aboutsummaryrefslogtreecommitdiff
AgeCommit message (Collapse)Author
2020-02-05schedule.js prevents off-by-one error by popping minimum amount in step 3Holden Rohrer
2020-02-05included bullet art request in jarvisHolden Rohrer
2020-02-05fixed a BIG bugHolden Rohrer
So, this is pretty embarrassing, but I didn't realize that I was pushing the entire stack whenever 0 chars were supposed to be pushed because I thought that javascript's built-in functions were ok with negative zeros (stupid me)
2020-01-31proper mathematical fix in schedulerHolden Rohrer
2020-01-29Merge branch 'socket'Holden Rohrer
Made the socket object used by jarvis a singleton MetaSocket object to maintain clean, modular code standards. This also means that future feature addition will be easier because all references are to one object, of which the behavior can be easily changed since most systems rely on EventEmitter or callback-style structures
2020-01-29generalized tools/search modelHolden Rohrer
search.js now provides for a .match and .unmatch method, so it doesn't need to be called every time (in-line with the singleton MetaSocket object model, meaning that callbacks are used and several spaces can be matched at a time)
2020-01-29jarvis uses mainworld nowHolden Rohrer
2020-01-29fixed mathematical model in tools/schedHolden Rohrer
Used to keep `num` the same and decrease the sum of weight, which created an error because the value of `num` was used in dequeuing job components. This patch prevents `dequeue()` from pushing more objects than it's actually able to.
2020-01-24jarvis changed in preparation for search refactorHolden Rohrer
2020-01-24jarvis integrated new changesHolden Rohrer
2020-01-24wwrap can be used for socket extensionHolden Rohrer
2020-01-24tools/claim now uses 'fallback' eventHolden Rohrer
2020-01-24Identity function built for .call useHolden Rohrer
In an effort to move jarvis to using one class, MetaSocket, for all of its interactions with the main socket, ident is now only able to be used as id.call(socket); this should simplify the behavior and means that it can use EventEmitter events.
2020-01-24made protectArea more efficientHolden Rohrer
jarvis now calls every protection update with the same movement instead of tile-by-tile, which would delay parts of the protection by a second
2020-01-24fixed vulnerability in schedHolden Rohrer
This vulnerability permitted external users to overwrite the alert for long periods of time
2020-01-21merged master to socketHolden Rohrer
2020-01-21two minor typo-like bugfixesHolden Rohrer
2020-01-20bug statementHolden Rohrer
2020-01-20minor jarvis changesHolden Rohrer
2020-01-20removed inaccurate claim.js commentHolden Rohrer
2020-01-20removed extraneous code from tools/searchHolden Rohrer
2020-01-20scheduler built for external accessHolden Rohrer
.dequeue has been made available, along with an infinite delay-time so that a secondary loop can manage multiple schedules at a time. This is paired with .size to make the external loop more efficient (hopefully, baudrate is used optimally)
2020-01-20added contactHolden Rohrer
2020-01-20scheduler has .pause functionHolden Rohrer
2020-01-20added a land claim systemHolden Rohrer
2020-01-20forgot to init var in writewrapHolden Rohrer
2020-01-20jarvis notif generation split outHolden Rohrer
On restoration of websocket, the notif actually gets pushed again, preventing pesky overwrites during downtime.
2020-01-19added wrapper for socket.js writesHolden Rohrer
2020-01-19socket returns a more useful number on .writeHolden Rohrer
2020-01-19uncaught typo I thought I fixedHolden Rohrer
2020-01-19schedule now rejects empty enqueuesHolden Rohrer
2020-01-18added new error case to socket.jsHolden Rohrer
2020-01-18simplified and corrected protectArea diffingHolden Rohrer
2020-01-18rectintersect typo for embarassingly long timeHolden Rohrer
2020-01-18jarvis.js reorgHolden Rohrer
2020-01-18schedule had a glaring errorHolden Rohrer
2020-01-19fixed other utils/ident bugsHolden Rohrer
2020-01-18typo in identHolden Rohrer
2020-01-18Schedule bugfixHolden Rohrer
It was adding too many prios to prios[] for the number of available jobs[] queues because they were added on every job insertion instead of on every new queue creation
2020-01-18moved deinit on close to ident.jsHolden Rohrer
2020-01-18jarvis externalized identify()Holden Rohrer
2020-01-18jarvis cursor detection was still brokenHolden Rohrer
2020-01-18padded uptime calculation in jarvisHolden Rohrer
2020-01-18Merge branch 'schedule'Holden Rohrer
Actually have a proper algorithmic scheduler in examples/jarvis.js now, allowing for low-latency responses and complex animation, etc.
2020-01-18Implemented bug fix for maxr in schedule.jsHolden Rohrer
Rankings may have been messed up because as new insertions get added, the data.length changes per write.
2020-01-18realized that .once doesn't work for cursorHolden Rohrer
2020-01-18queue-compat and bugfix in schedule.js---calls from jarvis still don't workHolden Rohrer
2020-01-18moved jarvis to scheduleHolden Rohrer
2020-01-17reorg comments in tools/schedule.jsHolden Rohrer
2020-01-17typo in tools/schedule.jsHolden Rohrer