| Commit message (Collapse) | Author | Age | Files | Lines |
|---|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Going to another member's dashboard showed its login page for a moment: a browser
without a session there was sent from `/admin/dashboard` to `/login` (by
`auth::basic_auth`), which began a session with the token handed over and went on to
the dashboard.
Now the dashboard's page is served without authentication, as the login page is (it
holds nothing: what it shows it reads with the browser's session). Opened with a token
handed over, the page begins the session with it (`POST /auth/session`) before it reads
anything; without a session, `obtainTokens` goes to the login page, as the redirect
did. A request without credentials is refused at once rather than after the 1.5 s
that slow down guessing passwords: nothing is guessed there, and a page without a
session goes to the login page without waiting twice for it.
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
The page renewed its token half way through its life by a timer. A timer fires late
where the computer slept or the tab was in the background, by when the token (and the
session cookie, which lives as long) may have ended: the page then fell back to
`POST /auth/login`, which takes a password the browser keeps (basic authentication),
and without one retried every 30 s, logged out without saying so.
Now whether the token is half way through its life is looked at again when the page
shows and every 30 s (`refreshIfDue`, one renewal at a time), so it is renewed while it
still can be, until the login ends (`--session-lifetime`). Where the login ended, or
there is none (400, 401, 403), the page goes to the login page; a node away is tried
again in 30 s. And the page no longer gets in by a password the browser keeps: only
by its session or through a proxy naming the user (`throughProxy`, `login: "header"`).
The node never challenges a browser for a password either (`WWW-Authenticate`, the
browser's prompt): not only on the dashboard's paths and for a page's own requests, but
for any request with `Sec-Fetch-*` or a `Mozilla/…` user agent. curl sends the password
unasked with `-u`, as before.
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
A node in focus showed this node's `/peers` entry for it and the links it reports:
what this node knows of it, never what it says of itself, though the page subscribes to
it. Now a subscribed node in focus shows its own figures first (`Node:`), as this
node's are (its `/admin/stats`), then this node's entry for it (`as this node sees it`)
and its links, each a branch. A link in focus shows each end's paths to the other as
that end has them: this node's, and a subscribed node's from its `/peers`.
The page reads them from the node every 2 s while it is in focus (`readFigures`,
`figuresOf`), with its narrow token, which may look at a node as any user with a
capability may; pages of other origins may now call `/admin/stats` and `/peers`
(`cors::open`), as they do a node's links and messages. Where they can't be read, the
box says why.
|
| |
|
|
|
|
|
|
|
|
| |
A subscribed node's messages come without being written out (their stream leaves the
text out: up to 8 KiB each, 500 at a time), so the figures box said only when, both
ends, type and size of one. Now the node answers for one message by its number,
`GET /admin/dashboard/messages/own/{seq}` (`{"text": …}`, `404` once it isn't kept),
which pages of other origins may call as they do its streams (`cors::open`), and the
page reads it when the message is shown, with its narrow token, and keeps it on the
message. Where it can't be read, the box says why.
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
The console said with whom a message went, not how: the message log was noted above
the identity layer (`CountingTransport`), where a message is to or from a node, and the
address it went to (the paths layer picks one, `Paths::send_first`) or came from is
gone. Only the paths layer's own PINGs and PONGs, logged under their addresses, had one.
Now the identity layer notes the node's messages (`MessageLog::set_addressed`), with
the address each went to (`IdentityTransport::send_proven` returns it) or came from;
`CountingTransport` leaves them to it, and still notes them where there is no identity
layer (tests). A console line ends with the transport (the address's scheme, in its
colour on the map, the address in the tooltip), and the figures box shows the address.
A subscribed node's messages carry theirs too, and the map runs every message along
its transport's line, not only the PINGs and PONGs.
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Setting a node's log filter took `write` on `_continuum`, all of the protocol data:
whoever could change one node's logging could change the swarm's members, remakes and
users too. Now what a node holds alone has paths of its own for capabilities, under
`_continuum/local/<node ID>/`: `log` (`/admin/log`), `config` (`/admin/config`) and
`debug` (`/debug`, always `write`), `read` to look and `write` to act
(`api::access::needed`, which takes the node's ID). So `_continuum/local/<node ID>` is
everything on one node, `_continuum/local/*/log` the log filter on every node, and
holders of `_continuum` keep it all.
Nothing is stored there: the store refuses writes under `_continuum/local`, and the
rules refuse a change touching it, as for epoch metadata. The quarantine stays with
`_continuum`: its entries are everyone's writes, which a node-local capability
mustn't show.
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
A capability's path covered by prefix only, so what is the same at every key of a map
(each node's log filter, say, at `_continuum/local/<id>/log` from the next patch on)
took a capability per key. Now a `*` segment stands for any one key, as it already does
in signed paths and sealed places (`envelope`, `sealed`, which now share one `agree`):
* covering (`Capability::covers`): the pattern no longer than the path, and agreeing
with it segment by segment;
* listings (`reaches`): pattern and path agreeing on the segments both have, so every
key on the way shows, nothing past where the pattern differs;
* narrowing (`Capability::holds`, for a token's limit narrowed again and a refresh's):
at or below, a `*` only under a `*`: a limit naming one key never widens to all.
A `*` first is outside protocol data, so it never covers `_continuum`, as `[]` doesn't.
A key that is `*` itself can't be named alone. Every node judges changes by
capabilities, so all must read them alike: they run the same version.
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
The node's keepalives take turns between the two sides (`PeerState::keepalive_due_at`),
but the paths layer pinged every path in use every `probe_every` (the keepalive interval,
30 s at least) whatever the peer did: a PING from the peer over the path didn't count.
So both sides pinged each path every interval, often within a second of each other
(seen on the dashboard: a PING answered, and our own PING right after it).
Now a PING from the peer over a path is noted (`PathState::peer_probe`), and a path is
pinged again (`rested`) an interval after the peer's PING, if that came last; after
ours, an interval and a half while the peer pings too (its PING within two intervals
before ours), so its turn comes first; else an interval. One exchange per interval
goes over a path, as the measurements' "two keepalives per neighbour per interval"
says, and each side measures every other interval; a shorter `--keepalive` measures
more often. A path whose way out broke while the peer's PINGs still come in is found by
our own PINGs half as often; a failed send or unanswered traffic (`suspect`) still
pings at once.
|
| |
|
|
|
|
|
|
|
|
|
|
|
| |
Requests weren't logged at all, short of what a handler logs itself. Now a middleware,
outermost on the API (and on a relay-only node's), logs a line per request as web
servers do: method and path (with the query), status, the milliseconds until the answer
began (a stream goes on after), where it came from (the address, or `unix`), and the
user (`-` where none was authenticated). The user comes back on the response:
`basic_auth` puts it there, for a token as for a password or a trusted proxy's header.
At debug, under the target `http` (`RUST_LOG=info,http=debug`, or `PUT /admin/log`):
an open dashboard alone would fill the log at info. With the target off, the request
passes untouched.
|
| |
|
|
|
|
|
|
|
| |
The remake dialog set the kind, its input, the settle time and the switch delay, but not
the replay window or the archive retention. Now each has two fields, epochs and an age,
their defaults as placeholders: both empty sends nothing (the node's defaults), one
given sends that limit alone, as the API reads it; in the body, or for an import in the
query (`replayWindow`, `archiveRetention`). Epochs that aren't a whole number are
refused in the page.
|
| |
|
|
|
|
|
| |
An import uploaded as `application/octet-stream` takes `rules`, `baseHeads`, `settle`
and `switchDelay` in the query, but not the retentions: for those the document had to
go base64-encoded in the JSON form. Now `replayWindow` and `archiveRetention` go in the
query too, as JSON like the rules (`{"epochs":1,"age":"7d"}`, as in the body).
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Every snapshot carried the metadata of all earlier epochs (`_continuum/epochs`)
forward, so it grew by an entry, a clock per actor among it, with every remake. Only a
rejoining node reads an earlier entry: a node on epoch P reads P+1's (its clock, when
it was proposed, its replay window) and every one up to the swarm's epoch (kind,
translation, whether it was strict). It replays only while P+1's replay window covers
it; beyond, its writes are quarantined whatever the entries say.
So a snapshot now leaves out the generations older than the oldest whose replay window
still covers a node on the epoch before it (`snapshot::left_out`), judged by the
remake's `proposedAt`: the same on every node, and earlier than any rejoining node's
clock, so nothing a node can still replay through goes. Metadata that doesn't read
stays, with all after it.
A strict remake's metadata stays for good: by it a rejoining node, however far behind,
tells a strict remake of a swarm strict already (`catchup::prepare`, which reads the
generations it finds). There is at most one, short of such a remake. No new key and no
change to the rejoin: a node beyond the window that finds some entries and not others
quarantines with "cannot translate" rather than "outside the replay window".
|
| |
|
|
|
|
|
|
| |
A key's line was stored as the key alone (`ssh-ed25519 <base64>`), whatever comment it
came with, both from `PUT /auth/keys/<name>` and from `--bootstrap-keys`. Now its
comment is kept after it (`PublicKey::kept_line`), the spaces between its words made
one. The comment checks nothing: every use of a key parses the line, and the parse
reads the key alone.
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
The credentials dialog listed a user's keys whole, to add one or
remove one but the last, and any failure replaced the dialog with the
error, the form lost.
* A key is shown short: its type, the first and last four characters
of what tells it apart (an Ed25519 key's base64 after the part all of
them share), its comment; the whole line in its tooltip.
* A key can be edited, its name and its line: saved under the name, a
new name written first and the old one removed.
* A user with a password on this node (`--htpasswd`) may remove its
last key: it logs in by the password. GET /auth/keys says so
(`password`), and DELETE /auth/keys/{id} takes it then; the rules no
longer keep the last key on every node (no other node can tell who
has a password where; under a strict policy the signed entry keeps
one still).
* What fails says why beside where it was asked (the add form, the
edited row, the removed one), which stays as it was, to be set right
and tried again.
|
| |
|
|
|
|
|
|
|
|
| |
A right press on a node started dragging it, as any press did: the map
took hold of the pointer, so the right-click went to the map and the
node's menu didn't open, and letting go put the node in focus. The
right button now opens the node's menu alone. Its entry for the node's
dashboard is `go to its dashboard`, greyed where the node publishes
none (and on this node's own), its tooltip saying where it goes or why
not.
|
| | |
|
| |
|
|
|
|
|
|
| |
With the entries' menus (1238), editing had three ways in: the
double-click, the `+` on a map's or list's row, and the menu. The
double-click and the `+` go; the menu does all of it. The title,
which stands for the root (its row isn't shown), now takes the
right-click too: its menu adds a key at the top level.
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
| |
The document box edited an entry by a double-click and added to a map
or list by the `+` on its row; a key couldn't be renamed, and a list's
`+` always asked for an index.
A right-click on an entry, or a finger held on it half a second, opens
its menu: `Edit value` (as the double-click), `Rename key` for a map's
key, and on a map `Add a key`, on a list `Append an item` or `Insert
an item at…`. Renaming reads the value, writes it under the new key
and deletes the old one (two writes, the value as JSON; a key that
exists already isn't overwritten). Inside a field being edited, and
where an entry offers nothing, the browser's own menu shows.
|
| |
|
|
|
|
|
|
|
|
|
|
|
| |
A stream's lag in the tooltip of `behind: …` was one figure, from when
a node logged a message to when the page had it: a second and a half
for a node a LAN away said nothing of where it went.
The node now stamps each `messages` event with when it sent it
(`{"sent": …, "messages": […]}`), and the tooltip splits each stream's
lag: how long its batches waited in the node, from logging the first
message to sending the batch (the node's own clock: exact), and how
long they took on the way, from sending to arriving. This node's own
stream is always listed, beside the three slowest.
|
| |
|
|
|
|
|
|
|
|
|
|
| |
The console appended messages as their streams brought them, a
subscribed node's later than this node's by its path, each line with
its node's own clock's time: lines of one moment came apart, and their
times read out of order.
Each message is now listed by when its node logged it, on this node's
clock, as the map runs them (`aheadOf`): one a slower stream brought
late goes in among those of its moment. The time shown is on this
node's clock; a subscribed node's own clock's time is in its tooltip.
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Each node's message stream comes its own way to the page, a subscribed
node's over the network to it: messages logged 16 ms apart on three
nodes ran over a second apart, as their streams arrived.
Messages now run when they went, on this node's clock: each at the
time its node logged it, taken back by how far that node's clock is
ahead of this node's (the map's clock chain, `aheadOf`), a while
behind real time (`replayDelay`): the most any stream lagged in the
last 30 s and a tenth of a second more, shown after the sliders
(`behind: …`), each stream's lag behind the fastest in its tooltip.
The browser's clock is against all streams alike and drops out. A
message known only from its receiver (its sender not subscribed) is
placed at its arrival, starting a crossing before. Messages lagging
more than 10 s (a gap) don't run, nor hold the map behind.
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
The node sent its messages to the page every 2 s (`messages` events,
this node's and a subscribed node's), and this node's ran on the map
from its counts, which come with the state: the map showed a message 2
to 4 s after it went, the messages of each 2 s spread evenly over the
next.
The message log now tells those waiting of each message logged
(MessageLog::logged_after), and both streams send what was logged a
tenth of a second after the first of it (MESSAGES_GATHERED): a busy
node about ten events a second, a quiet one each message soon after it
went. The page runs messages as their streams bring them, this node's
from its own log too (with the addresses its PINGs went to), as far
apart as they were sent; more than six on a link and way at once are
shared out by kind over the moment, and after a gap only the last 2 s
run. The counts still give the rates and what was delivered.
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
A message between this node and a subscribed one ran from this node's
counts both ways: one the subscribed node sent ran once this node had
counted it received, as it arrived. Each message now runs from its
sender where the page reads the sender (this node's counts, a
subscribed node's messages), so it runs the way it went, also where it
never arrived; only one from a node the page doesn't read runs as it
arrived. The console lists them as before.
Where the page reads both ends of a link, what one sent the other
should count received: a message sent and not received within 5 s
makes the link breathe red once (`breatheLink`), its tooltip saying
how many. Kinds the paths and identity layers count under addresses
are left out, as this node's counts leave them out; a subscribed
node's tick that came short (the latest 500 of more) starts the count
afresh, so what the page didn't see isn't taken as lost.
|
| |
|
|
|
|
|
|
|
|
|
|
| |
Dots and flashes ran straight from sender to receiver, whatever line
the message took: a PING over another transport, drawn as a curved
line beside the one in use, crossed the map off its line.
Each message now runs along the line it went over, as drawn (`route`:
the link's quadratic curve, its control point where renderGraph put
it): a message naming the address it went to (the paths layer's PINGs
and PONGs) over that address's transport's line, any other over the
line in use; a flash's streak bends with it.
|
| |
|
|
|
|
|
|
| |
On a dynamic map, links pull toward their lengths against the nodes
pushing each other apart, at one strength. A `pull` slider before the
layout's toggle, shown while dynamic, sets it from a quarter to four
times as hard (2 to its value, once at first), kept in the browser;
the map moves to the new balance smoothly.
|
| |
|
|
|
|
|
|
|
|
| |
A link was dotted where an end reported no links. Members used to
report their links to every dashboard (LINKS); the page now reads a
member's links only while subscribed to it, so a dotted link said no
more than that its other end isn't subscribed, which the node shows
already (hollow), and most links were dotted. The style and its
legend entry go; the link's tooltip says when one end alone reports
it.
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
A sync message carried the signatures its peer wasn't sent yet, by a
cursor per peer. Changes sent again, as after a message taken as
delivered and lost on the way, went without them: the peer was past
them. Under a policy requiring signatures they waited until the
writer's next signed change, or the next session.
A message now carries the signatures of the changes it carries
(Signatures::for_changes), whatever the peer was sent before: changes
sent again arrive vouched for as the first time. Past 256 (a message
carrying the batches of many writers), each writer's latest instead,
as a delta carries. The cursor goes, and with it resetting it after a
failed send or a new session.
|
| |
|
|
|
|
|
|
|
| |
api.md's Signed changes says what sent, all and none do, what each
costs, and which to choose: sent for nearly every swarm, all where each
change should carry its own proof, none to save signatures where only
protocol data and listed paths need them; and that signatures travel
with their changes, a message carrying none taking none. access.md no
longer says signatures come on their own within a second.
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
| |
Signatures sent on their own (SIGNATURES, every second, to each peer
not sent them yet) flooded every link: a signature made after its
change left, or one that came with changes, went to every neighbour,
whether or not it got the change along with its signature elsewhere.
With every change of a node signed as it leaves (1217) and signatures
going where their changes go (1218), none is left for them to carry:
the message, its task and its traffic counter go (wire version 46).
Signatures waiting for their changes are offered as before when sync
messages bring them, and the changes waiting for signatures judged
again with the next batch.
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
A node signed its changes no rule signs a while after them (--sign-every
1s): the change left at once, unsigned, and its signature followed on
its own a second later, to every neighbour, over every link. A few
writes made a cascade of SIGNATURES. Under a policy requiring
signatures it signed every change as it was made instead.
A node now signs its latest change, if no signature vouches for it, as
a sync message takes it (--sign-every sent, the default): the
signature leaves in the same message, and goes where the change goes.
One signature per message, however many changes it carries, each
vouching for those before it: a slow writer's every change is signed,
a fast writer's batches are, at no threshold to pick. A policy
requiring signatures signs what is sent too, even with `none`. `all`
still signs every change as it is made; changes to protocol data,
listed paths and ?signed=true writes are signed as they are made, as
before. The interval, and the task signing on it, go.
|
| |
|
|
|
|
| |
Review 3: since 1184 a proposal is written as the user who asks, so
under a policy the node doesn't know it answers 409 like every user
write. remakes.md says so.
|
| |
|
|
|
|
|
|
|
| |
Review 3, R5: a node given its root (--genesis-root) refused a peer
naming another only change by change while its document was empty,
each with an entry among the refused; the handshake went through, as
there was no root of its own yet to compare with. The root given is
now compared too: a peer naming another is refused at the handshake,
logged once, as one whose document starts from another root.
|
| |
|
|
|
|
|
|
|
|
|
|
| |
Review 3, R5: --refused-writes-per-minute bounds each user, but the
cost it guards, a document copy under the write lock, is the node's:
ten users, or ten accounts of one, are ten times the budget.
The node now takes six times a user's limit within a minute in all
(refused::NODE_FACTOR); past it, every user with a write refused
within the minute is answered 429 untried, until the oldest leave it.
A user none of whose writes was refused writes on, so honest writers
aren't held up by others' refusals.
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
| |
Review 3, R4: whether a batch was judged followed the receiving
document's policy now (and whether the batch writes one). After the
policy was removed, or replaced by a version the node doesn't know,
changes written under it and received late went in unjudged, while
nodes that had them earlier judged them; REFUSALS gave up likewise.
Screening now remembers that the document held a policy, known or not
(Screening::policed), and judges what it receives while it does, each
change by the policy at its dependencies; REFUSALS judges then too
(not under a policy the node doesn't know, as before). Not across a
restart, which access.md says.
|
| |
|
|
|
|
|
|
| |
Review 3, R3, the replace side: a replace or import remake writes its
content as it is, at signed places too, so under a strict policy a
member majority can put plain values where envelopes were. access.md
says that remakes are outside the rule; keeping envelopes at signed
places in every kind is postponed to future-work.md.
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Review 3, R3: 1186 kept honest nodes from proposing or voting for a
strict remake of a swarm strict already, but a majority of members,
each with keys of its own, could still commit one, and switch and
catchup followed it without asking what kind of remake made the
epoch: the registry, policy and bans of a strict swarm changing by a
member majority, the honest minority following.
A node now refuses to switch to such an epoch (its document's policy
strict, the remake's kind strict), and catchup refuses to rejoin one
whose chain of epochs holds a strict remake after the swarm was strict
(as the old document is, or once a remake since made it so). access.md
says so.
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Review 3, R2: the root pin was set by the first HELLO_REPLY, and the
first sync, a round trip later, took that root; a peer naming another
afterwards was refused as another history. So the seed that answered
first decided, not "none while they disagree", and a malicious seed
makes itself fastest.
A node joining with an empty document of the first epoch now takes no
root (RootPin::Awaiting, roots left for later) until each of its seeds
named one, or ROOT_WAIT (10 s) passed since the first did; then the
one they name, none while they differ, as before. With --genesis-root
it takes that one alone, at once. access.md, the option's help and
example.toml say so; switching roots later, while the joiner holds
nothing of its own, is postponed.
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Review 3, R1: what a node received was vouched for only while its own
document's policy required signatures. A joiner's first sync, and
every node's message bringing `required`, went in by the rules alone:
a malicious first peer could plant changes under other members' actors
at their next seqs, their real ones then failing as duplicates for
good.
The signatures a message carries are always offered now, and the
screen always has them: a change waits for one where the policy at the
heads it is judged at requires it (rules::Reads::required, read once
per heads), whatever the receiving document holds. Changes written
under a policy not requiring signatures wait for none, as on every
other node; the check for forged changes runs always. Changes waiting
for a signature are judged again whenever signatures come.
The test of signatures required now takes the history into an empty
document; a new one takes changes written under a policy not requiring
them without any.
|
| |
|
|
|
|
|
|
|
|
| |
A write every node would refuse is rolled back by forking the document
at the heads before it, under the write lock, at a cost that grows with
the document's history. A user could ask for such writes in a loop.
Count each user's refused writes over the last minute. Past
--refused-writes-per-minute (default 10), its writes are answered 429
without being tried, until the oldest refusal leaves the minute.
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Review 2, N7, the care it raised: a joining node's empty document took
as the first epoch's root whatever dependency-less change came first,
from whichever peer it joined through; the first epoch's ID is all
zeros, so nothing pinned it. A remake's epoch has its ID, the snapshot
change's hash, which the rules hold its root to.
* Every node names its epoch's root in HELLO and HELLO_REPLY (the
epoch's ID for a remake's, its genesis change for the first; wire
version 45). A node whose document starts from another root of the
same epoch refuses the handshake, logging it: another history under
the swarm's ID, a forged root or a genesis started again.
* A joining node with an empty document takes as its root only the one
it was given (--genesis-root, new), else the one its peers name, and
none while they name different ones (access::screen::RootPin, kept
in Screening across documents): one peer hands it another root only
if it is all the joiner hears, or never, with --genesis-root. A root
refused is listed with the refused, what builds on it with it.
* The genesis node logs the root (`The swarm's root`); GET
/admin/epochs shows it (`root`).
A genesis started again under a remembered swarm ID used to merge with
the nodes of it still running; they now refuse it (another root), and
the docs say to join them with seeds instead.
The epoch's identity stays as it is (generation 0 all zeros): records
and epoch metadata name it so, and every comparison of epochs with
them.
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Review 2, N7: a node screened received changes only while its own
document had a policy it enforces. A joiner's empty document has none,
so its first sync, as much history as the node it joins through chose,
was applied unjudged, and under signatures required unvouched; the same
for any node while the policy's own change arrived with later ones.
Sync messages, deltas and downloaded changes are now always screened,
each change judged by the policy at its dependencies (1135), so the
verdicts are the swarm's. While neither the document nor the changes
hold a policy (none touches a "policy" key: inspect::touch_key), only
an epoch's root is judged and nothing is placed: bench_screening, after
1188, measured placing everything at about as much again as applying
(564 ms + 543 ms for 100 000 changes), judging by the rules at about
five times (+2.5 s), which a document with a policy pays anyway.
access.md no longer lists the joiner's first sync, nor nodes agreeing
only as the policy reaches them (judged at the dependencies, they
agree), as left to do; changes::load_delta is for tests now.
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
| |
Review 2, N9: the API checked a write's capability on the path asked
(servers/list/3/name), the rules on the list it is in (servers/list):
a user whose capability lay inside a list had every such write pass the
API and refused by the node's own judgement (403, after rolling the
document back).
inspect::placed places a write's path in the document by the rule the
inspector places operations by (Place::child: cut after the first list
or text on the way), and the store's writes check capabilities there.
One rule for both; a test checks placed against the paths the inspector
finds in the change such a write makes.
|
| |
|
|
|
|
|
|
|
|
|
| |
Review 2, N7, measured first: a node whose document has no policy (a
joiner's empty one) applies what it receives unjudged. Screening it too
would cost finding every change's paths and judging it. bench_screening
takes 1 000, 10 000 and 100 000 changes into an empty document, as one
delta brings them: unjudged, as Automerge applies them, against
screened (access::screen::apply), with no policy in them (each passes,
read at its dependencies, after its paths were found) and with the
policy their first change wrote (each judged by the rules).
|
| |
|
|
|
|
|
|
|
|
|
|
|
| |
Review 2, N2: nothing checked that a strict remake's base wasn't strict
already. In a strict swarm the registry, the policy and the bans change
only as a user who may change them signs; a strict remake committed by
a majority of members, each with keys of its own, could drop any of
them (unban, remove managers, empty userSigned and the sealed
declarations), and the honest minority would follow.
A node now takes a strict remake (proposing or voting) only in a swarm
not strict yet: turning strict is done once, by a remake; after it, by
signatures.
|
| |
|
|
|
|
|
|
|
|
|
|
| |
Review 2, N8: a node with one neighbour (a swarm of two, a leaf) never
hears a second member name the change: its big drop was held for good,
an error logged at every REFUSALS. It now waits only where it has
another neighbour than the one telling it; with one, one word is all
there is.
access.md also says what the wait is: it guards against one peer, not
two members saying the same (any node with the swarm key registers),
and never makes a node drop what its own rules take.
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Review 2, N3: 1150 let any member write an abort decision once a remake
had expired a grace period ago, judged by the change's own time, the
writer's word. A member could race the coordinator's legitimate commit
with an abort dated as it liked; both taken, Automerge shows
one, and nodes that acted on the commit split from those that didn't.
The two writers of such aborts are operator requests whose endpoints
already need write on _continuum/remakes: aborting an expired remake
(DELETE /admin/remakes/{id}) and superseding one (POST
/admin/remakes?force=supersede). Both now commit as the user who asks
(the proposal and its vote with them, for a supersede), which the rules
take as one who may write the decision. The clause judging an abort by
its time goes, and with it the time the rules carried. An expired
remake is still aborted through any member, by an operator.
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
| |
Review 2, N1: a strict manifest wasn't bound to its base. Any member
could propose an old, legitimately signed one again at the current
heads; nodes would verify it and vote, and every value at a signed
place changed or added since (a ban, a new user, a rotated key, a
changed userSigned or sealed declaration) would be deleted, nothing
after the base being left to replay.
The manifest now names the epoch and the heads it was made at, and a
node takes it (proposing or voting) only for a remake of that epoch
with exactly that base. GET /admin/strict-manifest makes it so;
continuum-signed strict proposes at those heads already.
|
| |
|
|
|
|
|
|
|
| |
Review 2, N11: a value at _continuum/logouts/<name> that isn't a time
(read as damaged: every token revoked), or one far ahead, refuses every
token of the user, new ones too, until it is replaced. Any member can
write it (finding 10). The rules can't see the value (only that it is
written), so access.md says it, and how it is set right: the user logs
out again by password, or one who manages users revokes its tokens.
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
A strict remake takes a majority of members whose --strict-keys take
its manifest's signer: no node votes for one without, so
--follow-strict-remakes (1173) goes. A node that doesn't vote still
switches once the others commit, as with any remake.
So that an operator can tell why one fails: a strict remake aborted
for want of a majority (the voting period over, or a majority out of
reach) says in its reason (GET /admin/remakes/{id}, and the
coordinator's log) which members didn't vote, and that a member votes
for one only with its manifest signed by a key of its --strict-keys.
Each of those logs why itself (`Not voting for a strict remake`).
The sim a_strict_remake_too_few_take: only a takes the manifest; the
remake aborts, naming b and c.
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Until a swarm is strict, the user a change names is its node's word, so
any member could have planted registry entries, policy, bans or
user-signed values as any user, plain or signed by a key it put there
itself; writing the strict policy trusted all of it as it was.
A remake of a new kind, `strict`, turns a swarm strict instead:
* Its content is a manifest (access::strict): the swarm, the strict
policy, and each value kept at the places values are signed, by the
SHA-256 of its JSON, signed by an operator (ssh-keygen -Y sign -n
continuum-strict).
* Every node builds the next epoch as compaction does, keeping at the
signed places only the values the manifest names, as they were at
the base, and writing its policy (RemakeKind::protocol, the one kind
that changes protocol data). Writes after the base reach the new
epoch only by replays, judged by the strict policy.
* A node votes for one only with its manifest signed by one of
--strict-keys, keys given out of the document; without any, for none
(logged once, `Not voting for a strict remake`), unless
--follow-strict-remakes. A coordinator proposes only what it would
vote for (403). Once a majority commits one, every node switches, as
with any remake: refusing to switch would only cut a node off.
* GET /admin/strict-manifest gives the manifest keeping everything held
now, the values for review and the heads they are at;
continuum-signed strict shows them, signs once confirmed and proposes
at those heads.
The sim a_swarm_turning_strict_through_a_remake: a member plants a
registry entry as a manager; the operator's manifest leaves it out; a
and b, given the key, commit; c, without, follows; the entry is gone
and the policy strict everywhere.
|
