chore: development v0.2.125 - comprehensive testing complete [auto-commit]

Author: githubrobbi

CHANGELOG.md

@@ -23,7 +23,7 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0
 - Cleaned up all TTAPI references from justfile and build scripts
 - Updated justfile header and recipes for UFFS
 
-## [0.2.124] - 2026-01-27
+## [0.2.125] - 2026-01-27
 
 ### Added
 - Baseline CI validation for modernization effort
@@ -46,7 +46,7 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0
 ### Fixed
 - Various MFT parsing edge cases
 
-[Unreleased]: https://github.com/githubrobbi/UltraFastFileSearch/compare/v0.2.124...HEAD
-[0.2.124]: https://github.com/githubrobbi/UltraFastFileSearch/compare/v0.2.114...v0.2.124
+[Unreleased]: https://github.com/githubrobbi/UltraFastFileSearch/compare/v0.2.125...HEAD
+[0.2.125]: https://github.com/githubrobbi/UltraFastFileSearch/compare/v0.2.114...v0.2.125
 [0.2.114]: https://github.com/githubrobbi/UltraFastFileSearch/releases/tag/v0.2.114
 

Cargo.lock

@@ -39,7 +39,7 @@ exclude = [
 # Workspace Package Metadata (inherited by all crates)
 # ─────────────────────────────────────────────────────────────────────────────
 [workspace.package]
-version = "0.2.124"
+version = "0.2.125"
 edition = "2024"
 rust-version = "1.85"
 license = "MPL-2.0 OR LicenseRef-UFFS-Commercial"

Cargo.toml

@@ -77,21 +77,27 @@ Fixing identified differences between C++ and Rust UFFS outputs to achieve featu
 
 ---
 
-## Fix 4: Descendant Count Difference ⏳ DEFERRED
+## Fix 4: Descendant Count Difference ✅ FIXED
 
 ### What Failed
 - C++ G:\: 15,115 descendants
 - Rust G:\: 15,087 descendants (28 fewer)
 
 ### Why It Failed
-- Likely due to ADS handling differences
-- C++ may count ADS as separate entries in descendant count
-- Rust tree metrics are computed per-FRS, not per-stream
+- C++ counts each ADS (Alternate Data Stream) as a separate descendant
+- Rust tree metrics were computed per-FRS, counting each file as 1 regardless of stream count
+- C++ also includes all streams' sizes in treesize/tree_allocated
 
-### Status
-- Deferred for now - requires Windows testing to verify
-- The difference is small (0.18%) and may be intentional design difference
-- ADS are not separate file records, so not counting them in descendants may be correct
+### How Fixed
+Modified `compute_tree_metrics()` in `crates/uffs-mft/src/index.rs`:
+
+1. **Sum all streams' sizes**: In the first pass, iterate through all streams (default + ADS) to sum total size and allocated size, not just the first stream
+2. **Count streams as descendants**: When accumulating into parent, use `stream_count` instead of `1` to count each ADS as a separate descendant
+
+Key changes:
+- Extended `parent_info` tuple to include `stream_count`
+- Added loop to follow linked list of additional streams and sum their sizes
+- Changed accumulation from `1 + child_descendants` to `stream_count + child_descendants`
 
 ---
 
@@ -100,4 +106,5 @@ Fixing identified differences between C++ and Rust UFFS outputs to achieve featu
 | Commit | Description |
 |--------|-------------|
 | v0.2.124 | fix: C++ parity - Size on Disk, Directory Size, ADS Name |
+| v0.2.125 | fix: C++ parity - Descendant count includes ADS |
 

LOG/2026_01_27_18_00_CHANGELOG_HEALING.md

@@ -21,7 +21,7 @@ Traditional file search tools (including `os.walk`, `FindFirstFile`, etc.) work
 
 **UFFS reads the MFT directly** - once - and queries it in memory using Polars DataFrames. This is like reading the entire phonebook once instead of looking up each name individually.
 
-### Benchmark Results (v0.2.124)
+### Benchmark Results (v0.2.125)
 
 | Drive Type | Records | Time | Throughput |
 |------------|---------|------|------------|
@@ -33,7 +33,7 @@ Traditional file search tools (including `os.walk`, `FindFirstFile`, etc.) work
 
 | Comparison | Records | Time | Notes |
 |------------|---------|------|-------|
-| **UFFS v0.2.124** | **18.7 Million** | **~142 seconds** | All disks, fast mode |
+| **UFFS v0.2.125** | **18.7 Million** | **~142 seconds** | All disks, fast mode |
 | UFFS v0.1.30 | 18.7 Million | ~315 seconds | Baseline |
 | Everything | 19 Million | 178 seconds | All disks |
 | WizFile | 6.5 Million | 299 seconds | Single HDD |

README.md

@@ -1746,7 +1746,9 @@ impl MftIndex {
 
         // Phase 1: Build parent links and count pending children
         // Also initialize base metrics (each node's own contribution)
-        // First pass: initialize base metrics and collect parent info
+        // First pass: collect parent info and sum ALL streams' sizes (for C++ parity)
+        // C++ counts each ADS as a separate descendant and includes ADS sizes in
+        // treesize
         let parent_info: Vec<_> = self
             .records
             .iter()
@@ -1755,14 +1757,37 @@ impl MftIndex {
             .map(|(idx, record)| {
                 let frs = record.frs;
                 let parent_frs = record.first_name.parent_frs;
-                let size = record.first_stream.size.length;
-                let allocated = record.first_stream.size.allocated;
-                (idx, frs, parent_frs, size, allocated)
+                let stream_count = record.stream_count;
+
+                // Sum sizes across ALL streams (default + ADS) for C++ parity
+                let mut total_size = record.first_stream.size.length;
+                let mut total_allocated = record.first_stream.size.allocated;
+
+                // Follow the linked list of additional streams
+                let mut next_entry = record.first_stream.next_entry;
+                while next_entry != NO_ENTRY {
+                    if let Some(stream) = self.streams.get(next_entry as usize) {
+                        total_size = total_size.saturating_add(stream.size.length);
+                        total_allocated = total_allocated.saturating_add(stream.size.allocated);
+                        next_entry = stream.next_entry;
+                    } else {
+                        break;
+                    }
+                }
+
+                (
+                    idx,
+                    frs,
+                    parent_frs,
+                    total_size,
+                    total_allocated,
+                    stream_count,
+                )
             })
             .collect();
 
         // Second pass: initialize base metrics
-        for (idx, _frs, _parent_frs, size, allocated) in &parent_info {
+        for (idx, _frs, _parent_frs, size, allocated, _stream_count) in &parent_info {
             if let Some(record) = self.records.get_mut(*idx) {
                 record.descendants = 0;
                 record.treesize = *size;
@@ -1771,7 +1796,7 @@ impl MftIndex {
         }
 
         // Third pass: build parent links
-        for (idx, frs, parent_frs, _size, _allocated) in &parent_info {
+        for (idx, frs, parent_frs, _size, _allocated, _stream_count) in &parent_info {
             // Skip root or self-parent
             if parent_frs == frs {
                 continue;
@@ -1816,17 +1841,25 @@ impl MftIndex {
 
             let parent_idx_usize = parent_idx_u32 as usize;
 
-            // Read child's metrics
-            let (child_descendants, child_treesize, child_tree_allocated) =
+            // Read child's metrics and stream_count
+            // For C++ parity, each stream (default + ADS) counts as a separate descendant
+            let (child_descendants, child_treesize, child_tree_allocated, child_stream_count) =
                 if let Some(child) = self.records.get(child_idx) {
-                    (child.descendants, child.treesize, child.tree_allocated)
+                    (
+                        child.descendants,
+                        child.treesize,
+                        child.tree_allocated,
+                        child.stream_count,
+                    )
                 } else {
                     continue; // Safety: skip if index is invalid
                 };
 
             // Accumulate into parent
+            // Use stream_count instead of 1 to count each ADS as a separate descendant
+            // This matches C++ behavior where ADS are expanded as separate rows
             if let Some(parent) = self.records.get_mut(parent_idx_usize) {
-                parent.descendants += 1 + child_descendants;
+                parent.descendants += u32::from(child_stream_count) + child_descendants;
                 parent.treesize += child_treesize;
                 parent.tree_allocated += child_tree_allocated;
             }

crates/uffs-mft/src/index.rs

@@ -1 +1 @@
-v0.2.124
+v0.2.125

dist/latest

@@ -0,0 +1,4 @@
+90a0d6912a8154557ae1a22074c3a3658d3a654dba7149c621f048c4b549f031  dist/v0.2.125/uffs/uffs-windows-x64.exe (68656128 bytes)
+1855d47de6f1fb2548797ffd78f44d791f5e544030572175f14d4c14db38b497  dist/v0.2.125/uffs_mft/uffs_mft-windows-x64.exe (22208512 bytes)
+f3914b48b2ed7c786ed0419f63f6979e7b4fb4d2c63d04ee1052a9722edfb6b4  dist/v0.2.125/uffs_tui/uffs_tui-windows-x64.exe (66257920 bytes)
+70e26362ed324a38d0cede503b58a5635790abd86ed1808479d3d3636ac985af  dist/v0.2.125/uffs_gui/uffs_gui-windows-x64.exe (1682432 bytes)