aboutsummaryrefslogtreecommitdiffstats
Commit message (Collapse)AuthorAgeFilesLines
* api: support serving under a path prefixtv22 hours2-1/+78
|
* api: /stream & dashboard send X-Accel-Bufferingtv22 hours4-1/+22
|
* api: allow auth delegationtv22 hours9-38/+243
|
* dashboard: shorten long figures titletv24 hours1-0/+6
|
* dashboard: add B/stv24 hours2-23/+47
|
* dashboard: add msg/stv24 hours2-1/+58
|
* dashboard: message dot speed is configurabletv24 hours2-21/+45
|
* dashboard: add console filtertv24 hours2-2/+50
|
* dashboard: focus wins over hovertv27 hours2-24/+35
|
* dashboard: nodes and edges tint related messagestv27 hours2-3/+44
|
* dashboard: messages tint related nodes and edgestv27 hours2-5/+40
|
* add per-node traffic countstv27 hours4-5/+67
|
* add documentationtv32 hours15-0/+3575
|
* document: put user in commit message trailertv32 hours3-10/+80
|
* remake: read ops, not diffstv33 hours2-6/+6
| | | | This gives bench_replay a speedup of ~8x/64x/280x.
* api: /admin/changes read ops, not diffstv33 hours2-10/+90
| | | | | | | | | | | | The new store::changed_by reads each change's own operations instead of diffing the document before and after it. Each object's path is found once per change, as it was right after that change. So the cost now grows with the size of the change, not the size of the map it touched. - It names paths the same way the diff did: a map entry by its key, and list and text edits by the list or text. - It relies on import_obj, which Automerge marks as hidden from its docs, to turn an operation's object ID back into an object.
* api: drop $float as JSON number is OKtv33 hours4-35/+50
|
* api: add type caststv33 hours6-10/+269
|
* add remake benchmarkstv33 hours4-3/+94
|
* api: types are stickytv33 hours5-99/+317
|
* api: add application/vnd.continuum.typed+jsontv34 hours4-192/+383
|
* api: add application/vnd.continuum.* scalar typestv34 hours7-130/+809
|
* add document benchmarkstv45 hours3-0/+175
|
* api: add x-continuum-heads to writestv45 hours3-27/+63
|
* probe when reachabletv45 hours10-1/+69
|
* probe wakes peer workertv45 hours1-2/+3
|
* document: create new compound values at oncetv45 hours1-11/+29
| | | | | This speeds up large writes as new maps and lists are created in one batch operation instead of one operation per value.
* document: introduce patch logtv45 hours5-22/+148
|
* api: add support for counterstv45 hours4-8/+108
|
* api: allow reading past versionstv45 hours6-10/+325
|
* api: introduce /admin/changestv46 hours4-1/+248
|
* record node ID and API user for every writetv46 hours3-14/+77
|
* api: PUT changes only what differstv46 hours2-27/+153
|
* dashboard: always try showing ping/pong aliastv46 hours3-2/+32
|
* we're a replicated document, not KV storetv46 hours3-5/+5
|
* ping & pong are tied by nonce onlytv46 hours3-7/+22
|
* large messages over slow links are deltastv47 hours4-14/+85
|
* introduce per-transport sync reply timeouttv47 hours9-13/+56
|
* path probes follow keepalivetv47 hours3-9/+41
|
* libp2p: idle timeout follows keepalivetv47 hours5-2/+22
|
* tor: read max sync messagetv47 hours1-3/+5
|
* links: wait for subscriber instead of tickertv47 hours1-6/+16
|
* remake: download wakes on end instead of tickertv47 hours2-6/+15
|
* peer worker forget records on changetv47 hours6-4/+32
|
* peer worker sleeps until deadline, not fixed timetv47 hours10-32/+170
|
* bootstrap: replace pull by progress notifytv47 hours11-2/+52
|
* dashboard: intended edges are loosetv48 hours1-7/+14
|
* dashboard: allow message inspectiontv48 hours4-31/+292
|
* dashboard: 24h-clocktv48 hours1-1/+4
|
* dashboard: one spring per edgetv48 hours1-4/+6
|