perf(sync): the fork's signed commit with fast merged deletes and batched inserts

The offline sync package's shared part now comes from the fork's branch
halkyon/batch-inserts (949a664, signed), adding two changes to the pin.
A merge only deleting rows skips the projection pass when every visible
row below them goes with them (a card deleted on one device, in a
library of 8,000 cards: 5.4 s for the server to merge, now about 35 ms;
4 s on each other device, now about 0.1 s). And a merge the pass leaves
out writes its new rows up to 1,000 to a statement, its proof also
taking nested studysets (a first push of that library: 35 to 41 s on
the server, now about 5; a device receiving it, 60 s, now 44).

A new sync test deletes a card on one device while another rates it:
the merged delete declines the shortcut there, and both devices end
without the card or its rating.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Signed-off-by: Tizian 「ティツィアーン」 <raisondetredev@pm.me>
This commit is contained in:
Tizian 「ティツィアーン」 2026-10-17 22:24:30 +02:00
commit fd69719652
Signed by: raisondetredev
GPG key ID: 07F6FCF405488B64

View file

@ -17,23 +17,26 @@ dev_dependencies:
leancode_lint: ^27.0.0
dependency_overrides:
# The offline sync package's shared part from our fork: v0.0.10 with five
# The offline sync package's shared part from our fork: v0.0.10 with seven
# changes. Byte payloads compared as bytes: v0.0.10 compares them byte by byte
# through a UUID parse that throws for each byte, so a sync touching a few thousand
# cards kept the server busy for minutes, answering nobody, while its client timed
# out and sent everything again (2026-10-12). onInboundMerge, called only when
# changes from the remote peer were merged: onMergeSuccess fires for a space this
# device only pushed to, so every repository read everything again after each
# save's sync (2026-10-17). A local delete, and a merge only inserting new rows,
# skip the foreign key projection pass when they can't change it: the pass loaded
# nearly every synced row, 8 s for one delete and 9.5 s for each rating another
# device made, in a library of 8,000 cards (2026-10-17). And projections are
# save's sync (2026-10-17). A local delete, and a merge only inserting new rows
# or only deleting rows, skip the foreign key projection pass when they can't
# change it: the pass loaded nearly every synced row, 8 s for one delete and 9.5 s
# for each rating another device made, in a library of 8,000 cards (2026-10-17).
# Such a merge writes its new rows up to 1,000 to a statement: a first push of
# that library took the server 35 s, now 5 (2026-10-17). And projections are
# rebuilt once after the synced tables changed, not on every session and merge
# (upstream #150). Pinned to a signed commit of the fork's branch
# halkyon/fast-merges (on halkyon/fast-deletes, on halkyon/inbound-merges, on
# halkyon/byte-payloads); drop each change once a release of the package has it.
# halkyon/batch-inserts (on halkyon/fast-merge-deletes, on halkyon/fast-merges,
# on halkyon/fast-deletes, on halkyon/inbound-merges, on halkyon/byte-payloads);
# drop each change once a release of the package has it.
serverpod_offline_sync:
git:
url: https://code.halkyon.app/project-halkyon/serverpod-offline-sync.git
path: packages/serverpod_offline_sync
ref: f062725a16fffa1be8cf9141cdbfefae209598ea
ref: 949a6645e08af27bb61c96d7ddc2b6d145fed554