Changeset 1551

Timestamp:

Nov 3, 2012, 6:03:36 AM (11 years ago)

Author:

Peter Johansson

Message:

closes #518. svncopyright now works on characters rather than lines. That implies old cache files are obsolete and will be ignored from this revision.

Location:

trunk

Files:

• NEWS (modified) (1 diff)
• doc/readme.txt (modified) (3 diffs)
• lib/CopyrightStats.cc (modified) (6 diffs)
• lib/CopyrightStats.h (modified) (1 diff)
• test/copyright.cc (modified) (3 diffs)
• test/copyright_cache_test.sh (modified) (1 diff)
• test/ignore_revs_test.sh (modified) (5 diffs)

trunk/NEWS

@@ -4 +4 @@
 
 Version 0.11 (released NOT YET)
+  - svncopyright now works on characters rather than lines (ticket #518)
 
   A complete list of closed tickets can be found here [[br]]

trunk/doc/readme.txt

@@ -3 +3 @@
 Copyright (C) 2005 Jari Häkkinen
 Copyright (C) 2006, 2007, 2008 Jari Häkkinen, Peter Johansson
-Copyright (C) 2009, 2010, 2011 Peter Johansson
+Copyright (C) 2009, 2010, 2011, 2012 Peter Johansson
 
 This file is part of svndigest, http://dev.thep.lu.se/svndigest
@@ -202 +202 @@
 svncopyright updates the copyright statement in subversion controlled
 files. The program detects the copyright statement and replaces it
-with a copyright statement calculated from repository statistics. 
+with a copyright statement calculated from repository statistics.
 
 The copyright statement is detected as the first line containing the
@@ -211 +211 @@
 statement block is replaced with a new copyright statement generated
 from analyzing `svn blame output`. An author is considered to have
-copyright of the file if (s)he has added a line of code or comment
+copyright of the file if (s)he has added a non-whitespace character
 (excluding copyright statements). For an example of the format of the
 generated copyright statement, please have a look at the top of this
-file. 
+file.
 
 By default the `svn user name` of the author is printed in the

trunk/lib/CopyrightStats.cc

@@ -25 +25 @@
 #include "Configuration.h"
 #include "LineTypeParser.h"
-#include "SVNblame.h"
+#include "SkipWhiteSpaceIterator.h"
 #include "SVNinfo.h"
 #include "SVNlog.h"
+#include "SVNcat.h"
 #include "utility.h"
 
@@ -92 +93 @@
     std::string line;
     getline(is, line);
-    if (line!="SVNCOPYRIGHT CACHE VERSION 1")
+    if (line!="SVNCOPYRIGHT CACHE VERSION 2")
       return 0;
     getline(is, line);
@@ -149 +150 @@
     assert(commit->revision() >= first_rev);
     log_iterator end = log.commits().end();
+
     // loop over all commits
     for ( ; commit!=end; ++commit) {
@@ -170 +172 @@
         continue;
 
-      SVNblame svn_blame(path_, commit->revision());
-      LineTypeParser parser(path_);
-
-      // loop over lines
-      while (svn_blame.valid()) {
-        int lt = parser.parse(svn_blame.line());
-        if ((lt==LineTypeParser::code || lt==LineTypeParser::comment) &&
-            svn_blame.revision()==commit->revision()) {
-          year2user_[yearrev->first].insert(name);
-          break;
-        }
-        svn_blame.next_line();
-      }
-    }
-  }
+      SVNcat previous(path_, commit->revision()-1);
+      std::string previous_content;
+      remove_copyright_lines(previous.str(), previous_content);
+
+      SVNcat current(path_, commit->revision());
+      std::string current_content;
+      remove_copyright_lines(current.str(), current_content);
+
+      // if any character added between previous and current add
+      // author to set of authors
+      if (!subseq(current_content, previous_content)) {
+        year2user_[yearrev->first].insert(name);
+      }
+    }
+  }
+
+
+  void CopyrightStats::remove_copyright_lines(const std::string& src,
+                                              std::string& result) const
+  {
+    LineTypeParser ltp(path_);
+    std::istringstream is(src);
+    std::string line;
+    while (getline(is, line)) {
+      if (ltp.parse(line) != LineTypeParser::copyright) {
+        result += line;
+        result += '\n';
+      }
+    }
+  }
 
 
@@ -190 +207 @@
   {
     year2user_.clear();
+  }
+
+
+  bool
+  CopyrightStats::subseq(const std::string& sub, const std::string& seq) const
+  {
+    SkipWhiteSpaceIterator iter1(sub.begin(), sub.end());
+    SkipWhiteSpaceIterator end1(sub.end(), sub.end());
+    SkipWhiteSpaceIterator iter2(seq.begin(), seq.end());
+    SkipWhiteSpaceIterator end2(seq.end(), seq.end());
+
+    for ( ; iter1!=end1; ++iter1) {
+      iter2 = std::find(iter2, end2, *iter1);
+      if (iter2 == end2) {
+        return false;
+      }
+    }
+    return true;
   }
 
@@ -199 +234 @@
     assert(os.good());
 
-    os << "SVNCOPYRIGHT CACHE VERSION 1\n";
+    os << "SVNCOPYRIGHT CACHE VERSION 2\n";
     os << config_ << "\n";
     os << year2user_.size() << "\n";

trunk/lib/CopyrightStats.h

@@ -60 +60 @@
 
     void parse(svn_revnum_t rev, const std::map<int, svn_revnum_t>& year2rev);
+    void remove_copyright_lines(const std::string& src,
+                                std::string& result) const;
     void reset(void);
+
+    /**
+       sub is subsequence of seq if sub can be created from seq by
+       removing element in seq.
+
+       \return true if \a sub is subsequence of \a seq
+     */
+    bool subseq(const std::string& sub, const std::string& seq) const;
+
     void write_cache(void);
 

trunk/test/copyright.cc

@@ -66 +66 @@
     suite.add(false);
   }
-  suite.out() << "File contains 1 copyright line.\n"; 
+  suite.out() << "File contains 1 copyright line.\n";
 
-  // warn about missing Copyright statement only in verbose mode
-  if (suite.verbose()){
-    std::stringstream ss;
-    ss << "[copyright]\n"
-       << "missing-copyright-warning=yes\n"
-       << "[copyright-alias]\n"
-       << "jari = jh\n"
-       << "peter = pj\n";
-    Configuration& config = Configuration::instance();
-    config.load(ss);
-  }
+  std::stringstream ss;
+  ss << "[copyright]\n"
+     << "missing-copyright-warning=yes\n"
+     << "[copyright-alias]\n"
+     << "jari = jh\n"
+     << "peter = pj\n";
+  Configuration& config = Configuration::instance();
+  config.load(ss);
 
   suite.out() << "Create SVN instance" << std::endl;
@@ -90 +87 @@
   suite.out() << "Create File object" << std::endl;
   File file(0,filename,"");
-  
-  update_copyright(file, suite.verbose(), false);
+
+  update_copyright(file, true, true);
 
   is.open(filename.c_str());
@@ -100 +97 @@
   std::vector<std::string> copyright_correct;
   copyright_correct.push_back("Copyright (C) 2006 jh");
-  copyright_correct.push_back("Copyright (C) 2007, 2008 pj");
+  copyright_correct.push_back("Copyright (C) 2008 pj");
   if (copyrights.size()!=copyright_correct.size()) {
     suite.add(false);

trunk/test/copyright_cache_test.sh

@@ -43 +43 @@
 test -s stderr && exit_fail
 test -r $cache_file || exit_fail
+cat $cache_file
 # check that we cache user name not alias
 grep peter $cache_file || exit_fail

trunk/test/ignore_revs_test.sh

@@ -31 +31 @@
 cat >> config << _EOF
 [svn-props]
-README = svncopyright:ignore=$ignore
+Node.h = svncopyright:ignore=$ignore
 _EOF
 cat config
@@ -38 +38 @@
 }
 
+# log for AUTHORS
+# r4  jari  2006
+# r21 jari  2006
+# r22 jari  2006
+# r23 jari  2006
+# r41 peter 2007
+# r42 peter 2007 only updated copyright
+# r44 peter 2007 only copyright
+# r47 peter 2007 only copyright
+# r67 peter 2009 only copyright
+
 # no ignore
 SVNCOPYRIGHT_run 0 -v --ignore-cache --root toy_project
-$GREP "2006 Jari" toy_project/README || exit_fail
-$GREP "2007, 2008 Peter" toy_project/README || exit_fail
+$GREP "2006 Jari" toy_project/lib/Node.h || exit_fail
+$GREP "2007 Peter" toy_project/lib/Node.h || exit_fail
 
-run_svncopyright_with_ignore 42
-$GREP "2006 Jari" toy_project/README || exit_fail
-$GREP "Copyright (C) 2008 Peter" toy_project/README || exit_fail
+run_svncopyright_with_ignore 21
+$GREP "2006 Jari" toy_project/lib/Node.h || exit_fail
+$GREP "2007 Peter" toy_project/lib/Node.h || exit_fail
 
-run_svncopyright_with_ignore "-42"
-$GREP "2006 Jari" toy_project/README && exit_fail
-$GREP "Copyright (C) 2008 Peter" toy_project/README || exit_fail
+run_svncopyright_with_ignore "-30"
+$GREP "2006 Jari" toy_project/lib/Node.h && exit_fail
+$GREP "2007 Peter" toy_project/lib/Node.h || exit_fail
 
-run_svncopyright_with_ignore "41"
-$GREP "2006 Jari" toy_project/README || exit_fail
-$GREP "Copyright (C) 2007, 2008 Peter" toy_project/README || exit_fail
+run_svncopyright_with_ignore "1"
+$GREP "2006 Jari" toy_project/lib/Node.h || exit_fail
 
-run_svncopyright_with_ignore "1-34"
-$GREP "2006 Jari" toy_project/README || exit_fail
-$GREP "Copyright (C) 2007, 2008 Peter" toy_project/README || exit_fail
+run_svncopyright_with_ignore "1-22"
+$GREP "2006 Jari" toy_project/lib/Node.h || exit_fail
 
-run_svncopyright_with_ignore "1-35"
-$GREP "2006 Jari" toy_project/README && exit_fail
-$GREP "Copyright (C) 2007, 2008 Peter" toy_project/README || exit_fail
+run_svncopyright_with_ignore "1-23"
+$GREP "2006 Jari" toy_project/lib/Node.h && exit_fail
 
-run_svncopyright_with_ignore "42-"
-$GREP "2006 Jari" toy_project/README || exit_fail
-$GREP "Copyright (C) 2007, 2008 Peter" toy_project/README && exit_fail
-
-run_svncopyright_with_ignore "0-999"
-$GREP "Copyright (C)" toy_project/README || exit_fail
+run_svncopyright_with_ignore "41-"
+$GREP "2006 Jari" toy_project/lib/Node.h || exit_fail
+$GREP "2007 Peter" toy_project/lib/Node.h && exit_fail
 
 # test that nonsense argument fails with grace
@@ -74 +79 @@
 cat > config << _EOF
 [svn-props]
-README = svncopyright:ignore=nonsense
+Node.h = svncopyright:ignore=nonsense
 _EOF
 cat config
@@ -89 +94 @@
 peter = Peter
 [svn-props]
-README = svncopyright:ignore=0-35
+Node.h = svncopyright:ignore=41
 toy_project = svncopyright:ignore=42
 _EOF
@@ -95 +100 @@
 # do not use --ignore-cache so we also test that cache is ignored automatically
 SVNCOPYRIGHT_run 0 -v --root toy_project --config-file=config
-$GREP "Copyright (C) 2006" toy_project/README && exit_fail
-$GREP "Copyright (C) 2008" toy_project/README || exit_fail
+$GREP "Copyright (C) 2006" toy_project/lib/Node.h || exit_fail
+$GREP "Copyright (C) 2007" toy_project/lib/Node.h && exit_fail
 
 exit_success