4.4.26: Static Linux Builds, Memory-Aware Batching & Cross-Compilation Fixes

This release includes critical build system improvements and enhanced batching capabilities for more predictable resource usage.

### Cross-Compilation & Static Builds
- Fixed cross-compilation for Linux by introducing architecture-specific `CGO_CFLAGS` (x86_64-linux-gnu and aarch64-linux-gnu), ensuring correct header resolution during static linking.
- Enabled fully static Linux binaries using musl, producing self-contained executables with no external runtime dependencies.

### Memory-Aware Batching (New Feature)
- Added `--batch-mem` CLI option to control batching based on estimated memory usage (e.g., 128K, 64M, 1G), in addition to size-based limits.
- Introduced configurable min/max batch sizes and memory thresholds, with conservative memory estimation per sequence to avoid over-allocation.
- Implemented intelligent flushing logic that triggers when *either* byte or record count limits are exceeded, ensuring predictable memory behavior.
- Improved garbage collection after large batch discards to reduce memory pressure during large-scale processing.

### Build & Toolchain Improvements
- Updated Go toolchain to 1.26.1 and bumped key dependencies (e.g., golang.org/x/net v0.38.0).
- Enhanced build error reporting: logs are now displayed before cleanup on failure.
- Fixed Makefile quoting for `LDFLAGS` containing spaces.
- Updated install script to properly configure GOROOT, GOPATH, and GOTOOLCHAIN, with added progress feedback for downloads.

All batching behavior is backward-compatible and uses sensible defaults (128 MB memory, min: 1 record, max: 2000 records) to ensure smooth upgrades.

.github/workflows/release.yml

@@ -16,7 +16,7 @@ jobs:
       - name: Setup Go
         uses: actions/setup-go@v5
         with:
-          go-version: "1.26"
+          go-version: "1.23"
       - name: Checkout obitools4 project
         uses: actions/checkout@v4
       - name: Run tests
@@ -32,10 +32,12 @@ jobs:
             goos: linux
             goarch: amd64
             output_name: linux_amd64
+            cgo_cflags: "-I/usr/include/x86_64-linux-gnu -I/usr/include"
           - os: ubuntu-24.04-arm
             goos: linux
             goarch: arm64
             output_name: linux_arm64
+            cgo_cflags: "-I/usr/include/aarch64-linux-gnu -I/usr/include"
           - os: macos-15-intel
             goos: darwin
             goarch: amd64
@@ -54,7 +56,7 @@ jobs:
       - name: Setup Go
         uses: actions/setup-go@v5
         with:
-          go-version: "1.26"
+          go-version: "1.23"
 
       - name: Extract version from tag
         id: get_version
@@ -62,6 +64,12 @@ jobs:
           TAG=${GITHUB_REF#refs/tags/Release_}
           echo "version=$TAG" >> $GITHUB_OUTPUT
 
+      - name: Install build tools (Linux)
+        if: runner.os == 'Linux'
+        run: |
+          sudo apt-get update -q
+          sudo apt-get install -y musl-tools zlib1g-dev
+
       - name: Install build tools (macOS)
         if: runner.os == 'macOS'
         run: |
@@ -69,30 +77,21 @@ jobs:
           xcode-select --install 2>/dev/null || true
           xcode-select -p
 
-      - name: Build binaries (Linux)
-        if: runner.os == 'Linux'
-        env:
-          VERSION: ${{ steps.get_version.outputs.version }}
-        run: |
-          docker run --rm \
-            -v "$(pwd):/src" \
-            -w /src \
-            -e VERSION="${VERSION}" \
-            golang:1.26-alpine \
-            sh -c "apk add --no-cache gcc musl-dev zlib-dev zlib-static make && \
-                   make LDFLAGS='-linkmode=external -extldflags=-static' obitools"
-          mkdir -p artifacts
-          tar -czf artifacts/obitools4_${VERSION}_${{ matrix.output_name }}.tar.gz -C build .
-
-      - name: Build binaries (macOS)
-        if: runner.os == 'macOS'
+      - name: Build binaries
         env:
           GOOS: ${{ matrix.goos }}
           GOARCH: ${{ matrix.goarch }}
           VERSION: ${{ steps.get_version.outputs.version }}
+          CC: ${{ matrix.goos == 'linux' && 'musl-gcc' || '' }}
+          CGO_CFLAGS: ${{ matrix.cgo_cflags || '' }}
         run: |
-          make obitools
+          if [ "$GOOS" = "linux" ]; then
+            make LDFLAGS='-linkmode=external -extldflags=-static' obitools
+          else
+            make obitools
+          fi
           mkdir -p artifacts
+          # Create a single tar.gz with all binaries for this platform
           tar -czf artifacts/obitools4_${VERSION}_${{ matrix.output_name }}.tar.gz -C build .
 
       - name: Upload artifacts

pkg/obiiter/distribute.go

@@ -57,21 +57,34 @@ func (dist *IDistribute) Classifier() *obiseq.BioSequenceClassifier {
 }
 
 // Distribute organizes the biosequences from the iterator into batches
-// based on the provided classifier. It returns an IDistribute instance
-// that manages the distribution of the sequences.
+// based on the provided classifier and batch sizes. It returns an
+// IDistribute instance that manages the distribution of the sequences.
 //
-// Batches are flushed when either BatchSizeMax() sequences or BatchMem()
-// bytes are accumulated per key, mirroring the RebatchBySize strategy.
-func (iterator IBioSequence) Distribute(class *obiseq.BioSequenceClassifier) IDistribute {
-	maxCount := obidefault.BatchSizeMax()
-	maxBytes := obidefault.BatchMem()
+// Parameters:
+//   - class: A pointer to a BioSequenceClassifier used to classify
+//     the biosequences during distribution.
+//   - sizes: Optional integer values specifying the batch size. If
+//     no sizes are provided, a default batch size of 5000 is used.
+//
+// Returns:
+// An IDistribute instance that contains the outputs of the
+// classified biosequences, a channel for new data notifications,
+// and the classifier used for distribution. The method operates
+// asynchronously, processing the sequences in separate goroutines.
+// It ensures that the outputs are closed and cleaned up once
+// processing is complete.
+func (iterator IBioSequence) Distribute(class *obiseq.BioSequenceClassifier, sizes ...int) IDistribute {
+	batchsize := obidefault.BatchSize()
 
 	outputs := make(map[int]IBioSequence, 100)
 	slices := make(map[int]*obiseq.BioSequenceSlice, 100)
-	bufBytes := make(map[int]int, 100)
 	orders := make(map[int]int, 100)
 	news := make(chan int)
 
+	if len(sizes) > 0 {
+		batchsize = sizes[0]
+	}
+
 	jobDone := sync.WaitGroup{}
 	lock := sync.Mutex{}
 
@@ -102,7 +115,6 @@ func (iterator IBioSequence) Distribute(class *obiseq.BioSequenceClassifier) IDi
 					slice = &s
 					slices[key] = slice
 					orders[key] = 0
-					bufBytes[key] = 0
 
 					lock.Lock()
 					outputs[key] = MakeIBioSequence()
@@ -111,20 +123,14 @@ func (iterator IBioSequence) Distribute(class *obiseq.BioSequenceClassifier) IDi
 					news <- key
 				}
 
-				sz := s.MemorySize()
-				countFull := maxCount > 0 && len(*slice) >= maxCount
-				memFull := maxBytes > 0 && bufBytes[key]+sz > maxBytes && len(*slice) > 0
-				if countFull || memFull {
+				*slice = append(*slice, s)
+
+				if len(*slice) == batchsize {
 					outputs[key].Push(MakeBioSequenceBatch(source, orders[key], *slice))
 					orders[key]++
 					s := obiseq.MakeBioSequenceSlice()
 					slices[key] = &s
-					slice = &s
-					bufBytes[key] = 0
 				}
-
-				*slice = append(*slice, s)
-				bufBytes[key] += sz
 			}
 		}
 

pkg/obilua/obiseqslice.go

@@ -31,8 +31,7 @@ func obiseqslice2Lua(interpreter *lua.LState,
 }
 
 func newObiSeqSlice(luaState *lua.LState) int {
-	capacity := luaState.OptInt(1, 0)
-	seqslice := obiseq.NewBioSequenceSlice(capacity)
+	seqslice := obiseq.NewBioSequenceSlice()
 	luaState.Push(obiseqslice2Lua(luaState, seqslice))
 	return 1
 }

pkg/obioptions/version.go

@@ -3,7 +3,7 @@ package obioptions
 // Version is automatically updated by the Makefile from version.txt
 // The patch number (third digit) is incremented on each push to the repository
 
-var _Version = "Release 4.4.29"
+var _Version = "Release 4.4.26"
 
 // Version returns the version of the obitools package.
 //

pkg/obiseq/worker.go

@@ -104,11 +104,11 @@ func SeqToSliceWorker(worker SeqWorker,
 			for _, s := range input {
 				r, err := worker(s)
 				if err == nil {
-					if i+len(r) > cap(output) {
-						output = slices.Grow(output[:i], len(r))
-						output = output[:cap(output)]
-					}
 					for _, rs := range r {
+						if i == len(output) {
+							output = slices.Grow(output, cap(output))
+							output = output[:cap(output)]
+						}
 						output[i] = rs
 						i++
 					}

pkg/obitools/obidistribute/distribute.go

@@ -46,7 +46,8 @@ func CLIDistributeSequence(sequences obiiter.IBioSequence) {
 		formater = obiformats.WriteSequencesToFile
 	}
 
-	dispatcher := sequences.Distribute(CLISequenceClassifier())
+	dispatcher := sequences.Distribute(CLISequenceClassifier(),
+		obidefault.BatchSize())
 
 	obiformats.WriterDispatcher(CLIFileNamePattern(),
 		dispatcher, formater, opts...,

version.txt

@@ -1 +1 @@
-4.4.29
+4.4.26