feat(lanes): allow --squash on diverged components by davidfirst · Pull Request #10387 · teambit/bit

davidfirst · 2026-05-20T20:18:56Z

Why

bit lane merge <other-lane> --squash previously threw when components were diverged in history, forcing users to either do a two-way merge dance or fall back to --no-squash. With --no-squash on large lanes hosted in different scopes, exporting after the merge pulls the other scope's full snap history into the merging scope, which can blow up memory on both client and remote.

What

Diverged components under --squash now produce a single-parent merge snap. The dropped foreign head is recorded in Version.squashed (previousParents + laneId), so when running bit lane merge lane-b --squash from lane-a (where lane-a lives on scope-a and lane-b on scope-b):

Lane-b's intermediate snaps stay on scope-b — exporting lane-a only ships the merge snap to scope-a, not lane-b's chain.
bit log from a fresh consumer of scope-a is clean (no ParentNotFound walking into scope-b).
Re-merging lane-b after it advances works correctly: the existing VersionHistory graph treats squashed.previousParents as edges (version-history.ts:199, 263), so diverge calculation finds the last-merged head as the base instead of re-merging history that was already squashed.

No remote-side changes were needed.

Implementation

UnmergedComponent gains shouldSquash?: boolean.
mergeSnaps/applyVersionMultiple/applyVersion thread shouldSquash through and set the flag on the unmerged entry when divergent.
Squash guard in merge-lanes.main.runtime.ts no longer throws on diverged — it returns false so the snap-time path handles it.
At snap creation (snapping.main.runtime.ts), the unmerged-component branch checks shouldSquash: if true, calls setSquashed with both pre-squash parents and skips addParent of the foreign head.

Test plan

E2e tests in e2e/harmony/lanes/merge-lanes-squash-diverge.e2e.ts (21 cases, all passing):

single-scope diverged squash → single-parent snap, squash metadata, clean bit log
multi-scope diverged squash → scope-a does NOT contain lane-b snaps; scope-b still does
fresh consumer imports lane-a from scope-a → bit log and bit status don't throw
re-merge of lane-b after it advances → second merge snap records new dropped head, not the previously-squashed one; re-export succeeds
sanity: diverged without --squash still produces two-parent merge snap
sanity: squash on fast-forward unchanged

Previously `bit lane merge <other-lane> --squash` threw when components were diverged in history. Now it produces a single-parent merge snap with the dropped other-lane head captured in Version.squashed metadata. This keeps the merged lane's history self-contained on its own scope, so exporting after merging from a foreign scope does not pull that scope's intermediate snap objects. The existing VersionHistory graph already treats squashed.previousParents as edges, so subsequent re-merges from the same source lane correctly identify the previously-merged head as the diverge base.

Copilot

Pull request overview

Enables bit lane merge <other-lane> --squash to work even when component histories are diverged, by creating a single-parent merge snap and recording the dropped foreign head in Version.squashed metadata (preventing cross-scope history pull-in on export).

Changes:

Thread a shouldSquash flag through lane-merge → merge engine → snap creation, and persist it on UnmergedComponent.
Adjust squash handling so diverged components no longer throw in the pre-squash path; instead they’re handled at snap-creation time via squashed metadata.
Add comprehensive E2E coverage for single-scope and multi-scope diverged squash, re-merge behavior, and non-squash sanity checks.

Reviewed changes

Copilot reviewed 5 out of 5 changed files in this pull request and generated 1 comment.

Show a summary per file

File	Description
scopes/lanes/merge-lanes/merge-lanes.main.runtime.ts	Stops throwing on diverged `--squash` and passes `shouldSquash` through to the merge engine.
scopes/component/merging/merging.main.runtime.ts	Propagates `shouldSquash` into `UnmergedComponent` entries for diverged merges.
scopes/component/snapping/snapping.main.runtime.ts	On snap creation, uses `shouldSquash` to record `Version.squashed` metadata and avoid adding the foreign head as a second parent.
components/legacy/scope/lanes/unmerged-components.ts	Extends `UnmergedComponent` schema with optional `shouldSquash`.
e2e/harmony/lanes/merge-lanes-squash-diverge.e2e.ts	Adds E2E tests covering diverged squash scenarios and regressions.

…rged-lanes

setSquashed appends the passed log to version.modified[] by reference. Previously we passed version.log directly and then mutated version.log.message a few lines below — which corrupted the modified-log entry. Build a dedicated squash log (matching the pattern in merge-lanes.main.runtime.ts:getLogForSquash) so the audit trail stays correct. Addresses Copilot review on PR #10387.

Copilot

Copilot encountered an error and was unable to review this pull request. You can try again by re-requesting a review.

Four cleanups from the /simplify pass: 1. (reuse) Move getLogForSquash to @teambit/harmony.modules.get-basic-log so snapping and merge-lanes share one helper instead of two copies of the squash-log message format. 2. (simplify) Drop the defensive currentParent ternary in snapping — the diverged path always has a current-lane parent (that's what makes it diverged); the empty-parents fallback was unreachable. 3. (simplify) Shorten the verbose modified-log comment in snapping. 4. (simplify) Drop the redundant inner Boolean() in the shouldSquash coercion in merging — outer Boolean() is enough. Skipped findings (out of scope for /simplify): - E1: hoist getBasicLog out of the per-component path — needs plumbing through the merge flow. - A1: unify the two squash code paths (squashSnaps vs new branch) — major refactor. - A2/A3/A4: move shouldSquash off UnmergedComponent / make it mutually exclusive with unrelated — correctness/design concerns, not simplification. All 21 squash-on-diverged e2e tests pass.

Copilot

Pull request overview

Copilot reviewed 7 out of 7 changed files in this pull request and generated 1 comment.

getBasicLog calls getBitCloudUser, which hits api.<cloudDomain>/user every invocation. When called per-component during a merge or large snap, that's N HTTP requests — slow and rate-limit-prone. Cache the promise resolving to {username, email} at the module level. Date stays fresh per call. Single user, single process → single network round-trip regardless of caller count. Addresses Copilot review on PR #10387.

Copilot

Pull request overview

Copilot reviewed 7 out of 7 changed files in this pull request and generated no new comments.

Mirrors the lane-to-lane diverged-squash describe block, but merges dev into main when both have advanced from their common ancestor. Locks down the symmetry: the snap-time shouldSquash branch produces a single-parent merge snap on main with the dropped lane head recorded in Version.squashed, just as it does for a lane destination. Asserts: single-parent merge snap on main, squashed metadata records the lane head, bit log on main shows main chain + merge snap (no lane intermediates), and exporting main after the merge succeeds.

…r verification dev lives on scope-b, main on scope-a. dev advances by 3 snaps (L1..L3) and main advances independently by 3 snaps (M1..M3), so the histories are diverged AND multi-snap on both sides. bit lane merge -b/dev --squash --auto-merge-resolve theirs is run from main. After export of the merge snap to scope-a, a fresh consumer adds ONLY scope-a as a remote (not scope-b), imports comp1, and verifies: - bit status / bit log don't throw - log includes common ancestor + main chain + merge snap - log does NOT include L1, L2, or L3 — the dev lane intermediates stayed on scope-b, never shipped to scope-a, and the consumer can read the full main history without needing scope-b reachable

Copilot

Pull request overview

Copilot reviewed 7 out of 7 changed files in this pull request and generated no new comments.

Copilot AI review requested due to automatic review settings May 20, 2026 20:18

Copilot started reviewing on behalf of davidfirst May 20, 2026 20:19 View session

Copilot AI reviewed May 20, 2026

View reviewed changes

Comment thread scopes/component/snapping/snapping.main.runtime.ts Outdated

davidfirst added 2 commits June 4, 2026 11:33

Merge remote-tracking branch 'origin/master' into feat/squash-on-dive…

37fd7b5

…rged-lanes

Copilot AI review requested due to automatic review settings June 4, 2026 17:53

Copilot started reviewing on behalf of davidfirst June 4, 2026 17:53 View session

Copilot AI reviewed Jun 4, 2026

davidfirst added 2 commits June 4, 2026 14:06

Merge branch 'master' into feat/squash-on-diverged-lanes

d2c5404

Copilot AI review requested due to automatic review settings June 4, 2026 20:54

GiladShoham approved these changes Jun 4, 2026

View reviewed changes

Copilot started reviewing on behalf of davidfirst June 4, 2026 20:54 View session

Copilot AI reviewed Jun 4, 2026

View reviewed changes

Comment thread scopes/harmony/modules/get-basic-log/basic-log.ts

davidfirst added 2 commits June 4, 2026 17:14

Merge branch 'master' into feat/squash-on-diverged-lanes

c7cde4d

Copilot AI review requested due to automatic review settings June 5, 2026 17:36

Copilot started reviewing on behalf of davidfirst June 5, 2026 17:36 View session

Copilot AI reviewed Jun 5, 2026

View reviewed changes

davidfirst added 2 commits June 5, 2026 13:50

Copilot AI review requested due to automatic review settings June 5, 2026 18:37

Copilot started reviewing on behalf of davidfirst June 5, 2026 18:37 View session

Copilot AI reviewed Jun 5, 2026

View reviewed changes

davidfirst merged commit f67d560 into master Jun 5, 2026
13 checks passed

davidfirst deleted the feat/squash-on-diverged-lanes branch June 5, 2026 19:22

Navigation Menu

Uh oh!

feat(lanes): allow --squash on diverged components - #10387

feat(lanes): allow --squash on diverged components#10387
davidfirst merged 9 commits into
masterfrom
feat/squash-on-diverged-lanes

davidfirst commented May 20, 2026 •

edited

Loading

Uh oh!

Copilot AI left a comment

Uh oh!

Uh oh!

Copilot AI left a comment

Uh oh!

Copilot AI left a comment

Uh oh!

Uh oh!

Copilot AI left a comment

Uh oh!

Copilot AI left a comment

Uh oh!

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

3 participants

Uh oh!

Conversation

davidfirst commented May 20, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Why

What

Implementation

Test plan

Uh oh!

Copilot AI left a comment

Choose a reason for hiding this comment

Pull request overview

Reviewed changes

Uh oh!

Uh oh!

Copilot AI left a comment

Choose a reason for hiding this comment

Uh oh!

Copilot AI left a comment

Choose a reason for hiding this comment

Pull request overview

Uh oh!

Uh oh!

Copilot AI left a comment

Choose a reason for hiding this comment

Pull request overview

Uh oh!

Copilot AI left a comment

Choose a reason for hiding this comment

Pull request overview

Uh oh!

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

3 participants

davidfirst commented May 20, 2026 •

edited

Loading