake rate per tick, with no target; a
pool gives at most a quarter of that rate; an emptied deposit disappears;
- a dead cell's detritus breaks into one bubble per non-empty channel;
- wildlife swims to the nearest non-empty deposit and feeds by touch.
Bundle v9 (habitat 2) and schema v43 admit it. Session v11 carries the owner's 512
nearest deposits (id, centre, radius, shape, a drained flag and 16-bit amount codes)
instead of patches; every other cellular message is v10's. The client asks for v11,
draws bubbles (nutrient-coloured, pulsing while drained, popping when gone) and pools as
puddles, slows near food and sends no feed target; the HUD summarises food in range.
Earlier rules, regions and v10 clients are unchanged.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
The region 6 playtest found a speed mutation adding 6,667 um/s (1/300 of the range,
33x a founder), so buds seemed to teleport; reproduction made buds bigger and dearer;
and wildlife never hunted. Rules 7:
- genome schema 4: a step is a tenth of the current value (at least 1); an acquired
capability starts at the founder's value instead of a minimum it could never grow from;
- development 4: a bud is always half the body, and reproduction takes value/2000 off
its construction energy (founders save a quarter; matter is never discounted);
- twins: a seeded roll (reproduction/2 per thousand) splits one affordable bud's
investment into two quarter-size buds;
- wildlife engulfs a touching cell the region step lets it take, players' included.
Founder 3, owner format 6 (no new state), bundle v8 and schema v42; the cellular-v6
opening provisions it. The client's mutation callout shrinks to a "Mutation [E]" pill and
its candidate texts describe the new steps.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
The ADR-0131 browser sign-in and grant broker requests link winhttp.dll, a Windows
system DLL, but the package's import allowlist predates them, so package-windows.ps1
refused evo-client.exe ("imports winhttp.dll, which is not an allowed system DLL").
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
DEV runs one region at a time and moves to a new one before test cells die. A life
left in the old region keeps its account's home there: the grant broker routes the
account to a region nobody serves (503 region_unavailable) and any other region
refuses its birth (existing life, one active life per world).
tools/ops/retire-region-lives.sql, run as the database owner, ends each published
life of the region at the region's latest commit (as end_cellular_lives does) and
cancels unpublished reservations. It refuses while a server holds the region's lease,
is a dry run unless confirm=yes, is idempotent, and leaves checkpoints and history
unchanged. Tested on a disposable migrated cluster.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
The playtest found mutation offers the arena could never pay for. A mutation's
matter cost is the real body difference (ADR-0118), so it cannot be discounted.
Rules 6 cut it honestly:
- genome schema 3 steps a third as far (1/300 of the range);
- development version 3 charges a third of the construction energy (13 instead of
40 pJ per 1000 fg).
Founder 2 uses both. The cellular-v5 opening creates rules-6 regions on bundle v7,
which schema v41 admits. Existing lineages, regions and committed history are
unchanged.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
An offer the cell cannot afford stayed open, and only one offer is issued at a
time, so banked feeding credit never became a new offer. A v10 choice naming
operator 0 now declines the exact outstanding offer. It is reserved and published
like any choice: schema v40 accepts operator 0 in the reserved input and trace
selections and stores result 3, declined. The owner consumes the offer and
changes nothing else, and the client reports it as committed. Old traces never
carry operator 0, so their replay is unchanged. The Evolution panel gets a
"Decline this offer" button.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
evo-grant-broker verifies realm aeternalis access tokens offline (RS256 against
the pinned JWKS; audience aevolution-grants, client aevolution-game, verified
email) and issues 10-minute session grants:
- POST /grant routes to the region holding the account's life, or the default.
- POST /logout makes every issued grant stale and expires current control.
- GET /healthz.
Schema v39 maps (issuer, subject) to stable accounts with strictly increasing
serials. The binary ships in the server image. The DEV manifests are in
deploy/kubernetes/grant-broker-dev and are not yet enabled in region/dev; that
waits for the image promotion and the out-of-band Secrets and Ingress.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
M11.7 (ADR-0128): a frozen step intent, set by the server for every owned life
whose account holds no session, leaves the cell unchanged but engulfable. Owner
format 6 / rules 5 track presence: lingering for 200 ticks, then suspended out of
the world without aging; an unfrozen intent resumes it in place or at the nearest
clear spot within three extents. Bundle v6, schema v38 and the cellular-v4
opening carry it. The server refuses an offline account's carried actions, ticks
while anyone lingers, and skips suspended cells in replication and perception.
Also fixes the TSan failure of CI run 374: the rules-4 owner test read the bud
through a reference into a replaced state.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
M11.6a (ADR-0127, schema v37): a published life is marked ended once, by a
commit of its region, through the lease-fenced end_cellular_lives. One active
life per account replaces one life per account, so after a death the account
reserves and publishes a new primitive founder while the corpse, archive and
descendants stay. The host records endings after the death's publication and
again for every founder corpse on recovery. The cellular server now refuses only
an unusable carried action, not the whole tick.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
M11.5c (ADR-0125): v10 cellular actions, the canonical input and the trace carry
target kind 3 (bud), matched to durable actions; below rules 4 the owner
ignores it. The v5 version bundle is admitted in C++ and by schema v36, which
also admits kind 3 in SQL action inputs; the provisioner offers cellular-v3.
A PostgreSQL test publishes a bud from a fed founder and recovers it after a
host restart. The publication tests now use commit ids no other suite produces.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
ADR-0121, schema v35. A checkpoint whose trace carries mutation
selections publishes with its canonical mutation block and one result
byte per choice. SQL checks the block's order, matches each record to
one reserved choice of this region and to its trace selection (tick,
life, offer, operator, successor identity), binds it to the commit with
its result, refuses a choice already bound elsewhere and checks exact
retries. The four-argument publisher becomes an internal helper behind
a wrapper that refuses traces with selections.
C++ publication takes the results from its own replay rather than a
caller, prepared commits carry block and results through retry and
resolution, and recovery rebuilds the bound choices and requires the
replay's results to match. A PostgreSQL test feeds a v4 cell to its
first offer, publishes a reserved choice, recovers it and restarts a
host. Host and session admission follow in 11.4c-2.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
ADR-0124: checkpoints pin cellular bundle v4 (rules 3, owner envelope 4,
revision archive) with the revision history profile and format-3 traces.
A format-4/rules-3 opening provisions on v4, and evo-world-clone accepts
--opening cellular-v2 for it. Migration 34 widens canonical action input
validation to target kind 2 (engulf, no coordinates), admits v4 manifests
through the existing v3 parser chain, and lets birth binding read
format-3 traces for v3 and v4. The migration runner and its checks extend
to version 34, with SQL and provisioning tests for v4.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
The dev overlay now serves region 2, so the selftest's literal
REGION_ID "1" anchor no longer exists and setup aborted. Anchor the
injected listener patch on REGION_WORLD instead, and refresh the dev
session window (3 h).
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Compile the PostgreSQL stages with PG_BUILD_JOBS=6 (was 2), including
PostgreSQL itself. postgres-check.sh gains --full (release),
--sanitizer (ASan: C++ suite plus backup/DR/upgrade process checks that
run the instrumented binaries) and --suite-only (clang-tidy build), so
SQL-only checks run once. Each run prints per-phase timings, and the
postgres job logs runner CPU, memory and disk headroom.
Also refresh the dev REGION_SESSION_WINDOW.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
postgres-check expected granting find_region_action_record to the
runtime role to fail the role audit, but migration 19 grants it and the
audit lists it as fenced, so the grant was a no-op and the audit passed.
Exercise the internal check_cellular_action_block validator instead,
mirroring the existing check_action_block case. Refresh the dev session
window.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
The upgrade rehearsal asserted the schema stayed at version 23 after a
deliberately failing migration, but migrations now reach 33, so it
reported "schema version moved" although nothing changed. Record the
version before the failing migration and compare against that. Refresh
the dev session window.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
- IntentLimits::watermark_coalesce (evo-server: every 4, coalesce 4); ADR-0092 amendment.
- m10 load-report splits poll/replicate and commit-tick/other-tick maxima.
- Budgets v2: work p95 <= 25 ms over all ticks, commit ticks <= 50 ms (ADR-0097);
moving the commit off the tick is a recorded follow-up.
- evo-load: time-based action ids, duplicate_actions, receipts before outcomes;
drill ramps at 500 ms and judges by v2.
- Refresh the dev session window.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
ADR-0097. The second evolution-dev load drill (image 7952185d) held 20
sessions with client budgets met (response p95 111 ms, outcome p95 about
1.2 s, 20 Hz, no starvation) but the server over its own: tick work p99
48-70 ms (25 ms budget), session layer 1.2-2.0 ms per session at p99
(0.5 ms budget), governor refusing and shedding to level 3 to little
effect. ReplicationServer::tick re-encoded every environment cell and
organism for each session. Sessions with the same requested interest
(no perception predicate) now share one encoding per tick; markers stay
per session. Locally 21 sessions: about 970 us -> 360 us per tick.
The drill now waits for two consecutive calm reports before the bots.
The dev overlay's REGION_RETAINED listed the deployed build itself
(61df6a1, f88bee2), which the deploy policy refuses (CI deploy failed on
those commits); it now lists only the other builds, keeping 6c96629e,
the build of the region head. Refreshes the dev session window.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>