Feature/bsd compilation

.github/workflows/c-cpp.yml

@@ -2,49 +2,89 @@ name: C/C++ CI
 
 on:
   push:
-    branches: [ master, develop, 'release/**' ]
+    branches: [master, develop, "release/**", "feature/**"]
   pull_request:
-    branches: [ master, develop  ]
+    branches: [master, develop]
 
 jobs:
   build:
     runs-on: ubuntu-latest
     steps:
-    - uses: actions/checkout@v3
-    - name: dependencies
-      run: |
-        sudo apt-get update
-        sudo DEBIAN_FRONTEND=noninteractive apt-get install -y yaggo gettext swig python3-dev ruby-dev libperl-dev
-    - name: autotools
-      run: autoreconf -fi
-    - name: configure
-      run: ./configure --enable-all-binding --enable-swig
-
-    - name: make
-      run: make -j$(nproc)
-
-    - name: make check
-      run: make -j$(nproc) check
-    - name: Check logs
-      if: failure()
-      uses: actions/upload-artifact@v4
-      with:
-        name: checklog
-        path: "**/*tests/*.log"
-
-    - name: make distcheck
-      run: make -j$(nproc) distcheck
-    - name: Distcheck logs
-      if: failure()
-      uses: actions/upload-artifact@v4
-      with:
-        name: distchecklog
-        path: |
-          **/*tests/*.log
-          **/*tests/**/tee.*
-
-    - name: Distribution tar ball
-      uses: actions/upload-artifact@v4
-      with:
-        name: disttarball
-        path: "mummer-*.tar.gz"
+      - uses: actions/checkout@v3
+      - name: dependencies
+        run: |
+          sudo apt-get update
+          sudo DEBIAN_FRONTEND=noninteractive apt-get install -y yaggo gettext swig python3-dev ruby-dev libperl-dev
+      - name: autotools
+        run: autoreconf -fi
+      - name: configure
+        run: ./configure --enable-all-binding --enable-swig
+
+      - name: make
+        run: make -j$(nproc)
+
+      - name: make check
+        run: make -j$(nproc) check
+      - name: Check logs
+        if: failure()
+        uses: actions/upload-artifact@v4
+        with:
+          name: checklog
+          path: "**/*tests/*.log"
+
+      - name: make distcheck
+        run: make -j$(nproc) distcheck
+      - name: Distcheck logs
+        if: failure()
+        uses: actions/upload-artifact@v4
+        with:
+          name: distchecklog
+          path: |
+            **/*tests/*.log
+            **/*tests/**/tee.*
+
+      - name: Distribution tar ball
+        uses: actions/upload-artifact@v4
+        with:
+          name: disttarball
+          path: "mummer-*.tar.gz"
+
+  testbsd:
+    runs-on: ubuntu-latest
+    name: Compile on FreeBSD
+    steps:
+      - uses: actions/checkout@v4
+      - name: Test in FreeBSD
+        id: test
+        uses: vmactions/freebsd-vm@v1
+        with:
+          usesh: true
+          prepare: |
+            pkg install -y gmake yaggo autoconf automake libtool bash gcc14
+          run: |
+            autoreconf -fi && \
+            ./configure MAKE=gmake CC=gcc14 CXX=g++14 LDFLAGS=-Wl,-rpath=/usr/local/lib/gcc14 && \
+            gmake -j $(sysctl -n hw.ncpu) check
+
+  testmacos:
+    runs-on: macos-latest
+    steps:
+      - uses: actions/checkout@v3
+      - name: dependencies
+        run: |
+          brew install autoconf automake libtool md5sha1sum
+          gem install yaggo
+      - name: autotools
+        run: autoreconf -fi
+      - name: configure
+        run: ./configure CC=gcc-14 CXX=g++-14
+      - name: make
+        run: make -j$(sysctl -n hw.ncpu)
+      - name: make check
+        run: make -j$(sysctl -n hw.ncpu) check
+      - name: Check logs
+        if: failure()
+        uses: actions/upload-artifact@v4
+        with:
+          name: checklog
+          path: "**/*tests/*.log"

ChangeLog

@@ -1,5 +1,16 @@
 -- Version history --
 
+4.0.0  48bit suffix array. Go past previous sequence length limitations
+       nucmer is all C++ (not Perl script) and multi-threaded
+       nucmer can save the SA index for later reuse
+       Added SAM output format to nucmer
+       Moved to autoconf compilation system
+       Bindings to scripting language: Python, Perl, Ruby
+       Improved mummerplot
+       Better unittesting and conformance testing
+       Provide a library with pkg-conig configuration
+       Moved to artistic license 2.0
+       Moved to C++17 standard.
 3.23 - Added -D and --banded option to nucmer. These options can be used to
        specify the largest indel that will be included in an alignment
        segment without breaking it in two pieces.

INSTALL.md

@@ -1,48 +1,91 @@
-# MUMmer4 INSTALLATION README
+# MUMmer4 Compilation README
 
 ## Dependencies
 
-If compiling from a [release source tarball](../../releases) you need a
-recent version of the GCC compiler (g++ version >= 4.7) and other
-essential tools for compilation (GNU make, ar, etc. Install
-`build-essentials` on Debian or Ubuntu derivatives).  Additional
-requirements are needed to compile the SWIG script bindings. See the
-[SWIG installation guide](swig/INSTALL.md).
+If compiling from a [release source tarball](../../releases) (recommended), you need a recent version of the GCC compiler (see below, only GCC is supported) and other essential tools for compilation (GNU make, ar, etc).
+Additional requirements are needed to compile the SWIG script bindings.
+See the [SWIG installation guide](swig/INSTALL.md).
 
-If compiling from the github development tree, additionally you need autotools (autoconf, automake and libtools),
-[yaggo](https://github.com/gmarcais/yaggo/releases).
-You should compile from a [release source tarball](../../releases), unless you plan on modifying the code of MUMmer.
+If compiling from the github development tree, additionally you need autotools (autoconf, automake and libtools), [yaggo](https://github.com/gmarcais/yaggo/releases).
 
-On Ubuntu:
+### On Ubuntu
 
+From the tarball:
 ```Shell
-sudo apt install git build-essential yaggo autoconf automake libtool gettext
+sudo apt install build-essential
 # For the bindings to scripting, additionally install
 sudo apt install swig python3-dev ruby-dev libperl-dev
 ```
 
+From the git tree:
+```Shell
+sudo apt instaoo build-essential git yaggo autoconf automake libtool gettext
+# For the bindings to scripting, additionally install
+sudo apt install swig python3-dev ruby-dev libperl-dev
+```
+
+### On Mac OS
+
+MUMmer must be compiled with GCC, not Clang (and not the Apple provided `gcc` which is really `clang`).
+Install with Brew:
+
+```Shell
+brew install autoconf automake libtool md5sha1sum
+gem install yaggo
+```
+
+### On FreeBSD
+
+MUMmer must be compiled with GCC, not Clang.
+Install with Brew:
+
+```Shell
+brew install autoconf automake libtool md5sha1sum bash
+gem install yaggo
+```
+
+
 ## Compilation & Installation
 
-To compile and install from a [release source tarball](../../releases):
+If compiling from the release tarball (recommended), then the first command `autoreconf -fi` is not necessary.
+
+### On Ubuntu
 
 ```Shell
+autoreconf -fi # Optional, on if compiling from git tree
 ./configure --prefix=/path/to/installation
 make
+make check # Optional
 make install
 ```
 
-If `--prefix` is omitted, the software is installed in
-`/usr/local`. One may need `sudo make install` if installing in a
-system location.
+If `--prefix` is omitted, the software is installed in `/usr/local`.
+One may need `sudo make install` if installing in a system location.
+
+### On MacOS
+
+Compile with `gcc-14`.
 
-To compile from the github tree, `autoreconf` must additionally be run:
 ```Shell
-autoreconf -fi
-./configure --prefix=/path/to/installation
+autoreconf -fi # Optional, on if compiling from git tree
+./configure --prefix=/path/to/installation CC=gcc-14 CXX=g++-14
 make
+make check # Optional
 make install
 ```
 
+### On FreeBSD
+
+Compile with `gcc14`.
+
+```Shell
+autoreconf -fi # Optional, on if compiling from git tree
+./configure --prefix=/path/to/installation MAKE=gmake CC=gcc14 CXX=g++14 LDFLAGS=-Wl,-rpath=/usr/local/lib/gcc14
+gmake
+gmake check # Optional
+gmake install
+```
+
 ## SOFTWARE REQUIREMENTS
 
 The MUMmer4.x package requires the following to run successfully. In

include/thread_pipe.hpp

@@ -48,6 +48,7 @@ class ostream_buffered : public consumer<ostream_buffered, stringstream_wrapper>
   std::ostream& os_;
 public:
   ostream_buffered(std::ostream& os) : os_(os) { }
+  ~ostream_buffered() { close(); }
   bool operator()(stringstream_wrapper& e) {
     bool res = false;
     auto rdbuf = e.p_->rdbuf();

src/tigr/show-aligns.cc

@@ -696,7 +696,8 @@ void append(ColoredBuffer& Buff1, ColoredBuffer& Buff2, std::string &Buff3,
 
 void add_prefix(ColoredBuffer& Buff, long int pos, long int seqlen, int frame) {
   char b[LINE_BUFFER_LEN + 1];
-  sprintf(b, PREFIX_FORMAT, toFwd(pos, seqlen, frame));
+  snprintf(b, sizeof(b), PREFIX_FORMAT, toFwd(pos, seqlen, frame));
+  b[sizeof(b)-1] = '\0';
   Buff.clear();
   Buff += b;
 }

tests/check_cigar.cc

@@ -6,13 +6,12 @@
 
 typedef std::map<std::string, std::string> name_seq_map;
 
-name_seq_map read_fasta(const char* path) {
+name_seq_map read_fasta(std::istream& is) {
   name_seq_map res;
-  std::ifstream is(path);
 
   int c = is.peek();
   if(c != '>')
-    check_cigar_cmdline::error() << "Invalid fasta file '" << path << '\'';
+    check_cigar_cmdline::error() << "Invalid fasta file";
   std::string line;
   while(c != EOF) {
     std::getline(is, line);
@@ -26,6 +25,48 @@ name_seq_map read_fasta(const char* path) {
   return res;
 }
 
+name_seq_map read_fasta(const char* path) {
+  std::ifstream is(path);
+  return read_fasta(is);
+}
+
+name_seq_map read_fastq(std::istream& is) {
+  name_seq_map res;
+
+  std::string line;
+  int c = is.peek();
+  while(c != EOF) {
+      std::getline(is, line);
+      if(c != '@')
+        check_cigar_cmdline::error() << "Invalid fastq file: " << line;
+      std::string& seq = res[line.substr(1, line.find_first_of(" \t", 1) - 1)];
+      for(c = is.peek(); c != EOF && c != '+'; c = is.peek()) {
+        std::getline(is, line);
+        seq += line;
+      }
+      if(c == '+') {
+          std::getline(is, line);
+          std::getline(is, line);
+          c = is.peek();
+      }
+  }
+
+  return res;
+}
+
+name_seq_map read_seq(const char* path) {
+  std::ifstream is(path);
+  switch(is.peek()) {
+    case '>': return read_fasta(is);
+    case '@': return read_fastq(is);
+    default:
+      check_cigar_cmdline::error() << "Invalid sequence file '" << path << '\'';
+  }
+  // should never get there
+  exit(1);
+}
+
+
 char comp(char c) {
   switch(c) {
   case 'a': case 'A': return 't';
@@ -97,7 +138,7 @@ long find_nm(const std::vector<std::string>& fields) {
 int main(int argc, char *argv[]) {
   check_cigar_cmdline args(argc, argv);
   auto ref_map = read_fasta(args.ref_arg);
-  auto qry_map = read_fasta(args.qry_arg);
+  auto qry_map = read_seq(args.qry_arg);
 
   std::ifstream is(args.sam_arg);
   if(!is.good())

tests/gcc10_uniform_dist.hpp

@@ -36,6 +36,10 @@
 #ifndef _GCC10_UNIFORM_DIST_H
 #define _GCC10_UNIFORM_DIST_H
 
+#ifndef __glibcxx_assert
+#define __glibcxx_assert(x)
+#endif
+
 #include <type_traits>
 #include <limits>
 

tests/genome.sh

@@ -1,2 +1,2 @@
-time  sed -e 's/^>\([^[:space:]]\+\).*/>\1/' $D/seed_reads_2.fa | tee genome | nucmer -G --delta /dev/stdout $D/seed_genome.fa /dev/stdin | \
+time  sed -E 's/^>([^[:space:]]+).*/>\1/' "${D}/seed_reads_2.fa" | tee genome | nucmer -G --delta /dev/stdout "${D}/seed_genome.fa" /dev/stdin | \
     tee genome.delta | tail -n +3 | test_md5 8328b1577d8656eaa53aa61a113d89b0

tests/mummer.sh

@@ -1,2 +1,2 @@
-mummer -mum $D/seed_reads_1.fa $D/seed_reads_0.fa | ufasta hsort -H | ufasta dsort | test_md5 4e8182c9f745abf59158f69a05b942f3
-mummer -maxmatch $D/seed_reads_1.fa $D/seed_reads_0.fa | ufasta hsort -H | ufasta dsort | test_md5 a459f93742d1c36819e53e7a4c128bf7
+mummer -mum "${D}/seed_reads_1.fa" "${D}/seed_reads_0.fa" | ufasta hsort -H | ufasta dsort | test_md5 4e8182c9f745abf59158f69a05b942f3
+mummer -maxmatch "${D}/seed_reads_1.fa" "${D}/seed_reads_0.fa" | ufasta hsort -H | ufasta dsort | test_md5 a459f93742d1c36819e53e7a4c128bf7

tests/sam.sh

@@ -16,7 +16,7 @@ for i in 1 2; do
     [ "$(grep '^@PG' $f | head -c $pglen)" = "$pgline" ]
 
     # Test sequence headers
-    diff -q <(grep '^@SQ' $f) <(ufasta sizes -H $D/seed_reads_1.fa | sed 's/\([0-9]\+\) \([0-9]\+\)/@SQ\tSN:\1\tLN:\2/')
+    diff -q <(grep '^@SQ' $f) <(ufasta sizes -H $D/seed_reads_1.fa | sed -E 's/([0-9]+) ([0-9]+)/@SQ\tSN:\1\tLN:\2/')
 
     # Test that samtools can parse our output
     if [ -n "$SAMTOOLS" ]; then

tests/test_md5

@@ -1,4 +1,9 @@
-#! /bin/bash
+#! /usr/bin/env bash
 
-md5sum -c <(echo "$1  -") 2>&1 || \
-    test $(md5) == "$1" 2>&1
+# The semantic of md5sum differs in some subtle way. Darwin doesn't support
+# /dev/stdin, FreeBSD doesn't support "-". Linux supports both.
+
+case "$(uname)" in
+  (Darwin) exec md5sum -c <(echo "$1  -") 2>&1 ;;
+  (*) exec md5sum -c <(echo "$1  /dev/stdin") 2>&1 ;;
+esac