Jade Ellis
403a90c64f
fix: Correct relative script location
2026-08-22 22:43:41 +01:00
timedout
8a7324216e
fix: Invert inverted power level comparisons
2026-08-22 22:08:10 +01:00
timedout
4d487afe26
fix: SEC30
...
Reviewed-By: Ginger <ginger@gingershaped.computer >
Reviewed-By: Jade <jade@ellis.link >
Reviewed-By: Jacob Taylor <jacob@explodie.org >
2026-08-22 20:48:37 +01:00
timedout
103dafd529
fix: Address review comments
2026-08-22 20:14:43 +01:00
timedout
a61cc2ac50
fix: Return correct new extremities set
2026-08-22 20:14:43 +01:00
timedout
2ee33f091f
fix: Don't only include referenced extremities
2026-08-22 20:14:43 +01:00
timedout
1cc6fb1130
chore: Add newsfrag and re-run cargo fmt
2026-08-22 20:14:43 +01:00
timedout
681daed5e8
fix: Mark prev events as referenced as we come across them
2026-08-22 20:14:43 +01:00
timedout
8d95163da4
perf: Remove unnecessary retain
...
The remove() at the start of the function already does this
2026-08-22 20:14:42 +01:00
timedout
0608f674a3
fix: Properly track orphaned extremities
2026-08-22 20:14:42 +01:00
Erwan Leboucher
df95abf5ab
fix(federation): Queue PDUs during remote joins
2026-08-22 18:29:54 +00:00
Jacob Taylor
43f117de46
fix: Update sender_retry_backoff_limit to match docs and actually be 1 week
2026-08-20 07:41:06 -07:00
Erwan Leboucher
82149397b0
fix(admin): Complete MSC4323 suspend and lock endpoints
2026-08-20 00:58:30 +00:00
timedout
f1a77e0404
fix: Address review comments
2026-08-20 00:57:11 +00:00
timedout
9af60bcf62
feat: Lower default backoff base for a gentler backoff curve
2026-08-20 00:57:11 +00:00
timedout
b73b3532f2
fix: Don't set retrying status when retrying is illegal
2026-08-20 00:57:11 +00:00
timedout
7352edc6e4
fix: Treat any non-ok response as a server error in sender
...
Otherwise we never back off sending failed requests
2026-08-20 00:57:11 +00:00
timedout
a916fdf09a
fix: Mark even more error codes as stale based on observed data
2026-08-20 00:57:11 +00:00
timedout
768933b025
fix: Don't stop tracking servers when there's nothing more to send despite being inflight
2026-08-20 00:57:11 +00:00
timedout
46f0483395
fix: Correctly insert running status only if the sender should attempt to send
2026-08-20 00:57:11 +00:00
timedout
dde1f12814
style: Reformat
2026-08-20 00:57:11 +00:00
timedout
22f9b9b91d
fix(sender): retry_after duration can be negative
2026-08-20 00:57:11 +00:00
timedout
f7773760a6
fix(sender): Don't attempt to send concurrent transactions to a remote
2026-08-20 00:57:11 +00:00
timedout
6104c68768
style: Centralise decision for stale marker
2026-08-20 00:57:11 +00:00
timedout
9ed8c230fa
feat: Update servers-in-backoff command to annotate stale info
2026-08-20 00:57:11 +00:00
timedout
ef5bf7c977
feat: Clear destinations cache when a remote is potentially stale
2026-08-20 00:57:11 +00:00
timedout
53db584741
feat: Add debug command to extract backoff info
2026-08-20 00:57:11 +00:00
timedout
2d713c1631
feat: Use an explicit backoff base value
2026-08-20 00:57:11 +00:00
timedout
79c9f7da30
chore: Tidy up backoff helper code
2026-08-20 00:57:11 +00:00
timedout
bc1826de66
feat: Hook transaction senders into health tracking
2026-08-20 00:57:11 +00:00
timedout
98466e5243
style: Remove unsafe arithmetic
2026-08-20 00:57:11 +00:00
timedout
8c1a84e907
fix: Don't honour unhealthy hits unless the block period has already expired
2026-08-20 00:57:11 +00:00
timedout
8f66cbb6b9
feat: Clear remote health tracking map when cache is cleared
2026-08-20 00:57:11 +00:00
timedout
e1866df682
feat: Trace why remotes get health marked
2026-08-20 00:57:11 +00:00
timedout
213bf6fb70
feat: Refuse to send requests to unhealthy servers
...
Also marks them as online if a request succeeds
2026-08-20 00:57:11 +00:00
timedout
785db0bdba
chore: Move health tracking to federation service
...
Otherwise federation would have to depend on sender
2026-08-20 00:57:11 +00:00
timedout
6ae66eed1b
feat: Mark servers as online/offline based on transaction responses
2026-08-20 00:57:11 +00:00
timedout
077338792f
feat: Mark servers as healthy when they successfully make an inbound request
2026-08-20 00:57:11 +00:00
timedout
0c8b51c0f4
feat: Add server health tracking
2026-08-20 00:57:11 +00:00
Erwan Leboucher
1ecfb21aea
fix(spaces): Correct hierarchy traversal
2026-08-19 03:33:41 +00:00
Erwan Leboucher
39b08f73f7
fix(state-res): Derive sender power from the create event when sorting
2026-08-18 20:22:38 +00:00
Erwan Leboucher
3efb66c76c
fix: Do not make late-arriving events forward extremities
2026-08-18 16:55:47 +00:00
timedout
7b3e3a26bb
fix: Don't panic when receiving PDUs with illegal auth events
...
Apparently some servers do this??
2026-08-18 17:50:18 +01:00
Erwan Leboucher
1cdf369a80
fix(sync): Populate sliding sync num_live
2026-08-18 15:44:52 +00:00
Erwan Leboucher
1d655fda53
fix(timeline): Keep extremities a local event did not reference
2026-08-18 15:38:43 +00:00
ginger
b502f74c34
fix: Remove mystery additional admin room notice
2026-08-17 17:55:59 +00:00
timedout
a148174c59
style: Rephrase error message
2026-08-15 16:10:39 +00:00
timedout
2dc6051d67
fix(federation): Make transactions infallible
...
Previously, a transaction panicking in the handle() call would cause the
transaction to be indefinitely logged as "running" in the transaction
handler. This means subsequent transactions from the sending server
would be met with the "You're still sending me a txn!" 429 response,
forever, effectively causing defederation between the two servers, until
continuwuity is fully restarted.
I could just fix the known panics in the handle function and internal
calls, however there's so many subsystems that could reasonably panic
that it is just better to have a safer approach and ensure that the
handle call is infallible.
2026-08-15 16:10:39 +00:00
Erwan Leboucher
918070c512
fix(spaces): Expose restricted children over federation
2026-08-14 12:51:09 +02:00
timedout
08d3d8445f
chore: Merge branch 'nex/feat/registration-messages'
2026-08-12 14:25:15 +01:00
Jade Ellis