Merge tag '1.0' into debian

1.0

5 files changed, 234 insertions, 125 deletions

diff --git a/Makefile b/Makefile
index a55a9dd..e8f40f6 100644
--- a/Makefile
+++ b/Makefile
@@ -7,7 +7,7 @@
 # the Free Software Foundation, either version 3 of the License, or
 # (at your option) any later version.
 #
-# Foobar is distributed in the hope that it will be useful,
+# hadori is distributed in the hope that it will be useful,
 # but WITHOUT ANY WARRANTY; without even the implied warranty of
 # MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
 # GNU General Public License for more details.
@@ -15,9 +15,9 @@
 # You should have received a copy of the GNU General Public License
 # along with hadori.  If not, see <http://www.gnu.org/licenses/>.
 
-LDFLAGS+=-lz -lboost_program_options
+LDLIBS=-lboost_program_options -lstdc++
 CXXFLAGS?=-O2 -Wall
-CXXFLAGS+=-std=c++0x
+CXXFLAGS+=-std=c++11
 CPPFLAGS+=-D_FILE_OFFSET_BITS=64
 
 all: hadori.1
@@ -26,12 +26,13 @@ hadori.1: hadori
 	help2man -n $< -o $@ -N ./$<
 
 hadori: hadori.o
-hadori.o: hadori.C inode.h version.h
+hadori.o: hadori.C version.h
 
 version.h:
 	test ! -d .git || git describe | sed 's/^\(.*\)$$/#define HADORI_VERSION "hadori \1"/' > $@
 
 clean:
 	rm -f hadori hadori.o hadori.1
+	test ! -d .git || git checkout -f -- version.h
 
 .PHONY: version.h clean
diff --git a/TODO b/TODO
deleted file mode 100644
index 7019419..0000000
--- a/TODO
+++ /dev/null
@@ -1,5 +0,0 @@
-TODO/possible optimizations
-===========================
-
-* more debug output when attributes mismatch in handle_file()
-* perhaps use hash_map, hash_multimap for better performance?
diff --git a/changelog b/changelog
new file mode 100644
index 0000000..c2e8b94
--- /dev/null
+++ b/changelog
@@ -0,0 +1,183 @@
+commit 75293a34994df27b84ee30d94b8db625b0ca40ea
+Author: Timo Weingärtner <timo@tiwe.de>
+Date:   2014-11-23 01:06:29 +0100
+
+    more const
+
+commit 77bcb2f6d8d0afa0465dd921fa2e7fb64f1433b6
+Author: Timo Weingärtner <timo@tiwe.de>
+Date:   2014-11-21 20:30:34 +0100
+
+    use boost::iterator_range instead of homebrew hack
+
+commit 3c42fb81eb2cc9e658afc2156afa5d62880f3c7e
+Author: Timo Weingärtner <timo@tiwe.de>
+Date:   2013-11-17 14:38:02 +0100
+
+    Makefile: reset version.h on clean when in git
+
+commit 2d2195cf1caaac1c5a3b081fc1a49d1f7c12bc96
+Author: Timo Weingärtner <timo@tiwe.de>
+Date:   2013-11-17 14:36:18 +0100
+
+    Makefile: C++11 has been ready for a while now
+
+commit 38c62a8019f1a19ec93ea16d34fa2f05c702a564
+Author: Timo Weingärtner <timo@tiwe.de>
+Date:   2013-11-17 14:33:18 +0100
+
+    Makefile: add libstdc++ to LDLIBS
+
+commit 75b33a7f64d2f04efeeb09f96e9b6be716a65526
+Author: Timo Weingärtner <timo@tiwe.de>
+Date:   2013-11-17 14:23:02 +0100
+
+    tighten type requirements for range adaptors
+
+commit 9933463097df9fe16e446a0744204dbe04223544
+Author: Timo Weingärtner <timo@tiwe.de>
+Date:   2013-11-17 14:05:54 +0100
+
+    use LDLIBS instead of LDFLAGS
+    
+    LDLIBS are put behind the prerequisites by make's built-in rules so linking
+    won't fail with --as-needed.
+
+commit d34f4e8e84f3c503a830d1495945f8c4da7c4abe
+Author: Timo Weingärtner <timo@tiwe.de>
+Date:   2013-06-17 17:43:33 +0200
+
+    fix copy&paste oversight in license
+
+commit 5d0c23bc87b5be492869616acc42e619914aae38
+Author: Timo Weingärtner <timo@tiwe.de>
+Date:   2013-04-24 16:31:15 +0200
+
+    inode.h is not so big anymore, merge it into hadori.C
+
+commit 5f9ccea4152b63018095e247066524c425e7824e
+Author: Timo Weingärtner <timo@tiwe.de>
+Date:   2013-04-23 18:44:12 +0200
+
+    std::move inode f into kept, we won't use it afterwards
+
+commit 7bb00e45cc6fcfa85514bf4e645b0cedebbc0216
+Author: Timo Weingärtner <timo@tiwe.de>
+Date:   2013-04-23 18:32:03 +0200
+
+    use agg init and drop unnecessary stuff
+
+commit 9d0562d22cef95b0db1af30e199a371670ca8050
+Author: Timo Weingärtner <timo@tiwe.de>
+Date:   2013-04-23 18:28:33 +0200
+
+    remove hashing stuff
+    
+    the cases in which that gave a speedup are rare to nonexistent
+
+commit fb3fe1b085b30fcc9defa8f29477226806db085b
+Author: Timo Weingärtner <timo@tiwe.de>
+Date:   2012-03-04 03:08:30 +0100
+
+    Use initializer braces instead of make_pair
+
+commit 5a901f9956801b7b9ed819da8b721a58ad93d095
+Author: Timo Weingärtner <timo@tiwe.de>
+Date:   2012-03-03 01:57:29 +0100
+
+    TODO replaced by issues on github
+
+commit fc2e6f497641f6189d4c46bb975c47fe5cdecc24 (tag: 0.2)
+Author: Timo Weingärtner <timo@tiwe.de>
+Date:   2012-02-26 16:11:21 +0100
+
+    Version 0.2
+
+commit 8079e98267742de01442f0294d41085e6af2353a
+Author: Timo Weingärtner <timo@tiwe.de>
+Date:   2012-02-26 16:03:06 +0100
+
+    Fix version.h generation for non-git
+
+commit 27b8fc1395bece0dc309512a3790a4892e0d29b1
+Author: Timo Weingärtner <timo@tiwe.de>
+Date:   2012-02-03 14:33:21 +0100
+
+    Add --version, generate version.h from git, clean manpage
+
+commit 894df89979fdd604588eae999f984b05e471ceb8
+Author: Timo Weingärtner <timo@tiwe.de>
+Date:   2012-02-03 13:52:04 +0100
+
+    Output help to stdout and exit with EX_OK.
+
+commit cee77d8be37b074fd6e039a8207e30f2f067c561
+Author: Timo Weingärtner <timo@tiwe.de>
+Date:   2012-01-17 19:24:06 +0100
+
+    Add copyright notice and license.
+
+commit 8f967e3c65d1ecf6c43a6ff41b70f0fe47e64480
+Author: Timo Weingärtner <timo@tiwe.de>
+Date:   2011-12-25 18:45:41 +0100
+
+    clarify description for --stdin
+
+commit a40bb6e7fd6e74e8b2d2222bba9421b580577e8e
+Author: Timo Weingärtner <timo@tiwe.de>
+Date:   2011-12-02 20:23:18 +0100
+
+    also use unordered_map for sizes
+
+commit cffdf8acae2b5f985eeb81cb6e0a5bbd218b9d32
+Author: Timo Weingärtner <timo@tiwe.de>
+Date:   2011-12-02 19:52:02 +0100
+
+    use equal_range and range-for
+    
+    needs an adapter because std::multimap doesn't deliver begin() and end() for
+    a std::pair of its iterators.
+
+commit 091b33363d0c067c9071414f97fda16102bc065b
+Author: Timo Weingärtner <timo@tiwe.de>
+Date:   2011-12-02 19:17:47 +0100
+
+    use unordered_map for kept and to_link
+    
+    sorting is not needed
+
+commit 1c803dcad160e85fe4b8d46b322374283c0edd89
+Author: Timo Weingärtner <timo@tiwe.de>
+Date:   2011-12-02 19:07:31 +0100
+
+    move kept, to_link, sizes as static into handle_file
+    
+    only used inside that function
+
+commit 568bd7e0d719d7754943c6cfebfa4e4179d62333
+Author: Timo Weingärtner <timo@tiwe.de>
+Date:   2011-11-30 17:51:32 +0100
+
+    use more C++11 features
+
+commit 7349d7a5f4c305f5ba350858cc810ab79ca81bb4
+Author: Timo Weingärtner <timo@tiwe.de>
+Date:   2011-11-28 01:11:16 +0100
+
+    forget/dont't memorize inode in to_link if st_nlink == 1
+    
+    Might have caused data loss if inode number is reallocted to a file inside
+    the tree(s) being scanned.
+    Also saves a little bit of memory.
+
+commit f52b13a15df60b02e10f88b656192a2db5077706
+Author: Timo Weingärtner <timo@tiwe.de>
+Date:   2011-11-28 01:11:16 +0100
+
+    TODO update
+
+commit 148e9eab1109789916a7b55ece957156a57ee4cf (tag: 0.1)
+Author: Timo Weingärtner <timo@tiwe.de>
+Date:   2011-11-06 00:00:31 +0100
+
+    initial commit
diff --git a/hadori.C b/hadori.C
index 44b6e97..b1f0e5f 100644
--- a/hadori.C
+++ b/hadori.C
@@ -8,7 +8,7 @@
  * the Free Software Foundation, either version 3 of the License, or
  * (at your option) any later version.
  *
- * Foobar is distributed in the hope that it will be useful,
+ * hadori is distributed in the hope that it will be useful,
  * but WITHOUT ANY WARRANTY; without even the implied warranty of
  * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
  * GNU General Public License for more details.
@@ -20,12 +20,15 @@
 #include <boost/program_options.hpp>
 namespace po = boost::program_options;
 
+#include <boost/range/iterator_range.hpp>
+
 #include <string>
 #include <vector>
 #include <queue>
 #include <unordered_map>
 #include <iostream>
 #include <sstream>
+#include <fstream>
 
 #include <cstdlib>
 #include <cstring>
@@ -35,39 +38,55 @@ namespace po = boost::program_options;
 #include <dirent.h>
 #include <sysexits.h>
 
-#include "inode.h"
 #include "version.h"
 
-// needed for equal_range and range-for
-namespace std {
-template<typename T> T& begin(pair<T,T> & ip) {
-	return ip.first;
-}
-template<typename T> T& end(pair<T,T> & ip) {
-	return ip.second;
-}
-}
-
 po::variables_map config;
 std::ostream debug(std::clog.rdbuf()), verbose(std::clog.rdbuf()), error(std::clog.rdbuf());
 
+struct inode {
+	std::string const filename;
+	struct stat const stat;
+};
+
+inline bool compare (inode const & l, inode const & r) {
+	char lbuffer[1 << 14];
+	char rbuffer[1 << 14];
+	std::ifstream lf(l.filename.c_str());
+	std::ifstream rf(r.filename.c_str());
+	
+	while (not lf.eof()) {
+		lf.read(lbuffer, sizeof(lbuffer));
+		rf.read(rbuffer, sizeof(rbuffer));
+		if (lf.gcount() != rf.gcount())
+			return false;
+		if (memcmp(lbuffer, rbuffer, lf.gcount()))
+			return false;
+	}
+	return true;
+}
+
+inline std::ostream& operator<< (std::ostream& os, inode const & i) {
+	os << "Inode " << i.stat.st_ino << ", represented by " << i.filename;
+	return os;
+}
+
 void do_link (inode const & i, std::string const & other) {
 	if (!link(i.filename.c_str(), other.c_str())) {
 		error << "linking " << i << " to " << other << " succeeded before unlinking (race condition)" << std::endl;
 		exit(EX_UNAVAILABLE);
 	}
 	if (errno != EEXIST) {
-		char * errstring = strerror(errno);
+		char const * const errstring = strerror(errno);
 		error << "error linking " << i << " to " << other << ": " << errstring << ", nothing bad happened." << std::endl;
 		exit(EX_UNAVAILABLE);
 	}
 	if (unlink(other.c_str())) {
-		char * errstring = strerror(errno);
+		char const * const errstring = strerror(errno);
 		error << "error unlinking " << other << " before linking " << i << " to it: " << errstring << std::endl;
 		exit(EX_UNAVAILABLE);
 	}
 	if (link(i.filename.c_str(), other.c_str())) {
-		char * errstring = strerror(errno);
+		char const * const errstring = strerror(errno);
 		error << "error linking " << i << " to " << other << ": " << errstring << ", destination filename was already unlinked." << std::endl;
 		exit(EX_UNAVAILABLE);
 	}
@@ -75,8 +94,8 @@ void do_link (inode const & i, std::string const & other) {
 
 void handle_file(std::string const & path, struct stat const & s) {
 	static std::unordered_map<ino_t, inode const> kept;
-	static std::unordered_map<ino_t, ino_t> to_link;
-	static std::unordered_multimap<off_t, ino_t> sizes;
+	static std::unordered_map<ino_t, ino_t const> to_link;
+	static std::unordered_multimap<off_t, ino_t const> sizes;
 	
 	debug << "examining " << path << std::endl;
 	if (kept.count(s.st_ino)) {
@@ -91,9 +110,9 @@ void handle_file(std::string const & path, struct stat const & s) {
 			to_link.erase(s.st_ino);
 		return;
 	}
-	inode f(path, s);
+	inode f{path, s};
 	debug << f << " is new to us" << std::endl;
-	for (auto const & it : sizes.equal_range(s.st_size)) {
+	for (auto const & it : boost::make_iterator_range(sizes.equal_range(s.st_size))) {
 		inode const & candidate = kept.find(it.second)->second;
 		debug << "looking if it matches " << candidate << std::endl;
 		if (candidate.stat.st_mode != s.st_mode)
@@ -105,21 +124,18 @@ void handle_file(std::string const & path, struct stat const & s) {
 		if (not config.count("no-time"))
 			if (candidate.stat.st_mtime != s.st_mtime)
 				continue;
-		if (config.count("hash"))
-			if (candidate.get_adler() != f.get_adler())
-				continue;
 		if (!compare(candidate, f))
 			continue;
 		verbose << "linking " << candidate << " to " << path << std::endl;
 		if (s.st_nlink > 1)
-			to_link.insert(std::make_pair(s.st_ino, it.second));
+			to_link.insert({s.st_ino, it.second});
 		if (not config.count("dry-run"))
 			do_link(candidate, path);
 		return;
 	}
 	debug << "we keep " << f << std::endl;
-	kept.insert(std::make_pair(s.st_ino, f));
-	sizes.insert(std::make_pair(s.st_size, s.st_ino));
+	kept.insert({s.st_ino, std::move(f)});
+	sizes.insert({s.st_size, s.st_ino});
 }
 
 void recurse (std::string const & dir, dev_t const dev) {
@@ -129,7 +145,7 @@ void recurse (std::string const & dir, dev_t const dev) {
 	std::queue<std::string> subdirs;
 	
 	if (!(D = opendir(dir.c_str()))) {
-		char * errstring = strerror(errno);
+		char const * const errstring = strerror(errno);
 		error << "opendir(\"" << dir << "\"): " << errstring << std::endl;
 		return;
 	}
@@ -138,7 +154,7 @@ void recurse (std::string const & dir, dev_t const dev) {
 		path += '/';
 		path += d->d_name;
 		if (lstat(path.c_str(), &s)) {
-			char * errstring = strerror(errno);
+			char const * const errstring = strerror(errno);
 			error << "lstat(\"" << path << "\"): " << errstring << std::endl;
 			continue;
 		}
@@ -167,7 +183,7 @@ void recurse_start (std::string const & dir) {
 	struct stat s;
 
 	if (lstat(dir.c_str(), &s)) {
-		char * errstring = strerror(errno);
+		char const * const errstring = strerror(errno);
 		error << "lstat(\"" << dir << "\"): " << errstring << std::endl;
 		exit(EX_NOINPUT);
 	}
@@ -184,13 +200,12 @@ void recurse_start (std::string const & dir) {
 		handle_file(dir, s);
 }
 
-int main (int const argc, char ** argv) {
+int main (int const argc, char const * const * const argv) {
 	po::options_description opts("OPTIONS");
 	opts.add_options()
 		("help,h",	"print this help message")
 		("version,V",	"print version information")
 		("no-time,t",	"ignore mtime")
-		("hash",	"use adler32 hash to speed up comparing many files with same size and mostly identical content")
 		("dry-run,n",	"don't change anything, implies --verbose")
 		("verbose,v",	"show which files get linked")
 		("debug,d",	"show files being examined")
@@ -234,9 +249,7 @@ int main (int const argc, char ** argv) {
 	} else {
 		if (not config.count("stdin") and not config.count("null"))
 			error << "no arguments supplied, assuming --stdin." << std::endl;
-		char delim = '\n';
-		if (config.count("null"))
-			delim = '\0';
+		char const delim = config.count("null") ? '\0' : '\n';
 		for (std::string dir; getline(std::cin, dir, delim);)
 			recurse_start(dir);
 	}
diff --git a/inode.h b/inode.h
deleted file mode 100644
index 26db923..0000000
--- a/inode.h
+++ /dev/null
@@ -1,83 +0,0 @@
-/*
- * Copyright (C) 2011 Timo Weingärtner <timo@tiwe.de>
- *
- * This file is part of hadori.
- *
- * hadori is free software: you can redistribute it and/or modify
- * it under the terms of the GNU General Public License as published by
- * the Free Software Foundation, either version 3 of the License, or
- * (at your option) any later version.
- *
- * Foobar is distributed in the hope that it will be useful,
- * but WITHOUT ANY WARRANTY; without even the implied warranty of
- * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
- * GNU General Public License for more details.
- *
- * You should have received a copy of the GNU General Public License
- * along with hadori.  If not, see <http://www.gnu.org/licenses/>.
- */
-
-#include <string>
-#include <ostream>
-#include <fstream>
-
-#include <cstring>
-#include <sys/stat.h>
-#include <sys/types.h>
-
-#include <zlib.h>
-
-class inode {
-public:
-	std::string const filename;
-	struct stat const stat;
-protected:	
-	uLong mutable adler;
-
-public:
-	inode (std::string const &, struct stat const);
-		
-	uLong get_adler () const;
-	
-	friend bool compare (inode const &, inode const &);
-	friend std::ostream& operator<< (std::ostream&, inode const &);
-};
-
-inline inode::inode (std::string const & __filename, struct stat const __stat) : filename(__filename), stat(__stat), adler(-1) {
-}
-
-inline uLong inode::get_adler () const {
-	if (adler == uLong(-1)) {
-		char buffer[1 << 14];
-		std::ifstream f(filename.c_str());
-		
-		adler = adler32(0L, Z_NULL, 0);
-		while (not f.eof()) {
-			f.read(buffer, sizeof(buffer));
-			adler = adler32(adler, (Bytef *) buffer, f.gcount());
-		}
-	}
-	return adler;
-}
-
-inline bool compare (inode const & l, inode const & r) {
-	char lbuffer[1 << 14];
-	char rbuffer[1 << 14];
-	std::ifstream lf(l.filename.c_str());
-	std::ifstream rf(r.filename.c_str());
-	
-	while (not lf.eof()) {
-		lf.read(lbuffer, sizeof(lbuffer));
-		rf.read(rbuffer, sizeof(rbuffer));
-		if (lf.gcount() != rf.gcount())
-			return false;
-		if (memcmp(lbuffer, rbuffer, lf.gcount()))
-			return false;
-	}
-	return true;
-}
-
-inline std::ostream& operator<< (std::ostream& os, inode const & i) {
-	os << "Inode " << i.stat.st_ino << ", represented by " << i.filename;
-	return os;
-}