Remove invalid subset loading check

[KyleAMathews]

Removed the check that prevented requestSnapshot from being called in full mode. It's valid to load subsets while loading the full log, which is necessary for LiveQuery joins with collections using progressive syncMode.

Fixes the error: "Snapshot requests are not supported in full mode, as the consumer is guaranteed to observe all data"

[codecov[bot]]

Codecov Report

:white_check_mark: All modified and coverable lines are covered by tests. :white_check_mark: Project coverage is 69.54%. Comparing base (f41db22) to head (cb91c91). :warning: Report is 21 commits behind head on main.

Additional details and impacted files

@@            Coverage Diff             @@
##             main    #3436      +/-   ##
==========================================
+ Coverage   69.45%   69.54%   +0.08%     
==========================================
  Files         182      182              
  Lines        9787     9786       -1     
  Branches      356      358       +2     
==========================================
+ Hits         6798     6806       +8     
+ Misses       2987     2978       -9     
  Partials        2        2              

Flag	Coverage Δ
elixir	66.51% <ø> (+0.09%)	:arrow_up:
elixir-client	74.47% <ø> (+0.52%)	:arrow_up:
packages/experimental	87.73% <ø> (ø)
packages/react-hooks	86.48% <ø> (ø)
packages/typescript-client	93.76% <ø> (+0.08%)	:arrow_up:
packages/y-electric	55.12% <ø> (ø)
postgres-140000	65.28% <ø> (-0.21%)	:arrow_down:
postgres-150000	65.52% <ø> (+0.21%)	:arrow_up:
postgres-170000	65.35% <ø> (-0.05%)	:arrow_down:
postgres-180000	65.45% <ø> (?)
sync-service	65.72% <ø> (+0.05%)	:arrow_up:
typescript	87.18% <ø> (+0.05%)	:arrow_up:
unit-tests	69.54% <ø> (+0.08%)	:arrow_up:

Flags with carried forward coverage won't be shown. Click here to find out more.

:umbrella: View full report in Codecov by Sentry.
:loudspeaker: Have feedback on the report? Share it here.

:rocket: New features to boost your workflow:

• :package: JS Bundle Analysis: Save yourself from yourself by tracking and limiting bundle sizes in JS merges.

[pkg-pr-new[bot]]

Open in StackBlitz

npm i https://pkg.pr.new/@electric-sql/react@3436

npm i https://pkg.pr.new/@electric-sql/client@3436

npm i https://pkg.pr.new/@electric-sql/y-electric@3436

commit: cb91c91

[kevin-dp]

Should we have an e2e test for this?

[icehaunter]

Out of curiosity, what's the use of a snapshot in a full mode, given the client already holds all the same data?

[KyleAMathews]

@icehaunter for the "progressive mode" https://tanstack.com/blog/tanstack-db-0.5-query-driven-sync#progressive-sync-fast-initial-paint--instant-client-queries

@kevin-dp you mean in the tanstack db repo?

[samwillis]

@KyleAMathews This will not fix the issue with Tanstack DB progressive mode, we need to fix it on the DB side rather than on the Electric Client side.

The electric client ensures that the results of a requestSnapshot call are injected into the shape stream at the point where they are causally correct - this is important for on-demand mode in DB to insure that live updates are applied on top of the snapshot and it does not overwrite newer changes. With this fix the snapshots will still not be injected until the end of the stream, not optimistically as we would like.

The intention of progressive mode is that while the initial sync is happening we use and show snapshots. We need to have a discussion on this mode as currently nothing is applied to the collection until the first up-to-date.

[icehaunter]

@icehaunter for the "progressive mode" https://tanstack.com/blog/tanstack-db-0.5-query-driven-sync#progressive-sync-fast-initial-paint--instant-client-queries

@kevin-dp you mean in the tanstack db repo?

No, I got the general idea, I'm just not sure why additional snapshots are needed on top of already full data. I thought the idea for progressive mode was to only use these subset snapshots, otherwise it seems to make little sense if the client already has all the data. Am I missing something?

[samwillis]

@icehaunter

I'm just not sure why additional snapshots are needed on top of already full data

DB progressive mode is supposed to show the snapshots before we load all the data. on-demand is working with the new scheme perfectly.