Try to parse TL as UserFn before DB in legacy binary (de)serializer - #1
Closed
StachuDotNet wants to merge 1 commit into
Closed
StachuDotNet wants to merge 1 commit into
StachuDotNet wants to merge 1 commit into
Conversation
Owner
Author
|
PR in wrong repo. |
StachuDotNet
pushed a commit
that referenced
this pull request
Mar 1, 2023
StachuDotNet
added a commit
that referenced
this pull request
Jan 30, 2024
…ension Attempt #1: Publish VS Code Extension
StachuDotNet
added a commit
that referenced
this pull request
Jul 26, 2026
The "promote a value back to source" primitive that unlocks live values (dream #1). `dvalToExpr` takes a runtime value and returns a ProgramTypes expression, or an Error with a reason when the value has no literal syntax. Mapped: unit, bool, all int widths, float (split into sign/whole/fraction for EFloat), char, string, list, tuple, dict — recursively, so nested collections round-trip. Fresh gid IDs, which don't affect rendered source. Fails (Error, caller falls back): DateTime and Uuid (no literal form — pointed at their parse fns), records and enums (not yet mappable — need PT type-name construction), and functions / DB refs / blobs (no source at all). Tests are the standard `.dark` execution testfile. Because IDs are random, they pattern-match the result structurally rather than comparing whole expressions, and cover each literal kind, nested list/tuple/dict shapes, and the Error path for unmappable values (DDB, DUuid). Claude-Session: https://claude.ai/code/session_015BbLwctE41vpgJksNa9X4C
StachuDotNet
added a commit
that referenced
this pull request
Aug 8, 2026
…->main isolation leak Two things, both load-bearing: 1. ISOLATION LEAK (real bug): authoring to main ran WipRefresh, which treated a branch's effective=0 ops as main WIP -- discardWipOps + re-insert flipped them effective=1 and folded them into main, so a branch's work leaked into main just by authoring on main. Fixed: getWipOps + discardWipOps exclude op_branches-tagged ops (branch-pending state is not main WIP). 2. REBASE on a RELOAD-STABLE base model. The base_ts timestamp watermark was meaningless (reload re-stamps origin_ts), so it false-positived every merge after any main movement. Replaced with PER-NAME content-hash bases: branch_name_bases records main's content-hash for each name a branch first touches (stable across reload). A conflict = main's CURRENT hash for that name differs from the base (main changed the same name). canMerge blocks only on real conflicts; `dark ops rebase <id>` accepts main's state + reports what moved, unblocking merge (the branch's ops LWW on top). Verified end to end: author on a branch, author a DIFFERENT name to main -> branch stays isolated (effective=0), merge succeeds (no false conflict); author the SAME name on both -> merge blocked -> rebase -> merge succeeds. Plain eval + val to main unchanged; 6 BranchOverlay tests green. Design: notes/fresh-arch/scm-spec.md (open-q #1 resolved: base op set, realized as per-name hashes). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01LYzWX7SQ1W1Cf3gZTu2mHY
StachuDotNet
added a commit
that referenced
this pull request
Aug 8, 2026
…terialises rt_dval) branchValueContentFoldIsolatesName only guarded the isolation half of the expression-value fix (431f903): a branch value's AddValue content folds into package_values but its SetName must NOT land in main's locations. The other half of that fix -- that an EXPRESSION body actually materialises its Dval into rt_dval via evaluateAllValues (applyOps alone stores NULL) -- had no automated guard; it was the original bug (expression-valued branch values read a NULL rt_dval and returned nothing) and was only verified by hand. Extend the same test to run the production evaluateAllValues path after the fold and assert rt_dval materialises to 3+4=7. One test now locks both halves, so a regression in either surfaces in CI instead of only under manual branch probing. Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01LYzWX7SQ1W1Cf3gZTu2mHY
StachuDotNet
added a commit
that referenced
this pull request
Aug 8, 2026
BranchOps.Tests.fs (545 lines) was disabled in BOTH the fsproj and the test list and referenced LibDB.Rebase / LibDB.BranchOpPlayback, modules that no longer exist -- it would not compile if re-enabled. It tested the carved commit / partial-commit / rebase model; the overlay model replaced that wholesale, with fresh coverage in BranchOverlay.Tests + BranchCli.Tests (isolation, convergence, merge, review, the bug #1 value fold). Nothing in it is salvageable, so drop the file and its two tombstone comments rather than leave a misleading orphan. Also fix LibDB/README.md: it still listed Rebase.fs / Merge.fs as SCM-ops files. Both were folded into Branches.fs by the overlay carve; point the doc there. Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01LYzWX7SQ1W1Cf3gZTu2mHY
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
What is the problem/goal being addressed?
If you
currently the user function is (seemingly consistently) 'parsed' as a DB TopLevel rather than a UserFn.
What is the solution to this problem?
In the legacy binary serialization server, attempt to parse UserFns before DBs.
How are you sure this works/how was this tested?
I've tested this in the UI, and confirmed no regressions in other tests.
As existing integration tests already covered this bug (when the integration tests target F#, which they shortly will by default), I did not write new tests to cover this.