- metrics_registry: remove toPrometheusText/appendCumulativeBuckets and the time-bucket histogram that nothing in production ever emitted (the future /metrics exporter can bring it back); keep counts/totals read by the Developer tab
- Fix +Inf bucket routing that never incremented, and its test that locked the bug in
- Instrument developer_command container (kind 6) in processCommandContainer and stats label resolution
- Read metrics/{slow_command_ms,stall_warn_ms} at the top of initServer() so stall_warn_ms=0 disables the watchdogs before pool threads start
- Shrink KindStride to 1280 (largest extension in use is 1206) with a static_assert; document scrape cost of getCardsInGamesTotal; note slow_command logging has no rate limit in servatrice.ini.example
Extend Response_GetServerStats with live counters from the in-process
MetricsRegistry: cards in games, event loop stall totals/worst,
total commands processed, average command time, active command types,
and game-start count/duration. Add a repeated CommandStats message
carrying per-command breakdowns (kind, extension number, resolved
protobuf name, count, total ms) for every type that has seen at
least one sample.
Server-side cmdGetServerStats() populates all new fields after the
existing DB uptime snapshot query, resolving protobuf extension names
via the descriptor pool for human-readable labels like
session/Command_Ping.
Expand TabDeveloper with two tables: an overview section (existing
DB stats plus the new live metrics) and a per-command breakdown table
(Command / Count / Total ms / Avg ms) sorted by total_ms descending
so the hottest commands surface first.
Took 55 minutes
Took 47 seconds
Add a lock-free MetricsRegistry that accumulates per-command processing
times in preallocated histogram slots (one per protobuf command type,
bucketed at 1/5/10/25/50/100/250/500/1000/2500/5000 ms +Inf). The
hot-path observeCommand() uses only relaxed atomic adds — no locks,
no allocations, no cache-line ping-pong beyond the unavoidable counter
updates.
Wire the registry into AbstractServerSocketInterface::processCommandContainer()
so every processed command is attributed with its container's wall-clock
time. When a container exceeds metrics/slow_command_ms (default 500),
a warning is logged including the connected username.
Add an EventLoopWatchdog heartbeat that runs on every socket pool thread.
If a heartbeat overshoots metrics/stall_warn_ms (default 2000 ms), the
overshoot is recorded in atomic counters and a warning is logged. Both
thresholds are configurable in servatrice.ini; setting stall_warn_ms to 0
disables the watchdogs entirely.
Track game-start durations via a separate histogram in MetricsRegistry.
Server_Game::startGameNow() measures the time from zone creation through
player materialization and reports it via Server::observeGameStartDurationMs().
Add a live card-count gauge: Server_Game exposes getCardsInGame() and
Servatrice::getCardsInGamesTotal() sums across all running games under
the appropriate read locks.
Include a standalone metrics_registry_test (Google Test) that validates
empty registries, single/multi-sample histograms, kind encoding,
overflow-slot collapse, negative-duration clamping, gauge rendering,
and the game-start histogram separation.
Took 10 minutes
A developer who omits log_location entirely — or sends only "chat" —
leaves chatType, gameType, roomType all false, so getMessageLogHistory
skips the target_type clause and returns every row, private messages
included. When !allowPrivateChat the server now forces game+room when
no surviving location was requested, guaranteeing the query always
carries a target_type restriction.
[Client] Demote mod+dev to moderator path in log-tab dispatch
The developer command family is strictly weaker than the moderator one
(no private chat, no sender_ip, ip filter ignored), so granting the
developer bit to an existing moderator must not silently strip their
capabilities. useDeveloperCommands is now true only when the user holds
the developer bit and not the moderator bit.
[Client] Hide the IP-address filter for developer log tab users
The developer path ignores the ip_address query field server-side.
Showing the field lets a developer type an IP and get results that are
silently unfiltered by it rather than an empty result set — reads as a
broken filter. Hide labelFindIPAddress/findIPAddress alongside the
privateChat checkbox.
The Command_GetLogHistory message was folded into Command_ViewLogHistory,
which deleted command_get_log_history.proto, but serversocketinterface
still #included its generated header. Fresh CI builds fail on the missing
file; local builds masked it by reusing a previously generated header.
Address ZeizaZach's review of the developer staff role:
- Nudge the developer log query to exclude private chat and sender IPs
(the ModeratorCommand path still sees everything).
- Deduplicate Command_GetLogHistory into Command_ViewLogHistory, which now
extends both ModeratorCommand (ext) and DeveloperCommand (dev_ext); the
client picks the DeveloperCommand-scoped extension by extendee, and the
server reads it via the extension number.
- Pull the uptime snapshot SQL into Servatrice_DatabaseInterface as
getLatestUptimeSnapshot() and widen the reported counters to 64-bit.
- Document the admin bitfield (1 admin, 2 moderator, 4 judge, 8 developer)
and add a server-side test for the developer command path.
Introduce a Developer staff level (proto flag 32, DB admin bit 8) that
sits between admin and moderator: no kick/ban/warn/report/admin powers,
but gets server log access via a new developer command container family
(GET_SERVER_STATS, VIEWLOG_HISTORY) and an idle-timeout exemption.
- Protocol: IsDeveloper flag, developer_commands.proto envelope,
Command_GetServerStats/Command_GetLogHistory, Response_GetServerStats,
Command_AdjustMod.should_be_developer
- Servatrice: fail-closed developer dispatcher, uptime snapshot handler,
shared log history handler reuse, bit-8 DB mapping
- Client: burgundy pawn/badge/labels/sort order, prepareDeveloperCommand,
minimal Developer stats tab, log tab access, promote/demote actions
Took 24 minutes
Took 18 seconds
# Commit time for manual adjustment:
# Took 15 minutes
* [Client] Send game invites from the user context menu via a private message
The user context menu gains an "Invite to Game" submenu listing the
inviteable games in the room (the inviter's own games, honoring the
buddy-only setting). Picking one opens a private message to the target
user with a cockatrice://joingame link naming the game, so the target
gets a clickable invite instead of a raw URL. Multi-game rooms offer a
picker; a single inviteable game sends directly. Sending a message to
an offline user no longer swallows the draft — it reports that the
user is offline and keeps the typed text.
Took 30 seconds
Took 1 minute
* [Client] Open the invite dialog taller by default without enforcing a minimum size
---------
Co-authored-by: Lukas Brübach <Bruebach.Lukas@bdosecurity.de>
Move the deck-to-.cod conversion prompt logic (format check, saved
preference handling, overwrite confirmation, dialog) out of
DeckPreviewWidget into dlg_convert_deck_to_cod_format so the deck
editor can reuse it without duplicating it.
Took 4 minutes
Took 4 minutes
Took 1 minute
# Commit time for manual adjustment:
# Took 3 minutes
Co-authored-by: Lukas Brübach <Bruebach.Lukas@bdosecurity.de>
* add a limit to the size of deckfiles cockatrice can load
the limit is 99999 or 100k -1 right now, which is kind of the limit of
what looks acceptable in the player
* format
* up limit to 100k because that's what the tests do
* [DeckList] Add custom deck zones to the deck tree
Introduce user-definable zones nested under a board zone (main, side
or maybeboard) so players can organize cards inside a board without
changing board semantics.
- addCustomZone, renameCustomZone, moveCustomZone and removeCustomZone
manage zones. Names are unique across the whole deck and the standard
zone names (main/side/maybeboard/tokens) stay reserved.
- Board zones are created lazily on first use.
- getZoneObjFromName resolves custom names to their nested node so
addCard and XML loading route cards into them. Unknown names keep
creating legacy top-level zones.
- deleteNode keeps empty custom zones alive and only prunes empty
board zones.
- New deck_list_zones test suite locks hash parity with flat decks,
sideboard size accounting, maybeboard exclusion from plain export
and native-format round-trips.
Took 17 minutes
Took 11 minutes
* Extract to function
Took 4 minutes
---------
Co-authored-by: Lukas Brübach <Bruebach.Lukas@bdosecurity.de>
* [VDS] Drive folder and preview widgets from the model (MVC views)
Took 16 minutes
Took 8 minutes
Took 3 minutes
Took 11 minutes
* Rebase whoopsie
Took 4 minutes
* Hide widgets instead of destroying, go back to signals, rename for consistency.
Took 13 minutes
Took 4 seconds
Took 26 minutes
Took 5 seconds
# Commit time for manual adjustment:
# Took 9 minutes
* Make VDS startup smooth: batch deck loads, guard preview resizes
- Move color identity computation into the background load task and apply
finished deck loads in bounded batches per event loop turn, so finishing
hundreds of loads at once cannot stall the UI thread
- Skip redundant resize work in DeckPreviewWidget when the banner width did
not change, and collect the clamped children once instead of searching the
widget tree on every layout pass
Took 19 minutes
# Commit time for manual adjustment:
# Took 3 minutes
* [VDS] Expose filter matches as a proxy role instead of dropping rows
The folder display scanned source-model rows and probed acceptance with
mapFromSource(...).isValid(), reaching into both models for one answer.
The proxy now keeps every row and exposes each row's search/tag/color
filter result through FilterMatchRole. The folder display and the tag
filter read everything off proxy indexes, and hidden previews keep their
sorted position in the flow layout instead of being appended at the end.
Took 11 minutes
* [VDS] Bound pending-load drain by time and make row lookups O(1)
The fixed DECK_LOADS_PER_TURN = 24 cap had no measured basis. It was guessed and existed
because every applied load emitted dataChanged into each DeckPreviewWidget,
whose handler resolved its own row with an O(n) linear scan per widget.
The model now maintains a file path -> row hash kept in sync across scans,
renames and deletions, so rowForFilePath is O(1) and the fan-out cost is
gone at its source. The drain applies finished loads until a small time
budget per event loop turn runs out, so throughput self-tunes instead of
relying on an arbitrary count.
* Actual minimal fix for resize squishing
Took 20 minutes
* Fix color widget sizing
Took 16 minutes
* [BannerWidget] Also set a max height
Took 4 minutes
---------
Co-authored-by: Lukas Brübach <Bruebach.Lukas@bdosecurity.de>
* [Client] Pin user list header length to the viewport width
The header stretch mode kept a resize section property, so after any
column grew past the viewport the list carried an invisible horizontal
pan range that scrolled rows sideways without visual feedback
Drop the leftover property so displayed length always equals viewport
width and horizontal panning is impossible
* Show columns 1 and 2
Took 12 minutes
Took 2 minutes
---------
Co-authored-by: Lukas Brübach <Bruebach.Lukas@bdosecurity.de>
* [App] Add onboarding wizard
Took 10 minutes
Took 3 minutes
Took 7 minutes
Took 9 minutes
Took 2 minutes
Took 1 minute
Took 7 minutes
* Adjust CI
Took 14 minutes
Took 56 seconds
Took 2 seconds
Took 3 seconds
* Adjust CI again
Took 14 minutes
Took 2 seconds
* Comments and fixes
Took 9 seconds
Took 1 minute
* Rebase.
Took 5 minutes
Took 50 seconds
Took 15 seconds
* Comments.
Took 7 minutes
* CI lol
Took 3 minutes
* CI again lol
Took 4 minutes
* Drop some settings, add some new ones.
Took 19 minutes
* Resize when expanding section
Took 4 minutes
Took 3 minutes
Took 7 minutes
---------
Co-authored-by: Lukas Brübach <Bruebach.Lukas@bdosecurity.de>
* [VDS] Cache mana symbol renders and skip redundant resizes
- Render each mana symbol once at a bounded master size and derive every
requested size from the cached master, avoiding repeated full-size SVG
rasterization on the GUI thread
- Share scaled results through a process-wide cache keyed by symbol and
size, so repeated widget creation and rescales don't redo the work
- Skip redundant resize work in ColorIdentityWidget and ManaSymbolWidget
when sizes did not change
Took 8 minutes
Took 50 seconds
* Move to pixmap generator
Took 8 minutes
Took 4 seconds
---------
Co-authored-by: Lukas Brübach <Bruebach.Lukas@bdosecurity.de>
* [DeckList] Make deck tree card traversal recursive
getCardNodes and forEachCard now descend into nested zones instead of
assuming a flat main/side/token layout. For today's flat trees this is
behavior-preserving; it also removes two latent crashes (unchecked
dynamic_cast dereference, null card nodes passed to forEachCard
callers). Nested zones are introduced by later custom-zones units.
Took 3 minutes
Took 7 seconds
Took 9 seconds
* Fix rebase mistake.
---------
Co-authored-by: Lukas Brübach <Bruebach.Lukas@bdosecurity.de>
* [Client] Replace playmat crop spinboxes with a direct manipulation preview
The numeric fields exposed raw parameters with no relation to what the
game draws, accepted values the renderer clamps away, and no way to see
the result before committing
The preview now renders through the exact game pipeline into a viewport
shaped like a fresh board stack plus table area, with dimmed strips
marking where a developed table crops further. Drag pans, wheel zooms,
arrow keys nudge, plus and minus zoom, Backspace or Esc restores the
crop as of focus gain and lets Esc close the dialog when unchanged.
Focus ring and accessible name and description cover keyboard and screen
reader users, new paints use palette roles so themes recolor them
Took 25 minutes
Took 3 minutes
Took 54 seconds
* Remove stale constant
Took 3 minutes
Took 26 seconds
* Rebase
Took 2 minutes
Took 3 seconds
* Add editor spinboxes again
Took 4 minutes
---------
Co-authored-by: Lukas Brübach <Bruebach.Lukas@bdosecurity.de>
Introduce DECK_ZONE_MAYBEBOARD and its visible name, and treat the
maybeboard as editor-only scratch space: plain-text export, DeckStats
and TappedOut uploads now skip cards living there. Zones of this name
are created by later custom-zones units; until then the skips are inert.
Took 20 minutes
Co-authored-by: Lukas Brübach <Bruebach.Lukas@bdosecurity.de>
The offset and zoom fields exposed raw numbers with no relation to the
strip, and the stored printing reset to the first local printing on
every open
The preview now paints through UserListPainter::drawCardArt itself, so
what you see is exactly what the user list renders. Drag pans the art
vertically at output scale, wheel zooms, arrow keys nudge, plus and
minus zoom, Backspace or Esc restores the crop as of focus gain and lets
Esc close the dialog when unchanged. Margins stay explicit spinboxes
since they trim the strip sides with no natural drag mapping. Legacy
stored zoom below the gesture floor is normalized once on open, focus
ring and accessible name and description cover keyboard and screen
reader users, all strings set in retranslateUi
Took 7 minutes
# Commit time for manual adjustment:
# Took 55 seconds
Took 41 seconds
Co-authored-by: Lukas Brübach <Bruebach.Lukas@bdosecurity.de>
cn.svg and ss.svg defined their star polygons with coordinates around
plus/minus 5e5 compensated by tiny scale transforms. Qt drops shapes whose
device space bounds exceed its rasterizer coordinate limit, so both flags
silently lost their stars when rendered wider than roughly 76px. That
threshold was always exceeded because loadSvg with expandOnly renders at
the declared 640x480 native size before scaling down to the icon size.
Fold the scale transforms into the polygon coordinates so the geometry is
unchanged while bounds stay small at every render size.
Also cap expandOnly and usericon render canvases at four times the
requested size to bound memory use and keep pathological theme svgs away
from the rasterizer limit.
Took 8 minutes
Took 57 seconds
Took 3 minutes
Co-authored-by: Lukas Brübach <Bruebach.Lukas@bdosecurity.de>
* [Server] Add match result strategy hook
Took 7 minutes
Took 18 minutes
* Rebase.
Took 2 minutes
Took 13 seconds
---------
Co-authored-by: Lukas Brübach <Bruebach.Lukas@bdosecurity.de>
Two gaps made opening and confirming the banner dialog silently switch
the banner to the default art of the first local printing. The caller
dropped the card provider id when constructing the initial params, and
the constructor left the printing combo wherever onCardNameChanged put
it, which is always the first printing
Pass the provider id through, restore it in the combo when it resolves
locally, and keep it verbatim when it does not
Took 2 minutes
Co-authored-by: Lukas Brübach <Bruebach.Lukas@bdosecurity.de>
* [Game] Derive playmat sampling window from shared clamped helpers
The crop formula computed the sampled window inline with an unclamped
zoom floor, so stored vertical offsets below half of travel were dead
and extreme zooms could sample outside the art
Remap verticalOffset to place the window top edge within its travel,
floor zoom at visible width over min card side with a 4.0 ceiling,
clamp pan along the margin sum constant segment, and expose
playmatClampedZoom, playmatWindowSide and aspectFitRect so the game
renderer and any editor share one geometry model
Zoom 1 rendering is bit identical to before
Took 7 seconds
* Comments.
Took 29 minutes
---------
Co-authored-by: Lukas Brübach <Bruebach.Lukas@bdosecurity.de>
* native builds + merge
* naming and ordering
* use ninja and cmake build
* add ccache and cache mounts
* formatting
* Update servatrice.cpp
* Revert "Update servatrice.cpp"
This reverts commit 3acc684135c6db9baa17d00deaceecf8f7079721.
* remove ccache again
cache mounts are not part of GHA caches from docker action
* comments and cleanup
Use buildx provided in runner, see https://github.com/actions/runner-images/blob/main/images/ubuntu-slim/ubuntu-slim-Readme.md
* more explicit
* comments, first pass
* ${{ runner.temp }}
* $(printf "$GHCR_IMAGE@sha256:%s " *)
* follow docker docs for latest and extract short semver from our tags
* not so pretty, but allows the easy inspect at the end
* add Servatrice name
* add links to runner images
* comments, second pass
* cleanup
Cached card art pixmaps carry the screen device pixel ratio, so both
banner painters did their crop math on scaled pixels and blitted the
result at raw over logical size, clipping art into its top left quadrant
on any display above 100 percent
Normalize a local copy to DPR 1 before crop math in UserListPainter and
the popup header, clamp srcX and srcY bounds against stored zoom below 1,
keep shared cache entries untouched
Took 15 minutes
Co-authored-by: Lukas Brübach <Bruebach.Lukas@bdosecurity.de>
* [Client] Show live connection latency in the status bar
Add a permanent status-bar label fed by ConnectionController's
pingStatsUpdated: shows the latest round-trip time, hides while
disconnected or without samples, and carries a tooltip with
Last/Median/95th percentile/Maximum over the rolling sample window
(mirrored into the accessible description). The server tab gets the
same stats as its tooltip.
Took 2 minutes
Took 2 minutes
Took 1 minute
Took 48 seconds
Took 25 seconds
Took 4 minutes
* [Client] Graph connection latency history in the status bar
Add LatencyGraphWidget, a size-agnostic bar sparkline over the rolling
sample window: heights scale to the window's own range while colors map
onto an absolute quality ramp, so a steady good ping stays green. Embed
it in the new LatencyStatusWidget together with the textual ping
readout and feed both through ConnectionController's forwarded signals;
the whole area hides while disconnected or without samples.
Took 9 minutes
Took 14 seconds
* [Client] Show latency details when clicking the ping display
Clicking the status bar ping area opens a popup with a larger instance
of the latency graph plus the numeric statistics, selectable and
mirrored into the accessible name. Qt::Popup closes it on any outside
click; contents refresh live while open.
Took 33 seconds
* Fixup from core commit
Took 6 minutes
Took 5 seconds
Took 5 minutes
* Lint.
Took 12 minutes
* Consolidate.
Took 6 minutes
---------
Co-authored-by: Lukas Brübach <Bruebach.Lukas@bdosecurity.de>
* [Client] Detect main-thread event loop stalls
LagMonitor ticks the GUI event loop every 500 ms and records gaps
beyond 2 s as stalls, warning with their duration and keeping a
bounded ring of recent records for diagnostics. Measurement uses a
monotonic QElapsedTimer so wall-clock steps and suspend do not
fabricate stalls. Recorded timestamps stay in wall time for
correlating with user reports.
Took 1 minute
Took 13 minutes
Took 2 minutes
* [Client] Rename LagMonitor constants to SCREAMING_SNAKE_CASE
Took 15 minutes
* [Client] Discard suspend-spanning gaps in LagMonitor
Windows counts sleep time in its monotonic clock, so a suspend would
fabricate one bogus stall per resume. Reset the clock on application
state changes and drop implausibly huge gaps; extract recordGap() for
testability.
Took 3 minutes
* [Client] Unit test LagMonitor stall recording
Drives recordGap() directly to cover the threshold, plausibility cap,
trim, and clear behavior without timing-dependent waits.
Took 36 seconds
---------
Co-authored-by: Lukas Brübach <Bruebach.Lukas@bdosecurity.de>
* [Network] Measure real server round-trip times
Time each command container from send to response with QElapsedTimer,
aggregate samples in a fixed-size ring buffer (last/median/p95/max),
and emit aggregated pingStatsUpdated at most once per second so the
hot path stays free of signal traffic. Stats are cleared on
disconnect. Forward the signal through ConnectionController for UI
consumers. Unit-tested in latency_tracker_test.
Took 37 minutes
Took 4 minutes
# Commit time for manual adjustment:
# Took 8 minutes
* Move params to struct, more informative debug
Took 56 seconds
Took 53 seconds
Took 2 minutes
Took 33 seconds
---------
Co-authored-by: Lukas Brübach <Bruebach.Lukas@bdosecurity.de>
