.gitignore

+.dub/
+dub.selections.json
+BioD/
+BioD
+./BioD
+undeaD/
+lz4/
+core
+bin/
+test*
 build/
 shunit*
 /*.sam
@@ -6,4 +16,9 @@ shunit*
 /*.cram
 /*.crai
 /*.txt
+*.hex
+*.zcat
+*.out
 /utils/ldc_version_info_.d
+profile.data
+profile.raw

.gitmodules

-[submodule "BioD"]
-	path = BioD
-	url = https://github.com/biod/BioD.git
 [submodule "htslib"]
 	path = htslib
 	url = https://github.com/lomereiter/htslib.git
-[submodule "lz4"]
-	path = lz4
-	url = https://github.com/Cyan4973/lz4
 [submodule "undeaD"]
 	path = undeaD
 	url = https://github.com/dlang/undeaD
+[submodule "BioD"]
+	path = BioD
+	url = https://github.com/biod/BioD.git
+[submodule "lz4"]
+	path = lz4
+	url = https://github.com/lz4/lz4.git

.travis.yml

+language: d
+d:
+  - ldc
+matrix:
+  # OSX testing is under development
+  allow_failures:
+    - os: osx
+  include:
+    - os: linux
+      compiler: gcc
+      addons:
+        apt:
+          packages:
+            # Our dev environment is a more recent GNU C++
+            # note that Debian liblz4-dev no longer supports LZ4 frames
+            # - liblz4-dev
+            - shunit2
+    - os: osx
+      compiler: clang
+script:
+  - make
+  # - make test - disable tests because shunit2 we use is older
\ No newline at end of file

INSTALL.md

+# INSTALL SAMBAMBA
+
+## Sambamba dependencies
+
+* D compiler
+* gcc tool chain
+* BioD
+* htslib
+* undeaD
+* libz
+* liblz4

Makefile

-D_COMPILER=dmd
-D_FLAGS=--compiler=dmd -IBioD -IundeaD/src -g -d#-O -release -inline # -version=serial
-LDMD=ldmd2
+# This is a minimalistic make file to build sambamba with ldc2 as per instructions on
+# https://github.com/biod/sambamba#compiling-sambamba
+#
+# Typical usage:
+#
+#    make LIBRARY_PATH=~/opt/ldc2-1.7.0-linux-x86_64/lib debug|profile|release|static
+#
+# Static release with optimization (for releases):
+#
+#   make LIBRARY_PATH=~/opt/ldc2-1.7.0-linux-x86_64/lib pgo-static
+#
+# Debug version
+#
+#   make LIBRARY_PATH=~/opt/ldc2-1.7.0-linux-x86_64/lib debug
 
-STATIC_LIB_PATH=-Lhtslib -Llz4/lib
-STATIC_LIB_SUBCMD=$(STATIC_LIB_PATH) -Wl,-Bstatic -lhts -llz4 -Wl,-Bdynamic
-RDMD_FLAGS=--force --build-only --compiler=$(D_COMPILER) $(D_FLAGS)
+D_COMPILER=ldc2
+DFLAGS      = -wi -I. -IBioD -IundeaD/src -g
 
-PLATFORM := $(shell uname -s)
+DLIBS       = $(LIBRARY_PATH)/libphobos2-ldc.a $(LIBRARY_PATH)/libdruntime-ldc.a
+DLIBS_DEBUG = $(LIBRARY_PATH)/libphobos2-ldc-debug.a $(LIBRARY_PATH)/libdruntime-ldc-debug.a
+LIBS        = htslib/libhts.a lz4/lib/liblz4.a -L-L$(LIBRARY_PATH) -L-lrt -L-lpthread -L-lm
+LIBS_STATIC = $(LIBRARY_PATH)/libc.a $(DLIBS) htslib/libhts.a $(LIBRARY_PATH)/liblz4.a
+SRC         = $(wildcard main.d utils/*.d thirdparty/*.d cram/*.d) $(wildcard undeaD/src/undead/*.d) $(wildcard BioD/bio/*/*.d BioD/bio/*/*/*.d BioD/bio2/*.d BioD/bio2/*/*.d) $(wildcard sambamba/*.d sambamba/*/*.d sambamba/*/*/*.d)
+OBJ         = $(SRC:.d=.o) utils/ldc_version_info_.o
+OUT         = bin/sambamba
 
-ifeq "$(PLATFORM)" "Darwin"
+STATIC_LIB_PATH=-Lhtslib -Llz4
 
-LINK_CMD=gcc -dead_strip -lphobos2-ldc -ldruntime-ldc -lm -lpthread htslib/libhts.a lz4/lib/liblz4.a build/sambamba.o -o build/sambamba
-DMD_STATIC_LIBS=htslib/libhts.a lz4/lib/liblz4.a
+.PHONY: all debug release static clean test
 
-define split-debug
-dsymutil build/sambamba -o build/sambamba.dSYM
-strip -S build/sambamba
-endef
+debug:                             DFLAGS += -O0 -d-debug -link-debuglib
 
-else
+profile:                           DFLAGS += -fprofile-instr-generate=profile.raw
 
-LINK_CMD=gcc -Wl,--gc-sections -o build/sambamba build/sambamba.o $(STATIC_LIB_SUBCMD) -lphobos2-ldc -ldruntime-ldc  -lrt -lpthread -lm -ldl
-DMD_STATIC_LIBS=-L-Lhtslib -L-l:libhts.a -L-l:libphobos2.a -L-Llz4/lib -L-l:liblz4.a
+release static profile pgo-static: DFLAGS += -O3 -release -enable-inlining -boundscheck=off
 
-define split-debug
-objcopy --only-keep-debug build/sambamba sambamba.debug
-objcopy --strip-debug build/sambamba
-objcopy --add-gnu-debuglink=sambamba.debug build/sambamba
-mv sambamba.debug build/
-endef
+static:                            DFLAGS += -static -L-Bstatic
 
-endif
+pgo-static:                        DFLAGS += -fprofile-instr-use=profile.data
 
-PREREQS := ldc-version-info htslib-static lz4-static
+all: release
 
-# DMD only - this goal is used because of fast compilation speed, during development
-all: $(PREREQS)
-	mkdir -p build/
-	rdmd --force --build-only $(D_FLAGS) $(DMD_STATIC_LIBS) -ofbuild/sambamba main.d
+lz4-static: lz4/lib/liblz4.a
 
-# This is the main Makefile goal, used for building releases (best performance)
-sambamba-ldmd2-64: $(PREREQS)
-	mkdir -p build/
-	$(LDMD) @sambamba-ldmd-release.rsp
-	$(LINK_CMD)
-	$(split-debug)
+lz4/lib/liblz4.a: lz4/lib/lz4.c lz4/lib/lz4hc.c lz4/lib/lz4frame.c lz4/lib/xxhash.c
+	cd lz4/lib && gcc -O3 -c lz4.c lz4hc.c lz4frame.c xxhash.c && $(AR) rcs liblz4.a lz4.o lz4hc.o lz4frame.o xxhash.o
 
-# For debugging; GDB & Valgrind are more friendly to executables created using LDC/GDC than DMD
-sambamba-ldmd2-debug: $(PREREQS)
-	mkdir -p build/
-	$(LDMD) @sambamba-ldmd-debug.rsp
-	$(LINK_CMD)
+htslib-static:
+	cd htslib && $(MAKE)
 
 ldc-version-info:
-	./gen_ldc_version_info.py $(shell which $(LDMD)) > utils/ldc_version_info_.d
+	./gen_ldc_version_info.py $(shell which ldmd2) > utils/ldc_version_info_.d
+	cat utils/ldc_version_info_.d
 
-htslib-static:
-	cd htslib && $(MAKE)
+utils/ldc_version_info_.o: ldc-version-info
+	$(D_COMPILER) $(DFLAGS) -c utils/ldc_version_info_.d -od=$(dir $@)
 
-lz4-static: lz4/lib/liblz4.a
+build-setup: lz4-static htslib-static ldc-version-info
+	mkdir -p bin/
 
-lz4/lib/liblz4.a: lz4/lib/lz4.c lz4/lib/lz4hc.c lz4/lib/lz4frame.c lz4/lib/xxhash.c
-	cd lz4/lib && $(CC) -O3 -c lz4.c lz4hc.c lz4frame.c xxhash.c && $(AR) rcs liblz4.a lz4.o lz4hc.o lz4frame.o xxhash.o
+default debug release static: $(OUT)
 
-# all below link to libhts dynamically for simplicity
+profile: release
+	./bin/sambamba sort /gnu/data/in_raw.bam -p > /dev/null
+	ldc-profdata merge -output=profile.data profile.raw
+	rm ./bin/sambamba ./bin/sambamba.o # trigger rebuild
 
-sambamba-flagstat:
-	mkdir -p build/
-	rdmd $(RDMD_FLAGS) -L-lhts -version=standalone -ofbuild/sambamba-flagstat sambamba/flagstat.d
+default: all
 
-sambamba-merge:
-	mkdir -p build/
-	rdmd $(RDMD_FLAGS) -L-lhts -version=standalone -ofbuild/sambamba-merge sambamba/merge.d
+# ---- Compile step
+%.o: %.d
+	$(D_COMPILER) $(DFLAGS) -c $< -od=$(dir $@)
 
-sambamba-index:
-	mkdir -p build/
-	rdmd $(RDMD_FLAGS) -L-lhts -version=standalone -ofbuild/sambamba-index sambamba/index.d
+singleobj:
+	$(info compile single object...)
+	$(D_COMPILER) -singleobj $(DFLAGS) -c -of=bin/sambamba.o $(SRC)
 
-sambamba-sort:
-	mkdir -p build/
-	rdmd $(RDMD_FLAGS) -L-lhts -version=standalone -ofbuild/sambamba-sort sambamba/sort.d
+# ---- Link step
+$(OUT): build-setup singleobj utils/ldc_version_info_.o
+	$(info linking...)
+	$(D_COMPILER) $(DFLAGS) -of=bin/sambamba bin/sambamba.o utils/ldc_version_info_.o $(LIBS)
 
-sambamba-view:
-	mkdir -p build/
-	rdmd $(RDMD_FLAGS) -L-lhts -version=standalone -ofbuild/sambamba-view sambamba/view.d
+test:
+	./run_tests.sh
 
-sambamba-slice:
-	mkdir -p build/
-	rdmd $(RDMD_FLAGS) -L-lhts -version=standalone -ofbuild/sambamba-slice sambamba/slice.d
+check: test
 
-sambamba-markdup:
-	mkdir -p build/
-	rdmd $(RDMD_FLAGS) -L-lhts -version=standalone -ofbuild/sambamba-markdup sambamba/markdup.d
+debug-strip:
+	objcopy --only-keep-debug bin/sambamba sambamba.debug
+	objcopy --strip-debug bin/sambamba
+	objcopy --add-gnu-debuglink=sambamba.debug bin/sambamba
+	mv sambamba.debug bin/
 
-sambamba-depth:
-	mkdir -p build/
-	rdmd $(RDMD_FLAGS) -L-lhts -version=standalone -ofbuild/sambamba-depth sambamba/depth.d
+pgo-static: profile static debug-strip
 
-sambamba-pileup:
-	mkdir -p build/
-	rdmd $(RDMD_FLAGS) -L-lhts -version=standalone -ofbuild/sambamba-pileup sambamba/pileup.d
+install:
+	install -m 0755 bin/sambamba $(prefix)/bin
 
-.PHONY: clean ldc-version-info
+clean: clean-d
+	cd htslib ; make clean
+	rm -f profile.data
+	rm -f profile.raw
 
-clean:
-	rm -rf build/ ; $(MAKE) -C htslib clean ; $(MAKE) -C lz4 clean
+clean-d:
+	rm -rf bin/*
+	rm -f $(OBJ) $(OUT) trace.{def,log}

Makefile.guix

-# GNU Guix makefile
+# GNU Guix makefile. GNU Guix is the package manager for GNU and we use it for
+# sambamba development and deployment. Normally use the standard Makefile
+# instead.
 #
 # To build sambamba on GNU Guix:
 #
@@ -6,33 +8,46 @@
 #
 # run with
 #
-#   ./build/sambamba
-
+#   ./bin/sambamba
+#
+# certain paths may need to be set, e.g.
+#
+#   make UNDEAD_PATH=../undeaD/src -f Makefile.guix
+#
+# For more information see INSTALL.md
+#
 # The following two are modified by the Guix package:
 D_COMPILER=ldc2
 LDC_LIB_PATH=$(HOME)/.guix-profile/lib
 
-DFLAGS = -wi -I. -IBioD -IundeaD/src
-DLIBS  = $(LDC_LIB_PATH)/libphobos2-ldc.a $(LDC_LIB_PATH)/libdruntime-ldc.a
-DLIBS_DEBUG = $(LDC_LIB_PATH)/libphobos2-ldc-debug.a $(LDC_LIB_PATH)/libdruntime-ldc-debug.a
+UNDEAD_PATH=../undeaD/src
+BIOD_PATH=../BioD
+
+DFLAGS = -wi -I. -I$(BIOD_PATH) -IundeaD/src -I$(UNDEAD_PATH)
+# DLIBS  = $(LDC_LIB_PATH)/libphobos2-ldc.a $(LDC_LIB_PATH)/libdruntime-ldc.a
+# DLIBS_DEBUG = -debuglib=phobos2-ldc-debug,druntime-ldc-debug $(LDC_LIB_PATH)/libphobos2-ldc-debug.a $(LDC_LIB_PATH)/libdruntime-ldc-debug.a -link-debuglib
+DLIBS_DEBUG = -debuglib=phobos2-ldc-debug-shared,druntime-ldc-debug-shared -link-debuglib -L-lphobos2-ldc-debug-shared -L-ldruntime-ldc-debug-shared
 RPATH  = -L--rpath=$(dir $(realpath $(LDC_LIB_PATH)/libz.so)):$(dir $(realpath $(LDC_LIB_PATH)/liblz4.so))
 LIBS   = htslib/libhts.a -L-L$(LDC_LIB_PATH) -L-lrt -L-lpthread -L-lm -L-lz -L-llz4
-LIBS_STATIC = $(LDC_LIB_PATH)/libc.a $(DLIBS) htslib/libhts.a $(LDC_LIB_PATH)/liblz4.a
-SRC    = $(wildcard main.d utils/*.d thirdparty/*.d cram/*.d) $(wildcard undeaD/src/undead/*.d) $(wildcard BioD/bio/*/*.d BioD/bio/*/*/*.d) $(wildcard sambamba/*.d sambamba/*/*.d sambamba/*/*/*.d)
-OBJ    = $(SRC:.d=.o) utils/ldc_version_info_.o
-OUT    = build/sambamba
+# LIBS   = htslib/libhts.a -L-lrt -L-lpthread -L-lm -L-lz -L-llz4
+LIBS_STATIC = $(DLIBS) htslib/libhts.a $(LDC_LIB_PATH)/liblz4.a
+SRC    = $(wildcard utils/*.d thirdparty/*.d cram/*.d) $(wildcard $(UNDEAD_PATH)/undead/**/*.d) $(wildcard $(UNDEAD_PATH)/undead/*.d) $(wildcard $(BIOD_PATH)/bio/*/*.d $(BIOD_PATH)/bio/*/*/*.d) $(wildcard sambamba/*.d sambamba/*/*.d sambamba/*/*/*.d) utils/ldc_version_info_.d
+OBJ    = $(SRC:.d=.o)
+OUT    = bin/sambamba
+
+.PHONY: all guix guix-debug debug release static profile clean test
 
 # The Guix targets resolve the RPATH automatically
 guix:        DFLAGS += -O -release -g # Guix strips debug flags
 
-guix-debug:  DFLAGS += -O0 -g -d-debug
+guix-debug:  DFLAGS += -O0 -g -d-debug -unittest
 
 # The following options are run in development from ~/.guix-profile and need to inject the RPATH
-debug:       DFLAGS += -O0 -g -d-debug $(RPATH) -link-debuglib
+debug:       DFLAGS += -O0 -g -d-debug $(RPATH) -link-debuglib -unittest
 
-release:     DFLAGS += -O -release $(RPATH)
+release static:     DFLAGS += -O3 -release $(RPATH) -enable-inlining -Hkeep-all-bodies -boundscheck=off
 
-static:      DFLAGS += -O -release -static -L-Bstatic
+static:      DFLAGS += -static -L-Bstatic -L-L/gnu/store/rmjlycdgiq8pfy5hfi42qhw3k7p6kdav-glibc-2.25/lib/
 
 profile:     DFLAGS += -g -O -profile $(RPATH)
 
@@ -42,21 +57,17 @@ static:         LIBS = $(LIBS_STATIC)
 
 guix-debug debug profile: LIBS += $(DLIBS_DEBUG)
 
-.PHONY: all guix guix-debug debug release static profile clean test
-
 all: debug
 
-htslib-static:
-	cd htslib && $(MAKE)
+# bio2/pileup.d: bio2/bam/reader.d
 
-ldc-version-info:
-	./gen_ldc_version_info.py $(shell which ldmd2) > utils/ldc_version_info_.d
+# sambamba/subsample.d: bio2/pileup.d
 
-utils/ldc_version_info_.o: ldc-version-info
-	$(D_COMPILER) $(DFLAGS) -c utils/ldc_version_info_.d -od=$(dir $@)
+utils/ldc_version_info_.d:
+	./gen_ldc_version_info.py $(shell which ldc2) > utils/ldc_version_info_.d
 
-build-setup: htslib-static ldc-version-info
-	mkdir -p build/
+utils/ldc_version_info_.o: utils/ldc_version_info_.d
+	$(D_COMPILER) $(DFLAGS) -c utils/ldc_version_info_.d -od=$(dir $@)
 
 guix guix-debug default debug release static profile: $(OUT)
 
@@ -65,26 +76,36 @@ guix guix-debug default debug release static profile: $(OUT)
 	$(D_COMPILER) $(DFLAGS) -c $< -od=$(dir $@)
 
 # ---- Link step
-$(OUT): build-setup $(OBJ)
-	$(D_COMPILER) $(DFLAGS) -of=build/sambamba $(OBJ) $(LIBS)
+$(OUT): $(OBJ)
+	cd htslib && $(MAKE)
+	mkdir -p bin/
+	$(D_COMPILER) $(DFLAGS) -of=bin/sambamba $(OBJ) $(LIBS)
 
-test:
+test:	clean-tests
 	./run_tests.sh
 
-check: test
+biod:
+	cd $(BIOD_PATH)/src_ragel && make
+
+check: all test
 
 debug-strip: debug
-	objcopy --only-keep-debug build/sambamba sambamba.debug
-	objcopy --strip-debug build/sambamba
-	objcopy --add-gnu-debuglink=sambamba.debug build/sambamba
-	mv sambamba.debug build/
+	objcopy --only-keep-debug bin/sambamba sambamba.debug
+	objcopy --strip-debug bin/sambamba
+	objcopy --add-gnu-debuglink=sambamba.debug bin/sambamba
+	mv sambamba.debug bin/
 
 install:
-	install -m 0755 build/sambamba $(prefix)/bin
+	install -m 0755 bin/sambamba $(prefix)/bin
 
-clean: clean-d
+clean-c:
 	cd htslib ; make clean
 
 clean-d:
-	rm -rf build/*
-	rm -f $(OBJ) $(OUT) trace.{def,log}
+	rm -v $(OBJ) $(OUT) trace.{def,log}
+	rm -v bin/*
+
+clean-tests:
+	rm -rf output/*
+
+clean: clean-d clean-tests

Makefile.old

+D_COMPILER=dmd
+D_FLAGS=--compiler=dmd -IBioD -IundeaD/src -g -d#-O -release -inline # -version=serial
+LDMD=ldmd2
+
+STATIC_LIB_PATH=-Lhtslib -Llz4/lib
+STATIC_LIB_SUBCMD=$(STATIC_LIB_PATH) -Wl,-Bstatic -lhts -llz4 -Wl,-Bdynamic
+RDMD_FLAGS=--force --build-only --compiler=$(D_COMPILER) $(D_FLAGS)
+
+PLATFORM := $(shell uname -s)
+
+ifeq "$(PLATFORM)" "Darwin"
+
+LINK_CMD=gcc -dead_strip -lphobos2-ldc -ldruntime-ldc -lm -lpthread htslib/libhts.a lz4/lib/liblz4.a build/sambamba.o -o build/sambamba
+DMD_STATIC_LIBS=htslib/libhts.a lz4/lib/liblz4.a
+
+define split-debug
+dsymutil build/sambamba -o build/sambamba.dSYM
+strip -S build/sambamba
+endef
+
+else
+
+LINK_CMD=gcc -Wl,--gc-sections -o build/sambamba build/sambamba.o $(STATIC_LIB_SUBCMD) -lphobos2-ldc -ldruntime-ldc  -lrt -lpthread -lm -ldl
+DMD_STATIC_LIBS=-L-Lhtslib -L-l:libhts.a -L-l:libphobos2.a -L-Llz4/lib -L-l:liblz4.a
+
+define split-debug
+objcopy --only-keep-debug build/sambamba sambamba.debug
+objcopy --strip-debug build/sambamba
+objcopy --add-gnu-debuglink=sambamba.debug build/sambamba
+mv sambamba.debug build/
+endef
+
+endif
+
+PREREQS := ldc-version-info htslib-static lz4-static
+
+# DMD only - this goal is used because of fast compilation speed, during development
+all: $(PREREQS)
+	mkdir -p build/
+	rdmd --force --build-only $(D_FLAGS) $(DMD_STATIC_LIBS) -ofbuild/sambamba main.d
+
+# This is the main Makefile goal, used for building releases (best performance)
+sambamba-ldmd2-64: $(PREREQS)
+	mkdir -p build/
+	$(LDMD) @sambamba-ldmd-release.rsp
+	$(LINK_CMD)
+	$(split-debug)
+
+# For debugging; GDB & Valgrind are more friendly to executables created using LDC/GDC than DMD
+sambamba-ldmd2-debug: $(PREREQS)
+	mkdir -p build/
+	$(LDMD) @sambamba-ldmd-debug.rsp
+	$(LINK_CMD)
+
+ldc-version-info:
+	./gen_ldc_version_info.py $(shell which $(LDMD)) > utils/ldc_version_info_.d
+
+htslib-static:
+	cd htslib && $(MAKE)
+
+lz4-static: lz4/lib/liblz4.a
+
+lz4/lib/liblz4.a: lz4/lib/lz4.c lz4/lib/lz4hc.c lz4/lib/lz4frame.c lz4/lib/xxhash.c
+	cd lz4/lib && $(CC) -O3 -c lz4.c lz4hc.c lz4frame.c xxhash.c && $(AR) rcs liblz4.a lz4.o lz4hc.o lz4frame.o xxhash.o
+
+# all below link to libhts dynamically for simplicity
+
+sambamba-flagstat:
+	mkdir -p build/
+	rdmd $(RDMD_FLAGS) -L-lhts -version=standalone -ofbuild/sambamba-flagstat sambamba/flagstat.d
+
+sambamba-merge:
+	mkdir -p build/
+	rdmd $(RDMD_FLAGS) -L-lhts -version=standalone -ofbuild/sambamba-merge sambamba/merge.d
+
+sambamba-index:
+	mkdir -p build/
+	rdmd $(RDMD_FLAGS) -L-lhts -version=standalone -ofbuild/sambamba-index sambamba/index.d
+
+sambamba-sort:
+	mkdir -p build/
+	rdmd $(RDMD_FLAGS) -L-lhts -version=standalone -ofbuild/sambamba-sort sambamba/sort.d
+
+sambamba-view:
+	mkdir -p build/
+	rdmd $(RDMD_FLAGS) -L-lhts -version=standalone -ofbuild/sambamba-view sambamba/view.d
+
+sambamba-slice:
+	mkdir -p build/
+	rdmd $(RDMD_FLAGS) -L-lhts -version=standalone -ofbuild/sambamba-slice sambamba/slice.d
+
+sambamba-markdup:
+	mkdir -p build/
+	rdmd $(RDMD_FLAGS) -L-lhts -version=standalone -ofbuild/sambamba-markdup sambamba/markdup.d
+
+sambamba-depth:
+	mkdir -p build/
+	rdmd $(RDMD_FLAGS) -L-lhts -version=standalone -ofbuild/sambamba-depth sambamba/depth.d
+
+sambamba-pileup:
+	mkdir -p build/
+	rdmd $(RDMD_FLAGS) -L-lhts -version=standalone -ofbuild/sambamba-pileup sambamba/pileup.d
+
+.PHONY: clean ldc-version-info
+
+clean:
+	rm -rf build/ ; $(MAKE) -C htslib clean ; $(MAKE) -C lz4 clean

README.md

+[![Build Status](https://travis-ci.org/biod/sambamba.svg?branch=master)](https://travis-ci.org/biod/sambamba) [![Anaconda-Server Badge](https://anaconda.org/bioconda/sambamba/badges/installer/conda.svg)](https://conda.anaconda.org/bioconda) [![DL](https://anaconda.org/bioconda/sambamba/badges/downloads.svg)](https://anaconda.org/bioconda/sambamba)
+
 # sambamba
-[![Anaconda-Server Badge](https://anaconda.org/bioconda/sambamba/badges/installer/conda.svg)](https://conda.anaconda.org/bioconda) [![DL](https://anaconda.org/bioconda/sambamba/badges/downloads.svg)](https://anaconda.org/bioconda/sambamba)
 
 ## Table of Contents
 
@@ -14,30 +15,50 @@
 <a name="intro"></a>
 # Introduction
 
-Sambamba is a high performance modern robust and fast tool (and
-library), written in the D programming language, for working with SAM
-and BAM files.  Current functionality is an important
-subset of samtools functionality, including view, index, sort,
-markdup, and depth. Most tools support piping: just specify `/dev/stdin`
-or `/dev/stdout` as filenames.
-
-For almost 5 years the main advantage over `samtools` was parallelized BAM reading.
-Finally in March 2017 `samtools` 1.4 was released, reaching parity on this.
-That said, we still have quite a few interesting features to offer:
-
-- faster `sort` (no benchmarks yet, sorry)
+Sambamba is a high performance highly parallel robust and fast tool
+(and library), written in the D programming language, for working with
+SAM and BAM files. Because of its efficiency is an important work
+horse running in many sequencing centres around the world
+today.
+
+Current functionality is an important subset of samtools
+functionality, including view, index, sort, markdup, and depth. Most
+tools support piping: just specify `/dev/stdin` or `/dev/stdout` as
+filenames. When we started writing sambamba (in 2012) the main
+advantage over `samtools` was parallelized BAM reading and writing.
+In March 2017 `samtools` 1.4 was released, reaching parity on this. A
+[recent performance comparison](https://github.com/guigolab/sambamBench-nf)
+shows that sambamba holds its ground and can do better in different
+configurations. Here are some comparison
+[metrics](https://public-docs.crg.es/rguigo/Data/epalumbo/sambamba_ws_report.html). For
+example for flagstat sambamba is 1.4x faster than samtools. For index
+they are similar. For Markdup almost 6x faster and for view 4x
+faster. For sort sambamba has been beaten generally, though sambamba
+is up to 2x faster on large RAM machines.
+
+In addition sambamba has a few interesting features to offer, in particular
+
+- faster large machine `sort`, see [performance](./test/benchmark/stats.org)
 - automatic index creation when writing any coordinate-sorted file
 - `view -L <bed file>` utilizes BAM index to skip unrelated chunks
 - `depth` allows to measure base, sliding window, or region coverages
   - [Chanjo](https://www.chanjo.co/) builds upon this and gets you to exon/gene levels of abstraction
 - `markdup`, a fast implementation of Picard algorithm
 - `slice` quickly extracts a region into a new file, tweaking only first/last chunks
+- and more
+
+Even though Sambamba started out as a samtools clone we are now in the
+process of adding new functionality - also in the
+[BioD project](https://github.com/biod/BioD). The D language is
+extremely suitable for high performance computing. At this point we
+think that the BAM format is here to stay for processing sequencing
+data and we aim to make it easy to parse and process BAM files.
 
 Sambamba is free and open source software, licensed under GPLv2+.
 See manual pages [online](https://lomereiter.github.io/sambamba/docs/sambamba-view.html)
 to know more about what is available and how to use it.
 
-For more information on Sambamba you can contact Artem Tarasov and Pjotr Prins.
+For more information on Sambamba contact the mailing list (see below).
 
 <a name="install"></a>
 # Binary installation
@@ -49,11 +70,11 @@ are Github source and binary
 the tarball, unpack it and run it. For example
 
 ```sh
-wget https://github.com/biod/sambamba/releases/download/v0.6.6/sambamba_v0.6.6_linux.tar.bz2
-tar xvjf sambamba_v0.6.6_linux.tar.bz2
-./sambamba_v0.6.6
+wget https://github.com/biod/sambamba/releases/download/v0.6.8/sambamba_v0.6.8_linux.tar.bz2
+tar xvjf sambamba_v0.6.8_linux.tar.bz2
+./sambamba_v0.6.8
 
-    sambamba 0.6.6
+    sambamba 0.6.8
 
         Usage: sambamba [command] [args...]
 
@@ -62,35 +83,6 @@ tar xvjf sambamba_v0.6.6_linux.tar.bz2
         To get help on a particular command, just call it without args.
 ```
 
-## Install latest pre-release
-
-A *latest* pre-release of sambamba 0.6.7 for Linux that includes debug
-information and *all* dependencies is available from this
-[link](http://test-gn2.genenetwork.org/ipfs/QmakasNfZhdbPA3xJYNxNX7at5FtYnS4hUNnvDbzxhZf2J). This
-24Mb download reflects the development edition and includes recent
-versions of libraries, samtools and bcftools. It should install on any
-Linux distribution, including old ones on HPC clusters.
-
-Install the tarball by unpacking it and running the contained install
-script with a target directory e.g.
-
-```sh
-wget http://test-gn2.genenetwork.org/ipfs/QmakasNfZhdbPA3xJYNxNX7at5FtYnS4hUNnvDbzxhZf2J/hb13hjys1064jmb6z17yc1f822hv9zsz-sambamba-0.6.7-pre1-7cff065-x86_64.tar.bz2
-tar xvjf QmakasNfZhdbPA3xJYNxNX7at5FtYnS4hUNnvDbzxhZf2J/hb13hjys1064jmb6z17yc1f822hv9zsz-sambamba-0.6.7-pre1-7cff065-x86_64.tar.bz2
-./install.sh ~/sambamba-test
-~/sambamba-test/bin/sambamba
-
-    sambamba 0.6.7-pre1
-
-    Usage: sambamba [command] [args...]
-
-        Available commands: 'view', 'index', 'merge', 'sort',
-                            'flagstat', 'slice', 'markdup', 'depth', 'mpileup'
-
-```
-
-Binaries are also available through the following packaging tools (note the version numbers):
-
 ## Bioconda install
 
 [![Install with CONDA](https://anaconda.org/bioconda/sambamba/badges/installer/conda.svg)](https://anaconda.org/bioconda/sambamba)
@@ -126,8 +118,7 @@ similar. Also try running the latest version of sambamba to make sure
 it has not been fixed already. Support/installation questions should
 be aimed at the mailing list. The issue tracker is for development
 issues around the software itself. When reporting an issue include the
-output of the program and the contents of the .log.txt file in the
-output directory.
+output of the program and the contents of the output directory.
 
 ## Check list:
 
@@ -139,7 +130,7 @@ output directory.
 6. [ ] If it is a support/install question I have posted it to the [mailing list](https://groups.google.com/forum/#!forum/sambamba-discussion)
 7. [ ] If it is software development related I have posted a new issue on the [issue tracker](https://github.com/biod/sambamba/issues) or added to an existing one
 8. [ ] In the message I have included the output of my sambamba run
-9. [ ] In the message I have included the relevant .log.txt file in the output directory
+9. [ ] In the message I have included the relevant files in the output directory
 10. [ ] I have made available the data to reproduce the problem (optional)
 
 To find bugs the sambamba software developers may ask to install a
@@ -168,35 +159,48 @@ which targets LLVM.
 
 ## Compilation dependencies
 
-- zlib
+- git (to check out the repo)
+- gcc compiler 4.9 or later (for htslib)
+- D compiler 1.7.0 or later (ldc2, see below)
+- python2 (parses D-compiler header for version info)
+- zlib (library)
+- lz4 (library)
 - htslib (submodule)
-- lz4 (submodule)
-- BioD (submodule)
-- undeaD (submodule)
+- BioD (source)
+- undeaD (source)
 - python2
 
 ## Compiling for Linux
 
-The LDC compiler's github repository also provides binary images. The current
-preferred release for sambamba is LDC - the LLVM D compiler (>= 1.1.0). After
-installing LDC:
+The LDC compiler's github repository provides binary images. The current
+preferred release for sambamba is LDC - the LLVM D compiler (>= 1.6.1). After
+installing LDC from https://github.com/ldc-developers/ldc/releases/ with, for example
+
+```sh
+cd
+wget https://github.com/ldc-developers/ldc/releases/download/v$ver/ldc2-1.7.0-linux-x86_64.tar.xz
+tar xvJf ldc2-1.7.0-linux-x86_64.tar.xz
+export PATH=$HOME/ldc2-1.7.0-linux-x86_64/bin:$PATH
+export LIBRARY_PATH=$HOME/ldc2-1.7.0-linux-x86_64/lib
+```
 
 ```sh
 git clone --recursive https://github.com/biod/sambamba.git
 cd sambamba
-    git clone https://github.com/dlang/undeaD
-    make sambamba-ldmd2-64
+make
 ```
 
-Installing LDC only means unpacking an archive and setting some
-environmental variables, e.g. unpacking into `$HOME`:
+To build a debug release run
 
 ```sh
-cd
-wget https://github.com/ldc-developers/ldc/releases/download/v$ver/ldc2-$ver-linux-x86_64.tar.xz
-tar xJf ldc2-$ver-linux-x86_64.tar.xz
-export PATH=~/ldc2-$ver-linux-x86_64/bin/:$PATH
-export LIBRARY_PATH=~/ldc2-$ver-linux-x86_64/lib/
+make clean && make debug
+```
+
+To run the test fetch shunit2 from https://github.com/kward/shunit2 and put it in the path so
+you can run
+
+```sh
+make check
 ```
 
 ### GNU Guix
@@ -209,6 +213,9 @@ guix package -i ldc
 
 ## Compiling for Mac OS X
 
+Note: the Makefile does not work. Someone want to fix that using the
+Makefile.old version? See also https://github.com/biod/sambamba/issues/338.
+
 ```sh
     brew install ldc
     git clone --recursive https://github.com/biod/sambamba.git
@@ -227,9 +234,16 @@ documentation](https://github.com/biod/sambamba-dev-docs).
 <a name="debug"></a>
 # Debugging and troubleshooting
 
+## Segfaults on certain Intel Xeons
+
+Important note: some popular Xeon processors segfault under heavy
+hyper threading - which Sambamba utilizes.  Please read
+[this](https://blog.cloudflare.com/however-improbable-the-story-of-a-processor-bug/)
+when encountering seemingly random crashes.
+
 ## Dump core
 
-In a crash sambamba can dump a core. To make this happen set
+In a crash sambamba can dump a core file. To make this happen set
 
 ```sh
 ulimit -c unlimited
@@ -274,7 +288,8 @@ the tarball and run the contained install.sh script with TARGET
 Run sambamba in gdb with
 
 ```
-gdb --args ~/sambamba-test/sambamba-*/bin/sambamba view --throw-error
+gdb -ex 'handle SIGUSR1 SIGUSR2 nostop noprint' \
+  --args ~/sambamba-test/sambamba-*/bin/sambamba view --throw-error
 ```
 
 <a name="license"></a>

RELEASE-NOTES.md

+## ChangeLog v0.6.8-pre1 (20180207)
+
+Minor release with a much faster binary. 10-20% faster than v0.6.6,
+due to ldc and LLVM improvements. Fixes speed regression of v0.6.7 for
+large files. See also [performance](https://github.com/biod/sambamba/blob/master/test/benchmark/stats.org)
+
++ Fixed Makefile for general use, see #332
++ Started benchmarking, see #283 and https://github.com/biod/sambamba/blob/master/test/benchmark/stats.org
++ Readded [Travis-ci support](https://travis-ci.org/biod/sambamba) for Linux (MacOS is disabled #338)
++ Updated BioD to latest https://github.com/biod/BioD/commit/5e56b2bb45324af2194b3339d298fd827c8003ae
++ Bug fixes:
+  * #328 Debug version: SAM output of CRAM file is populated with debug on pipe
+  * #331 Segmentation fault attempting to view header in json format
+  * #335 Intel Xeon bug may segfault Sambamba - this was tracked down to an Intel Xeon bug
++ Documentation updates

cram/wrappers.d

@@ -23,13 +23,13 @@ struct RcPtr(T, alias Free) {
             this.ptr = ptr;
             debug {
                 payload_id = ++payload_counter;
-                stderr.writeln("Init ", T.stringof, "* #", payload_id);
+                // stderr.writeln("Init ", T.stringof, "* #", payload_id);
             }
         }
 
         ~this() {
             debug {
-                stderr.writeln("Free ", T.stringof, "* #", payload_id);
+                // stderr.writeln("Free ", T.stringof, "* #", payload_id);
             }
             Free(ptr);
         }
@@ -51,7 +51,7 @@ struct RcPtr(T, alias Free) {
     this(this)
     {
         static if (is(T == cram_slice)) {
-            debug writeln("COPIED #", data.payload_id + 1);
+            // debug stderr.writeln("COPIED #", data.payload_id + 1);
         }
     }
 
@@ -159,7 +159,7 @@ struct CramContainerRange {
         auto err_msg = "Failed to read container header";
         while (true) {
             // read container header
-            debug writeln("cram_read_container");
+            // debug stderr.writeln("cram_read_container");
             auto ptr = nullChecked!cram_read_container(err_msg, _fd);
             if (ptr is null) {
                 empty = true;
@@ -240,8 +240,8 @@ class UndecodedSliceRange {
         _container.curr_slice++;
 
         auto err_msg = "Failure in cram_read_slice";
-        debug stderr.writeln("cram_read_slice (", _container.curr_slice,
-                             "/", _container.max_slice, ")");
+        // debug stderr.writeln("cram_read_slice (", _container.curr_slice,
+        //                      "/", _container.max_slice, ")");
         auto ptr = cram_read_slice(_fd);
         if (ptr is null) {
             throw new CramException(err_msg);

debian/changelog

-sambamba (0.6.7-3) UNRELEASED; urgency=medium
+sambamba (0.6.8-pre3-1) UNRELEASED; urgency=medium
 
+  * New upstream pre-release
+    Closes: #907489
   * Point Vcs fields to salsa.debian.org
   * Standards-Version: 4.2.1
-  * Fix normalize(weight) (applied patch from upstream)
-    Should close #907489 but it does not work - wait for new upstream release
   * Add build time test
+  * Allow pre-releases in watch file
 
- -- Andreas Tille <tille@debian.org>  Tue, 28 Aug 2018 20:07:55 +0200
+ -- Andreas Tille <tille@debian.org>  Sun, 23 Sep 2018 07:42:58 +0200
 
 sambamba (0.6.7-2) unstable; urgency=medium
 

debian/patches/01_add_meson.patch

@@ -10,7 +10,7 @@ Subject: [PATCH] Add Meson build file
 
 --- /dev/null
 +++ b/meson.build
-@@ -0,0 +1,135 @@
+@@ -0,0 +1,138 @@
 +project('Sambamba', 'd')
 +
 +project_version   = '0.6.6'
@@ -31,10 +31,13 @@ Subject: [PATCH] Add Meson build file
 +    'sambamba/flagstat.d',
 +    'sambamba/index.d',
 +    'sambamba/markdup.d',
++    'sambamba/markdup2.d',
 +    'sambamba/merge.d',
 +    'sambamba/pileup.d',
 +    'sambamba/slice.d',
 +    'sambamba/sort.d',
++    'sambamba/subsample.d',
++    'sambamba/validate.d',
 +    'sambamba/utils/common/bed.d',
 +    'sambamba/utils/common/file.d',
 +    'sambamba/utils/common/filtering.d',
@@ -114,7 +117,7 @@ Subject: [PATCH] Add Meson build file
 +# Targets
 +#
 +sambamba_exe = executable('sambamba',
-+    ['main.d',
++    ['sambamba/main.d',
 +     sambamba_src,
 +     utils_src,
 +     cram_src,

debian/patches/03_fix_normalize.patch

-From: Pjotr Prins <pjotr.public01@thebird.nl>
-Origin: https://github.com/biod/sambamba/commit/cb170d641c21f5aabeb04cedab3ced5b7262d007
-Date: Fri, 27 Jul 2018 23:28:29 +0000
-Bug-Debian: https://bugs.debian.org/907489
-Subject: [PATCH] Fixes normalize(weight)
-
-/gnu/store/4snsi4vg06bdfi6qhdjfbhss16kvzxj7-ldc-1.10.0/include/d/std/numeric.d(1845):
-Error: read-modify-write operations are not allowed for shared variables. Use core.atomic.atomicOp!"+="(s, e) instead.
----
- Makefile         | 2 +-
- Makefile.guix    | 6 +++---
- sambamba/merge.d | 8 ++++++--
- 3 files changed, 10 insertions(+), 6 deletions(-)
-
---- a/Makefile
-+++ b/Makefile
-@@ -1,5 +1,5 @@
- D_COMPILER=dmd
--D_FLAGS=--compiler=dmd -IBioD -IundeaD/src -g -d#-O -release -inline # -version=serial
-+D_FLAGS=--compiler=dmd -I../BioD -IBioD -IundeaD/src -g -d#-O -release -inline # -version=serial
- LDMD=ldmd2
- 
- STATIC_LIB_PATH=-Lhtslib -Llz4/lib
---- a/sambamba/merge.d
-+++ b/sambamba/merge.d
-@@ -1,6 +1,7 @@
- /*
-     This file is part of Sambamba.
-     Copyright (C) 2012-2016    Artem Tarasov <lomereiter@gmail.com>
-+    Copyright (C) 2012-2017    Pjotr Prins <pjotr.prins@thebird.nl>
- 
-     Sambamba is free software; you can redistribute it and/or modify
-     it under the terms of the GNU General Public License as published by
-@@ -403,8 +404,11 @@ int merge_main(string[] args) {
-             alias ReturnType!(BamReader.readsWithProgress!withoutOffsets) AlignmentRangePB;
-             auto alignmentranges_with_file_ids = new Tuple!(AlignmentRangePB, size_t)[files.length];
- 
--            auto weights = cast(shared)array(map!(pipe!(getSize, to!float))(filenames));
--            normalize(cast()weights);
-+            // auto weights = cast(shared)array(map!(pipe!(getSize, to!float))(filenames));
-+            auto weights1 = array(map!(pipe!(getSize, to!float))(filenames));
-+            normalize(weights1);
-+            // auto weights = cast(shared)weights1;
-+            immutable weights = cast(immutable)weights1;
- 
-             foreach (i; 0 .. files.length) {
-                 alignmentranges_with_file_ids[i] = tuple(

debian/patches/series

 01_add_meson.patch
 disable-assert.patch
-03_fix_normalize.patch

debian/watch

 version=4
 
-https://github.com/lomereiter/sambamba/releases .*/archive/v(\d[\d.-]+)\.(?:tar(?:\.gz|\.bz2)?|tgz)
+https://github.com/lomereiter/sambamba/releases .*/archive/v(\d[\d.pre-]+)@ARCHIVE_EXT@

doc/design.org

+* Sambamba new design
+
+** Introduction
+
+Because of its great multi-core performance Sambamba has served over
+five years in sequencing centers [[https://groups.google.com/d/msg/sambamba-discussion/fIgrrUa441o/XG7Rt3dFAQAJ][around the world]]. Here we start on a
+new design (sambamba2) that should improve performance and, perhaps
+more importantly, make the building blocks more composable. D has
+proven to be a great language for multi-core performance and code
+clarity, so we are happy to build on its latest language features. For
+example, see this [[http://forum.dlang.org/thread/gvtjhpxdqpboppoodmxm@forum.dlang.org][dicussion]] on streaming data, which, if you read it
+carefully, suggests we should use multiple implementations for file
+access.
+
+Here we document some of the choices we are making for the new
+design. Starting with markdup2, the new version is a prototype for new
+sambamba architecture using more canonical D language features,
+including immutable and improved laziness and a more functional
+programming style. It should provide improved performance and minimize
+RAM use, as well as better composability. Also we are preparing it for
+CRAM input.
+
+Another point to consider is chunking and built-in parallelism which
+may give different results on different hardware platforms. When an
+algorithm allows for it and there is enough RAM we should allow for
+chunking file and running chunks in parallel. Both current sambamba
+and samtools provide a taskpool, but have little options for tuning
+for hardware variations. Other hardware considerations are providing
+the latest LLVM support for just in time (JIT) compilation and GPUs.
+
+** Streaming input data
+
+Every input file should be on its own streaming thread. If there is
+only one input file, it should be possible to use a stdin pipe. When
+using a pipe we will pass reads in an uncompressed format (it makes no
+sense to compress and decompress in a pipe). This means we should be
+able to do something like this on a sorted BAM file:
+
+#+BEGIN_SRC bash
+sambamba unpack in.bam | sambamba markdup2 |sambamba pack -f bam > out.bam
+#+END_SRC
+
+Because of the pipes these tools can run in parallel.
+
+** Streaming output data
+
+The main thread is the output thread and writes data to stdout, or
+optionally a file. Compression may be handed out to other threads.
+
+** Composability
+
+Composability happens at two levels. First at the tool level. By
+providing support for Unix pipes we'll make sambamba easier to plug
+into other solutions. In particular we are interested in providing
+support for other programming languages, such as Python and Ruby.
+
+The second composability level is within the code base. It should be
+easy to plug in different file readers, for example. Using a more
+functional programming style and getting rid of (deprecated)
+std.stream should help there.
+
+** Tooling
+
+*** Metrics
+
+To get the best performance having metrics is extremely
+important. Luckily LLVM provides great tooling for metrics and we
+should use that.
+
+*** Logging
+
+Because of the pipes it is crucial sambamba gives clear error messages
+and should be capable of writing to a log file.

dub.json

+{
+  "name": "sambamba",
+  "targetType": "none",
+  "authors": [
+    "pjotrp"
+  ],
+  "dependencies": {
+    "undead": "~>1.0.9"
+  },
+  "description": "Sambamba",
+  "copyright": "Copyright © 2012-2017, Artem Tarasov and Pjotr Prins",
+  "license": "GPL3",
+}

run_tests.sh

@@ -3,4 +3,4 @@
 # download shunit2 in order to run tests:
 # curl -L "https://dl.dropboxusercontent.com/u/7916095/shunit2-2.0.3.tgz" | tar zx --overwrite
 
-./test_suite.sh | tee /dev/stderr | grep -q 'success rate: 100%'
+./test/test_suite.sh | tee /dev/stderr | grep -q 'success rate: 100%'

main.d → sambamba/main.d

 /*
     This file is part of Sambamba.
-    Copyright (C) 2012-2016    Artem Tarasov <lomereiter@gmail.com>
+    Copyright (C) 2012-2017    Artem Tarasov <lomereiter@gmail.com>
+    Copyright (C) 2012-2018    Pjotr Prins <pjotr.prins@thebird.nl>
 
     Sambamba is free software; you can redistribute it and/or modify
     it under the terms of the GNU General Public License as published by
@@ -17,16 +18,22 @@
     Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA  02111-1307  USA
 
 */
-import sambamba.view;
+
+import std.experimental.logger;
+
+import sambamba.depth;
 import sambamba.index;
-import sambamba.merge;
-import sambamba.sort;
+import sambamba.fixbins;
 import sambamba.flagstat;
-import sambamba.slice;
 import sambamba.markdup;
-import sambamba.depth;
+import sambamba.markdup2;
+import sambamba.merge;
 import sambamba.pileup;
-import sambamba.fixbins;
+import sambamba.sort;
+import sambamba.slice;
+import sambamba.subsample;
+import sambamba.validate;
+import sambamba.view;
 
 import sambamba.utils.common.ldc_gc_workaround;
 
@@ -37,32 +44,53 @@ import utils.ldc_version_info_ : LDC_VERSION_STRING, DMD_VERSION_STRING, LLVM_VE
 
 import std.stdio;
 
+
 void printUsage() {
-    stderr.writeln("sambamba " ~ VERSION);
-    stderr.writeln();
-    stderr.writeln("Usage: sambamba [command] [args...]");
-    stderr.writeln();
-    stderr.writeln("    Available commands: 'view', 'index', 'merge', 'sort',");
-    stderr.writeln("                        'flagstat', 'slice', 'markdup', 'depth', 'mpileup'");
-    stderr.writeln("    To get help on a particular command, just call it without args.");
-    stderr.writeln();
-    stderr.writeln("Leave bug reports and feature requests at");
-    stderr.writeln("https://github.com/lomereiter/sambamba/issues");
-    stderr.writeln();
+    stderr.writeln("
+Usage: sambamba [command] [args...]
+
+  Available commands:
+
+    view        view contents and convert from one format
+                to another (SAM/BAM/CRAM/JSON/UNPACK)
+    index       build index (BAI)
+    merge       merge files (BAM)
+    sort        sort file (BAM)
+    slice       slice file (BAM using BED)
+    markdup     mark or remove duplicates (BAM)
+    subsample   subsamble (BAM)
+    flagstat    output statistics (BAM)
+    depth       output statistics (BAM)
+    validate    simple validator (BAM)
+
+  Work in progress (WIP):
+
+    markdup2    mark or remove duplicates v2 (BAM)
+
+  No longer recommended:
+
+    mpileup     parallel execution of samtools (BAM)
+
+To get help on a particular command, call it without args.
+
+For bug reports and feature requests see
+
+       https://github.com/biod/
+");
 }
 
 void printVersion() {
-    stderr.writeln("sambamba " ~ VERSION);
     stderr.writeln();
-    stderr.writeln("This version was built with:");
-    stderr.writeln("    LDC " ~ LDC_VERSION_STRING);
-    stderr.writeln("    using DMD " ~ DMD_VERSION_STRING);
-    stderr.writeln("    using LLVM " ~ LLVM_VERSION_STRING);
-    stderr.writeln("    bootstrapped with " ~ BOOTSTRAP_VERSION_STRING);
+    stderr.writeln("sambamba " ~ VERSION ~ " by Artem Tarasov and Pjotr Prins (C) 2012-2017");
+    stderr.writeln("    LDC " ~ LDC_VERSION_STRING ~ " / DMD " ~ DMD_VERSION_STRING ~
+     " / LLVM" ~ LLVM_VERSION_STRING ~ " / bootstrap " ~ BOOTSTRAP_VERSION_STRING);
     stderr.writeln();
 }
 
 int main(string[] args) {
+    globalLogLevel(LogLevel.info);
+    printVersion();
+
     if (args.length == 1) {
         printUsage();
         return 1;
@@ -77,9 +105,12 @@ int main(string[] args) {
         case "sort":      return sort_main(_args);
         case "flagstat":  return flagstat_main(_args);
         case "slice":     return slice_main(_args);
-        case "markdup":  return markdup_main(_args);
+        case "markdup":   return sambamba.markdup.markdup_main(_args);
+        case "markdup2":  return sambamba.markdup2.markdup_main(_args);
+        case "subsample": return subsample_main(_args);
         case "depth":     return depth_main(_args);
         case "mpileup":   return pileup_main(_args);
+        case "validate":  return validate_main(_args);
 
         // hidden commands
         case "fixbins":   return fixbins_main(_args);