aboutsummaryrefslogtreecommitdiffstats
path: root/src/federation
Commit message (Collapse)AuthorAgeFilesLines
* federation: a checkpoint of 64 heads at most, a version of 96 partiestv15 hours3-5/+56
| | | | | | | | | | | No event is above more heads than it may name dependencies (64): a version naming more for its checkpoint is never written. A node voting by itself proposed one at whatever heads it held for a while, however many; it now proposes none past 64, as what is written merges them, and a party's client backs none such (`client::backable`). A version event names the votes passing a proposal, two thirds of its version's parties: past 96 parties, more than an event may name. A version names at most 96 now (`version::PARTIES`).
* store: a new event builds on as many heads as an event may nametv15 hours5-3/+37
| | | | | | | | | | | A party's client builds on the heads its node shows (`GET /events`), and a node on its own heads; an event names at most 64 dependencies. With more heads than that, where many write at once and nothing merges, every event built on them was none. `Store::building_on` gives the heads, as many as an event may name (`Fold::DEPENDENCIES`, the federation's 64), the earliest taken first, so that what is written merges what stayed apart the longest; `GET /events` shows those, and a node voting by itself builds on them.
* federation: votes on the events proposing a checkpoint above its headstv15 hours4-16/+147
| | | | | | | | | | | | | | | Anyone could propose a checkpoint's version again, by an event beside its heads. A node voting by itself, and a party's client, built a vote on every event proposing what it voted on, so every vote was beside the heads, and no version event written on them was taken: the checkpoint could never move the store. And the voters' rule refused it once that event was held for a while, as the heads left it out. They vote on the events proposing it that are above the heads it names (`State::proposing`), which `GET /events/counts` lists as `proposals`; where there are none, `none` says so, the node casts nothing and the client says why. The voters' rule leaves proposals and votes aside, as they lapse with the move. Row 70 is tested (`federation_attack_70`).
* fold: of a party's lists, those written under the latest version firsttv15 hours3-14/+55
| | | | | | | | | | | | | | | | | | A key a version names, stolen, could write its party's node list at a history from before the version that named the party's keys anew: written under the version before, it counted, as naming keys anew leaves a list's `since` alone, and beside the list the party wrote under the new version the lower ID won. The thief signed a few, kept the lowest, and did so again after every list the party wrote. Of a party's key sets, and of its node lists, those written under the latest version now come first, then what no other has in its history, then the lowest ID; the model alike. A list the party writes under the version that replaced the key stays, as `federation.md` says. Key sets were safe already, by their `since`; one rule serves both. This changes what counts where two of a party's lists, written under different versions, are beside each other: nodes read them alike only once all run it. Row 69 is tested (`federation_attack_69`).
* model: stops at a version it doesn't readtv21 hours2-14/+82
| | | | | | | | | | | | | The model stops where `federation.md` says a node does: at a version event whose proposal passed at its history, isn't a version as this continuum reads it, and changes the version the model reads, it takes and refuses nothing more (`Verdict::Stopped`); beside that version, such an event is refused, as the node refuses it of its own. The property checking the store against the model runs on histories where parties propose such versions too (a field this continuum doesn't know): written in order, events come after what they build on, so both must stop at the same event, having taken the same before it, the adversary's two versions changing one among them. A check that some history stopped keeps the property from passing vacuously.
* events: why no party's client backs a proposaltv21 hours3-0/+13
| | | | | | | | | | | | `GET /events/counts` gives each proposal on the current version `none`: why a party's client backs it not (`client::backable`), where it is no version as this node reads it, none changing the current one, or one naming the heads of the current one's checkpoint; else `null`. So a node shows that a proposal is none, and why, before it passes: one that passes all the same is never written, or, no version here, the node stops at its version event. The dashboard shows it with the proposal, and `continuum-party proposals` says it by the federation the client knows. Row 48's test checks what a node shows of the proposal.
* version: say why a version is nonetv21 hours2-38/+99
| | | | | | | | | | | | `Version::read` names what makes a version none: the field it lacks, or the one more it has, that this continuum doesn't know (of the version, a party's keys or node list, a rule), the federation it names beside the one this continuum knows, how many keys or nodes a party has against how many it may, the node ID that is none, and of a key why it is none (`keys::spelled`): its type, its point, or its spelling. A node that stops at a version says so in `stopped` (`GET /events`) and its log; a party's client in what it won't propose or back. The guide to a node that stopped says what `why` tells.
* store: a stopped node leaves out none of what it holds backtv21 hours2-7/+143
| | | | | | | | | | | | | | | A store that stopped can't tell which of what it holds back will be taken, nor which of what waits: `Store::held_back` gives both, each after what it builds on. A node leaves out none of it: it puts it again where its document lacks it (`Holding::lacking`), a remake's snapshot holds it (`Holding::at_base`), and after a switch or a rejoin it puts again what of it the document it left held (`Holding::left`). A remake while nodes are stopped no longer drops what was put since the stop, nor the event they stopped at. Tested: a node keeps what it holds back, what waited among it, through a snapshot, and a store read anew from that stops there again; and a simulation: a swarm stopped at a version compacts, a node away meanwhile rejoins and one started again with nothing reads the new epoch, each stopping there, holding back what came after, until their operators have them drop that event.
* federation: a joining node takes the first of two versions, as any nodetv22 hours1-1/+113
| | | | | | | | | | | | Of two versions changing the one it reads, a node takes the first it judges, joining or not, and reports the other: its operator chooses, by dropping and allowing. No rule holds a joining node back until it has synced with every seed: it learns of the other members as it syncs, and its seeds may go away. federation.md drops the joining rule from Changing the policy, and says so in what isn't promised and in row 63; the guides say what a joining node takes. Row 63 is tested (`federation_attack_63`): from the root, either side first, the operator switching sides; and pinned at a later checkpoint, the chain old keys grew never taken.
* federation: a node stops at a version it doesn't readtv22 hours7-31/+415
| | | | | | | | | | | | | | | | | | | | | | A version event writing a proposal passed on the version a node reads, which doesn't read as a version there (of a later federation, using what a later continuum takes, or none at all), the node can't judge: `Fold::stops`. The store stops at it: it takes nothing and refuses nothing from then on, holding back whatever comes, that event first (`Offered::Stopped`), and keeps none of it across a restart, where it stops there again. One beside the version the node reads is refused of its own, as any other. Its operator has it drop that event (`POST /admin/events/drop`), or one it builds on: it is refused as dropped, or as built on what was, any version event writing that version with it, kept in the store file; the node judges what it held back and goes on under the version before. Allowed again, it stops there once more. The node logs the stop and its end; `GET /events` names the event, why, and how many it holds back (`stopped`); a put answers `stopped`, written into the document all the same; a node voting by itself casts nothing while stopped. The test fold's eras tagged `new` stop a store. Row 39 is tested (`federation_attack_39`); of rows 9, 10 and 48, a node stops at the version event.
* store: a drop refuses whatever writes what a dropped event writestv24 hours6-22/+203
| | | | | | | | | | | | | A fold may say what an event writes that others may write alike (`Fold::writes`): of the federation, a version event's version. A node whose operator has it drop such an event drops the taken events writing the same with it, and refuses any writing it from then on, whoever wrote it anew: of a version, it takes no other version event naming it and writes none itself, though it holds the votes that passed it. The node keeps what each dropped event writes in its store file, across restarts; allowing one of them takes them all back. Row 51 of the federation's attacks has its test, and waits no more.
* attacks: rows 45, 58 and 65 tested beside the code they testtv24 hours4-5/+161
| | | | | | | | | | | The check that every attack row of the federation has its test looks beside the code a row is about too: a node's (`held`), a party's client's (`client`), and a node voting by itself (`vote`). Rows 45 (a version written on the side a node doesn't read, refused as its own and reported apart, the change that put it taken), 58 (a node following on checkpoints only, unless its operator chose any version) and 65 (a client keeping the nays that freed its `true`) have their tests there, and wait no more.
* continuum-party: proposes only what a party may backtv29 hours2-12/+54
| | | | | | | | | One check (`client::backable`) decides what a party's client proposes and what it votes `true` on: a version, changing the current one as a version must, whose checkpoint, where it names one, names none of the heads of the current version's checkpoint, as a version naming those again is never written. A proposal nobody may back is refused before it is signed, saying why.
* continuum-party keys, nodes: shown given no change, --set, --forcetv29 hours1-46/+110
| | | | | | | | | | `keys` and `nodes` alone (`--list`) show the party's key set or node list as it counts at the node, one a line, without a key. `--set` writes it whole, none twice, in place of the arguments that did; `--add` and `--remove` change it as it counts. A set or list left empty, whichever way (`--set` with none, `--remove` of the last), is written with `--force` only: none of the party's writes count without keys, nor its nodes as members without a list.
* continuum-party: votes and proposals checked for the key a version namestv29 hours1-11/+16
| | | | | | A vote or a proposal is signed with a key the current version names for the party, as a key set and a node list are: the client checks it before it signs, and records nothing where the key isn't one.
* continuum-party keys, nodes: --add and --remove, and the key they needtv29 hours1-18/+198
| | | | | | | | | | | | | `keys` and `nodes` write the party's key set or node list whole, or, with `--add` and `--remove`, the set or list as it counts at the node with those added or taken off, each of which must change it; the client says what it writes. Node IDs are checked as the federation spells them. A set or list is signed with a key the current version names for the party: the client checks that first, and where it isn't one, or the party is none of the version, says that only a version naming the party's keys or list anew changes them, which a party proposes and the parties pass.
* continuum-party propose: the proposal's hash, and --yestv29 hours1-6/+18
| | | | | `propose` says the proposal's hash, the version's, which `vote` takes; with `--yes`, it votes `true` on it too, by the record, as `vote` does.
* continuum-party propose: only a version changing the current onetv29 hours1-1/+8
| | | | | | The client proposes a version only where it changes the current one as a version must (its number, the hash it names, each `since`), and says why it doesn't otherwise: no other can pass.
* continuum-party next: the version after the current one, to changetv29 hours1-0/+44
| | | | | | | `next` prints the current version numbered on, naming its hash as the version it changes, and without the checkpoint it names, whose heads are behind: what a party changes and proposes. The guides begin a version with it.
* continuum-party: why waiting for the ssh-agent may blockecdsatv30 hours1-0/+4
| | | | | The client signs one event at a time and does nothing else meanwhile, and an agent may wait for its user as long as that takes.
* continuum-party: a key the ssh-agent holds, given itselftv30 hours2-19/+107
| | | | | | | | | `--key` gives a public key itself, an `authorized_keys` line, as well as a file of them: of a file of several, the first signs, with a warning naming it, and another is given itself to sign with it. An agent that doesn't hold the key given signs nothing, and the client stops before it records or puts what it would have signed, saying which key it asked for.
* keys: signing through an ssh-agent tested wherever tests runtv30 hours1-21/+21
| | | | | | The test fails without OpenSSH on the machine testing, rather than passing without asking anything. Its directory is a scratch directory, removed when the test ends, the agent stopped before it.
* files: a test's scratch directory goes when the test endstv30 hours1-1/+2
| | | | | | | `scratch_dir` gives a guard (`Scratch`) that removes the directory with all it holds when dropped, the test passing or failing. Each test keeps it for as long as it uses the directory; the dashboard's browser check writes its page into one.
* attacks: row 62 with an ECDSA signature spelled otherwisetv30 hours2-0/+34
| | | | | | | Row 62 names an ECDSA signature's twins, `s` high (the order less `s`) and `r` or `s` with zeros more in front, each none. Its test offers the ECDSA vector's write with each (`keys::respelled`) to a store whose root names that key, and has it take the write as it was signed.
* continuum-party: signing through an ssh-agentslopulus45 hours5-19/+250
| | | | | | | | | | | | | | A key held in hardware is never a file: only an ssh-agent signs with it. Given a public key as `--key` (a `.pub` file, `ssh-ed25519` or `ecdsa-sha2-nistp256`), the client asks the agent at `SSH_AUTH_SOCK` to sign what SSHSIG signs (a sign request with the key's blob, no flags) and writes the blob around the agent's signature as `ssh-keygen -Y sign` does. An ECDSA signature the agent gives with a high `s` is made low, the other of its two spellings, which the federation takes. What it makes is checked before it is used. A private key file signs as it did. A test starts `ssh-agent`, adds an ECDSA and an Ed25519 key to it and deletes their files, signs through it, and has both continuum and `ssh-keygen -Y verify` check each signature, one of an agent's high-S answers among them. Without OpenSSH it asks nothing.
* keys: a party's key may be ecdsa-sha2-nistp256, its signature low-S onlyslopulus45 hours4-33/+300
| | | | | | | | | | | | | | | | | | | | A party's keys may be held where no file has them, in a Secure Enclave or a phone, and those are ECDSA keys on NIST P-256. Everywhere a party's key counts (a version's keys, a key set, the key an event is signed with) such a key is taken beside an Ed25519 one, spelled as `ssh-keygen` writes it, without a comment: `ecdsa-sha2-nistp256`, a space, and the base64 of its SSH encoding, its point uncompressed and on the curve. Its signature is SSHSIG as an Ed25519 one is, hashed with `sha512`, its inner signature over SHA-256 as SSH signs with such a key. ECDSA signatures are malleable: of `s` and the curve's order less `s`, both check. So the federation takes one spelling alone: `s` low, at most half the order, `r` and `s` each SSH's shortest mpint, and the blob reading back to itself as an Ed25519 one does. What anyone makes of a party's signature is then none, as before. The vectors are a key `ssh-keygen -t ecdsa -b 256` made, a first version naming it, and a write and a key set `ssh-keygen -Y sign` signed with it, each taken; the signature's high-S twin, which checks as ECDSA, and its `r` spelled longer are none. Ed25519 keys and signatures are spelled and checked as they were.
* events: the store's summary names the node's party, and whether it votes for ittv2 days1-0/+5
| | | | | | `GET /events` names the party the node is of (`party`, `--store-party`) and whether it votes for it by itself (`votesByItself`, `--store-vote-key`), and so casts a vote by hand (`POST /admin/events/vote`): what a page offering votes through the node asks.
* continuum-party: options from configuration files, as the node'stv2 days1-44/+70
| | | | | | | | | `continuum-party` takes its options (`node`, `token`, `key`, `record`: `config::PARTY_OPTIONS`) on the command line and in configuration files of the node's format: `--config`, repeatable, arguments winning, `--unset`, includes and imports. The configuration's loader takes the options it reads as a table (`config::with_options`, `config::merged`, `file::load_with`), the node's or the client's, and a key naming none of them is an error in either; `node` and `token` are secrets, never logged.
* continuum-party proposals: where each proposal stands, and what it changestv2 days2-0/+125
| | | | | | | | `continuum-party proposals` lists the proposals on the current version at the node, each with whether it passed or is dead and the parties voting `true` and `false`, as `GET /events/counts` names them, and what its version changes of the current one, read from an event proposing it, its signature checked: each field whose value differs, and of `keys` and `nodes` each party's (`client::changes`).
* events: counts names who voted what, and whether a proposal passed or is deadtv2 days1-14/+71
| | | | | | | | `GET /events/counts` gives each proposal on the current version the parties voting `true` and `false` (`yes`, `no`), whether it passed and whether it is dead (`passed`, `dead`), as the node reads its votes (`State::tally`, which the fold judges by too), so that reading where a vote stands takes no fetch of every vote. A party's client still decides its votes by the votes it checks itself.
* continuum-party keys: a key set from authorized_keys lines, comments left outtv2 days1-3/+11
| | | | | | The federation spells a key without a comment, so a key set holding a line's comment is none. `continuum-party keys` takes each key as an `authorized_keys` line, one argument each, and writes it as the federation spells it.
* client: a record begun from the store's root casts on its first versiontv2 days2-9/+42
| | | | | | | | | | A record casts nothing on the version current when it began, as a `true` sent before may be held back somewhere. A federation's first version is current from its start, so a record begun at a node casts nothing on it, and with every party's begun so, no proposal on it can pass. `continuum-party init PARTY --root FILE` begins a record from the file the store's swarm starts from, for a party that never voted in it (`Record::before`): nothing was sent before it, and it casts on every version, the first too. A node voting by itself takes such a record as its `--store-vote-record`.
* held: the store's worker reads a document only while it is writabletv2 days1-8/+8
| | | | | | | | | | | A rejoin makes the node's document writable again only as it ends, in the lock of the swap, once the node checked the epoch (`Holding::follows`) and follows it or gives it up. The store's worker reads nothing, and writes nothing, while the document isn't writable, checked in the lock it reads under. Whether a rejoin runs isn't enough: its state is taken up for its last checks, while the epoch's document is still the node's, and a store started from that document would keep the checkpoint it read there, and put again what it took into the document the node goes back to.
* held: nothing of a rejoin's document is read before the node follows ittv2 days1-0/+6
| | | | | | | | | While a node rejoins, its document is that of the epoch it tries, which it follows only once it checks it (`Holding::follows`). The store's worker reads nothing of it, and writes nothing, while the node rejoins: else the store, starting there from the snapshot's checkpoint, would take one the node neither builds nor has pinned, and keep it should the node give up the rejoin.
* simtest: made-up epochs handed to a node holding a store (rows 15, 31)tv3 days1-1/+1
| | | | | | | | | | d serves an epoch of the next generation it made up, of a lower ID than the others', its snapshot holding the root as the document stood before the checkpoint they moved to. A node joining with nothing that reaches only d takes it, the root being what its operator named, and reads only what it holds (15); one reaching only the others takes not their epoch, whose checkpoint it neither builds nor has pinned (15); the others try d's, find its root before their document's checkpoint, and stay on theirs (31).
* held: a checkpoint a node can't build, handed by peers, only pinned (row 22)tv3 days1-1/+35
|
* simtest: the remake rows of a swarm holding a storetv3 days1-1/+1
| | | | | | | | | | | | | Simulations of the store's attack rows on remakes (`docs/store.md`, Attacks): a snapshot other than its base's isn't followed (11); one coordinator's two decisions for an epoch, shown to different nodes, end on the lower epoch (13); a node away while the others switch rejoins and puts again what it wrote meanwhile (16); after a compaction whose snapshot holds a checkpoint, a node away all along follows only once its operator pins it (24); a compaction whose base comes before a checkpoint, switched after nodes moved to it, leaves them there, and a node starting with nothing reads the new document to it (28); a member deciding many compactions at once moves every node to the lowest epoch, losing no event (30).
* simtest: events written beyond a compaction's base go again (row 12)tv3 days1-1/+1
| | | | | | A swarm holding a store compacts: events written beyond the base, on nodes that switch later, are put again in the new epoch, and every node holds every event, reads the same, and refused no change.
* remake: holding a store, a node follows an epoch only from a checkpoint it ↵tv3 days1-0/+122
| | | | | | | | | | | | | builds A node of a swarm holding a store rejoining an epoch (`docs/store.md`, Remakes) follows it only where its snapshot has the swarm's policy and store, and holds the node's document's checkpoint, one it builds from that document, or one its operator pinned; with an empty document, the root of its store too (`Holding::follows`). A made-up epoch holding another, the root after the store moved on, say, would leave it holding events whose history nobody puts again: it stays on its own, as when a rejoin's check fails, and tries again later.
* remake: holding a store, a node puts again what the document it left holdstv3 days1-36/+120
| | | | | | | | | | After a switch or a rejoin, a node of a swarm holding a store puts again what the document it left holds that a store starting at the new document's checkpoint takes, by its own refusals and drops, and the new one lacks (`docs/store.md`, Remakes, `Holding::left`): what it held beyond the base, and, where it moved to a later checkpoint since, what a store starting at the snapshot's takes to reach it. Its events aren't replayed into the new document, so none nobody takes go along.
* vectors: the snapshot after a checkpoint, as a run printed ittv3 days1-1/+4
|
* vectors: a snapshot after a checkpointtv3 days1-0/+75
| | | | | | | The snapshot of a compaction after the store moved to a checkpoint, built from a document every run builds alike: its change's hash and bytes, which a run of this code prints, as no script apart encodes Automerge. Left empty until a run prints them.
* remake: a holding swarm's snapshot holds the store read at its basetv3 days1-0/+149
| | | | | | | | | | | A compaction's snapshot in a swarm holding a store holds the store as the node reads it from the document at the base, by its own refusals and its operator's drops (`docs/store.md`, Remakes): the checkpoint it reaches there, where it isn't the root, at `_continuum/checkpoint`; the root, the events that checkpoint keeps and those after it, each at its ID; and of the entries, the members'. What the checkpoint forgot, events nobody takes, and the entries of nodes that aren't members stay behind (`Holding::at_base`, `events::snapshot`, `Host::holding`).
* held: a store starts from the checkpoint its document holdstv3 days1-19/+80
| | | | | | | | A remake's snapshot holds the store's checkpoint at `_continuum/checkpoint` where it isn't the root (`docs/store.md`, What the document holds), and none of what came before it. A node reading its store anew from such a document starts there, with what the checkpoint keeps; one pinning a checkpoint finds it there too, where it can't build it from the events.
* holding: a node writes its acks and the compactions it coordinatestv3 days1-1/+1
| | | | | | | | | | | A change of a swarm holding a store may write its node's own protocol data a remake needs (`docs/store.md`, What a change writes): its acks, a remake's record and proposal, where there is none, of a compaction it coordinates, and, as the coordinator, its decision, or the record of the remake that made the epoch, carried into it whole; and the maps holding them, where there are none. An abort asked for by another than the coordinator stays closed, as does any other kind of remake. Rows 14 and 21 are judged here.
* federation: the test of a node voting above a checkpoint's heads holds for a ↵tv3 days1-0/+2
| | | | | | | | | while With no time to hold an event for a while, lass's write, taken just before the proposal, counted as held for a while before it was proposed, and the node backed no heads leaving it out. The test's node now holds an event for an hour before it holds it for a while.
* federation: votes build on the events proposing what they vote ontv3 days2-5/+82
| | | | | | | | | | A node voting by itself, and a party's client, signed a vote on the heads the node held: where those held what is beside a checkpoint's heads (a party's writes made while the swarm was parted, say), the vote wasn't above them, and a version event written on it was refused, so the checkpoint never moved, though it passed. Both now vote on the events proposing what they vote on, and a node proposes a checkpoint on the heads it names: so their votes are above those heads wherever the proposal is.
* federation: the changes founding a swarm, fixedtv3 days1-29/+55
| | | | | | | | The changes founding the swarm of the vectors' first version's root, and those founding the fork whose root is the fork vector's, keeping tv's write, each by its hash and its bytes, as a run of this code printed them: Automerge encodes them, so no script apart computes them. Checked apart so far: each hash the SHA-256 of its change's chunk, each writer the first 16 bytes of its root's ID, the fork's second change built on its first.
* simtest: a fork whose root doesn't read as its heads do, found by who holds themtv3 days1-3/+2
| | | | | | | | | `federation_attack_33`: beside the fork a node holding the history makes, one whose root leaves out tv's write, which counts at its heads. b, holding the history, builds the one and not the other (`fork::check`). Each fork's founder starts its swarm from its root and the events it keeps, and a node of lass joins each by its ID, taking it on its operator's word, and reads what the root keeps: tv's write in the one, carried by the changes founding its swarm, nothing in the other.
* federation: the changes founding a swarm printed, to be fixed as vectorstv3 days1-0/+41
| | | | | | | An ignored test prints, each by its hash and its bytes, the changes founding the swarm of the vectors' first version's root, and those founding the fork whose root is the fork vector's, keeping tv's write: Automerge encodes them, so no script apart computes them, and they are fixed once printed by a run of this code.