source: trunk/lib/Configuration.cc @ 1237

// $Id: Configuration.cc 1237 2010-10-23 21:41:40Z peter $

/*
  Copyright (C) 2007, 2008, 2009, 2010 Jari Häkkinen, Peter Johansson

  This file is part of svndigest, http://dev.thep.lu.se/svndigest

  svndigest 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.

  svndigest 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 svndigest. If not, see <http://www.gnu.org/licenses/>.
*/

#include <config.h>

#include "Configuration.h"

#include "Colors.h"
#include "Functor.h"
#include "utility.h"

#include "yat/split.h"

#include <algorithm>
#include <cassert>
#include <cstdlib>
#include <fstream>
#include <map>
#include <string>
#include <sstream>
#include <stdexcept>
#include <utility>

namespace theplu{
namespace svndigest{

  Configuration* Configuration::instance_=NULL;


  Configuration::Configuration(void)
  {
  }


  void Configuration::add_codon(std::string key, std::string start, 
                                std::string end)
  {
    std::pair<std::string, std::string> p(start,end);
    String2Codons::iterator iter = string2codons_.end();
    for (String2Codons::iterator i=string2codons_.begin();
         i!=string2codons_.end(); ++i)
      if (i->first == key)
        iter = i;
    
    if (iter==string2codons_.end())
      string2codons_.push_back(std::make_pair(key, VectorPair(1,p)));
    else
      iter->second.push_back(p);
  }


  const std::map<std::string, std::string>&
  Configuration::author_colors(void) const
  {
    return author_color_;
  }

  std::string Configuration::author_str_color(const std::string& author) const
  {
    std::string res;
    std::map<std::string, std::string>::const_iterator iterator;
    if ( (iterator=author_color_.find(author)) != author_color_.end())
      res = iterator->second;
    return res;
  }


  const std::vector<std::pair<std::string, std::string> >* 
  Configuration::codon(std::string file_name) const 
  {
    if (const std::pair<std::string,std::string>* dict=dictionary(file_name))
      file_name = translate(file_name, *dict);
    for (String2Codons::const_iterator i(string2codons_.begin());
         i!=string2codons_.end(); ++i) {
      if (fnmatch(i->first.c_str(), file_name.c_str()))
        return &i->second;
    }
    return NULL;
  }


  const std::map<std::string,Alias>& Configuration::copyright_alias(void) const
  {
    return copyright_alias_;
  }


  const std::string&  Configuration::copyright_string(void) const
  {
    return copyright_string_;
  }

  const std::pair<std::string,std::string>* 
  Configuration::dictionary(std::string lhs) const
  {
    for (size_t i=0; i<dictionary_.size(); ++i)
      if (fnmatch(lhs.c_str(), dictionary_[i].first.c_str()))
        return &dictionary_[i];
    return NULL;
  }


  bool Configuration::equal_false(std::string str) const
  {
    transform(str.begin(), str.end(), str.begin(), tolower);
    return str=="false" || str=="no" || str=="off" || str=="0";
  }


  bool Configuration::equal_true(std::string str) const
  {
    transform(str.begin(), str.end(), str.begin(), tolower);
    return str=="true" || str=="yes" || str=="on" || str=="1";
  }


  const std::string& Configuration::image_anchor_format(void) const
  {
    return image_anchor_format_;
  }


  const std::string& Configuration::image_format(void) const
  {
    return image_format_;
  }


  void Configuration::load(void)
  {
    set_default();
    validate_dictionary();
  }


  void Configuration::load(std::istream& is)
  {
    assert(is.good());

    bool parsing_found=false;
    bool dictionary_found=false;
    std::string line;
    std::string section;
    std::string tmp;
    while (getline(is, line)) {
      line = ltrim(line);
      if (line.empty() || line[0]=='#')
        continue;
      std::stringstream ss(line);
      if (line[0] == '[') {
        getline(ss, tmp, '[');
        getline(ss, section, ']');
        continue;
      }
      std::string lhs;
      getline(ss, lhs, '=');
      lhs = trim(lhs);
      std::string rhs;
      getline(ss, rhs);
      rhs = trim(rhs);
      if (rhs.empty()){
        throw Config_error(line, "expected format: <lhs> = <rhs>");
      }
      if (section == "copyright-alias"){
        std::map<std::string,Alias>::iterator iter = 
          copyright_alias_.lower_bound(lhs);
        if (iter!=copyright_alias_.end() && iter->first==lhs){
          std::stringstream mess;
          mess << "in copright-alias section " << lhs << " defined twice.";
          throw Config_error(line, mess.str());
        }
        
        // insert alias
        copyright_alias_.insert(iter,std::make_pair(lhs, Alias(rhs,copyright_alias_.size())));
      }
      else if (section == "trac"){
        if (lhs=="trac-root")
          trac_root_=rhs;
        else {
          std::stringstream mess;
          mess << "in trac section" << lhs + " is invalid option.";
          throw Config_error(line, mess.str());
        }
      }
      else if (section == "output") {
        if (lhs=="blame-information") {
          if (equal_false(rhs))
            output_blame_information_ = false;
          else if (equal_true(rhs))
            output_blame_information_ = true;
          else {
            throw Config_error(line, "");
          }
        }
        else if (lhs=="file") {
          if (equal_false(rhs))
            output_file_ = false;
          else if (equal_true(rhs))
            output_file_ = true;
          else {
            throw Config_error(line, "");
          }
        }
        else if (lhs=="tab-size") {
          tab_size_ = strtoul(rhs.c_str(), NULL, 10);
          if (tab_size_==0) {
            throw Config_error(line, 
                       "invalid value: tab-size must be a positive integer");
          }
          else if (tab_size_ == ULONG_MAX) {
            throw Config_error(line, "invalid value: out of range");
          }
        }
      }
      else if (section == "copyright") {
        if (lhs=="missing-copyright-warning") {
          if (equal_false(rhs))
            missing_copyright_warning_ = false;
          else if (equal_true(rhs))
            missing_copyright_warning_ = true;
          else {
            throw Config_error(line, "");
          }
        }
        else if (lhs=="copyright-string") {
          copyright_string_ = rhs;
        }
      }
      else if (section == "author-color") {
        unsigned char r,g,b;
        try {
          str2rgb(rhs, r,g,b);
        }
        catch (std::runtime_error& e) {
          throw Config_error(line, e.what());
        }
        author_color_[lhs] = rhs;
      }     
      else if (section == "parsing-codons") {
        if (!parsing_found) {
          parsing_found=true;
          // clearing the default setting
          string2codons_.clear();
        }
        
        if (codon(lhs)) {
          std::stringstream mess;
          mess << "clashes with previous given file name pattern: ";
          // find previous file-name-pattern
          for (String2Codons::const_iterator i(string2codons_.begin());
               i!=string2codons_.end(); ++i) {
            if (fnmatch(lhs.c_str(), i->first.c_str())) {
              mess << "`" << i->first << "'";
              break;
            }
          }
          throw Config_error(line, mess.str());
        }
        std::stringstream ss(rhs);
        std::string start;
        while (getline(ss, start, ':')) {
          start = trim(start);
          std::string end;
          getline(ss, end, ';');
          end = trim(end);
          if (start.empty() && end.empty())
            continue;
          try {
            if (start.empty() || start=="\"\"") {
              throw std::runtime_error("start-code is empty");
            }
            else if (start.size()<3) {
              std::stringstream mess;
              mess << "start-code `" << start << "' is invalid";
              throw std::runtime_error(mess.str());
            }
            start = trim(start, '"');
            if (end.empty() || end=="\"\"") {
              throw std::runtime_error("end-code is empty");
            }
            else if (end.size()<3) {
              std::stringstream mess;
              mess << "end-code `" << end << "' is invalid";
              throw std::runtime_error(mess.str());
            }
            end = trim(end, '"');
          }
          catch (std::runtime_error& e){
            throw Config_error(line, e.what());
          }
          replace(start, "\\n", "\n");
          replace(end, "\\n", "\n");
          add_codon(lhs, start, end);
        }
      } 
      else if (section == "file-name-dictionary") {
        if (!dictionary_found) {
          dictionary_found=true;
          // clearing the default setting
          dictionary_.clear();
        }
        
        if (const std::pair<std::string, std::string>* entry=dictionary(lhs)) {
          std::stringstream mess;
          mess << "clashes with previous given file name pattern: "
               << "`" << entry->first << "'";
          throw Config_error(line, mess.str());
        }
        lhs = trim(lhs);
        rhs = trim(rhs);
        if (!lhs.empty() && !rhs.empty()) 
          dictionary_.push_back(std::make_pair(lhs, rhs));
        else if (!lhs.empty() || !rhs.empty()) {
          throw Config_error(line, "");
        }
      } 
      else if (section == "svn_props") {
        svn_props_.push_back(std::make_pair(lhs, empty_str_map_));
        std::vector<std::string> vec;
        yat::utility::split(vec, rhs, ';');
        for (size_t i=0; i<vec.size(); ++i) {
          std::vector<std::string> vec2;
          yat::utility::split(vec2, vec[i], '=');
          std::string key = trim(vec2[0]);
          std::string value("");
          if (vec2.size() >= 2)
            value = trim(vec2[1]);
          svn_props_.back().second[key] = value;
        }
      }
      else if (section == "image") {
        if (lhs == "format") {
          try {
            image_format(rhs);
          }
          catch (std::runtime_error e) {
            throw Config_error(line, 
                               "unknown format: " + rhs + "\n" + e.what());
          }
        }
        else if (lhs == "image_format") {
          try {
            image_anchor_format(rhs);
          }
          catch (std::runtime_error e) {
            throw Config_error(line, 
                               "unknown format: " + rhs + "\n" + e.what());
          }
        }
      }
    }
    validate_dictionary();
  }


  void Configuration::image_anchor_format(const std::string& format)
  {
    if (format!="none" && format!="pdf" && format!="png" && format!="svg") {
      std::ostringstream oss;
      oss << "Valid arguments are:\n"
          << "  - `none'\n"
          << "  - `pdf'\n"
          << "  - `png'\n"
          << "  - `svg'";
      throw std::runtime_error(oss.str());
    }
    image_anchor_format_ = format;
  }


  void Configuration::image_format(const std::string& format)
  {
    if (format!="none" && format!="png" && format!="svg") {
      std::ostringstream oss;
      oss << "Valid arguments are:\n"
          << "  - `none'\n"
          << "  - `png'\n"
          << "  - `svg'";
      throw std::runtime_error(oss.str());
    }
    image_format_ = format;
  }


  Configuration& Configuration::instance(void)
  {
    if (!instance_){
      instance_ = new Configuration;
      instance_->load();
    }
    return *instance_;
  }


  bool Configuration::missing_copyright_warning(void) const
  {
    return missing_copyright_warning_;
  }


  std::string 
  Configuration::translate(const std::string& str,
                           const std::pair<std::string, std::string>& dic) const
  {
    std::string res;
    std::vector<std::string> vec;
    if (!regexp(dic.first, str, vec)) {
      std::stringstream mess;
      mess << "invalid config file: "
           << "expression " << dic.first << " is invalid";
      throw std::runtime_error(mess.str());       
    }
    for (std::string::const_iterator i(dic.second.begin()); 
         i!=dic.second.end(); ++i) {
      if (*i == '$') {
        std::stringstream ss(std::string(i+1, dic.second.end()));
        size_t n = 0;
        ss >> n;
        if (n>vec.size() || n==0){
          std::stringstream mess;
          mess << "invalid config file: "
               << "expression " << dic.second << " is invalid";
          if (n)
            mess << "because " << n << " is a too large.";
          throw std::runtime_error(mess.str());       
        }
        res += vec[n-1];
        ++i;
        if (n>9){
          ++i;
          if (n>99)
            ++i;

        }
      }
      else
        res += *i;
    }

    return res;
  }


  std::string trans_end_code(std::string str)
  {
    if (str.size()>0 && str[str.size()-1]=='\n')
      return str.substr(0, str.size()-1) + std::string("\\n");
    return str;
  }


  std::string trans_beg_code(std::string str)
  {
    if (str.size()>0 && str[0]=='\n')
      return std::string("\\n") + str.substr(1); 
    return str;
  }


  std::string trim(std::string str, char c)
  {
    if (str.size()<2 || str[0]!=c || str[str.size()-1]!=c){
      std::stringstream mess;
      mess << "expected `" << str << "' to be surrounded by `" << c << "'";
      throw std::runtime_error(mess.str());
    }
    return str.substr(1, str.size()-2);
  }


  void Configuration::set_default(void)
  {
    copyright_alias_.clear();
    copyright_string_="Copyright (C)";
    missing_copyright_warning_=false;
    trac_root_ = "";
    tab_size_ = 2;

    add_codon("*.ac", "#", "\n");
    add_codon("*.ac", "dnl", "\n");
    add_codon("*.am", "#", "\n");
    add_codon("*.as", "#", "\n");
    add_codon("*.as", "dnl", "\n");
    add_codon("*.bat", "\nREM", "\n");
    add_codon("*.bat", "\nrem", "\n");
    add_codon("*.c", "//", "\n");
    add_codon("*.c", "/*", "*/");
    add_codon("*.cc", "//", "\n");
    add_codon("*.cc", "/*", "*/");
    add_codon("*.cpp", "//", "\n");
    add_codon("*.cpp", "/*", "*/");
    add_codon("*.css", "<!--", "-->");
    add_codon("*.cxx", "//", "\n");
    add_codon("*.cxx", "/*", "*/");
    add_codon("*.h", "//", "\n");
    add_codon("*.h", "/*", "*/");
    add_codon("*.hh", "//", "\n");
    add_codon("*.hh", "/*", "*/");
    add_codon("*.hpp", "//", "\n");
    add_codon("*.hpp", "/*", "*/");
    add_codon("*.html", "<%--", "--%>");
    add_codon("*.java", "//", "\n");
    add_codon("*.java", "/*", "*/");
    add_codon("*.jsp", "<!--", "-->");
    add_codon("*.m", "%", "\n");
    add_codon("*.m4", "#", "\n");
    add_codon("*.m4", "dnl", "\n");
    add_codon("*.pl", "#", "\n");
    add_codon("*.pm", "#", "\n");
    add_codon("*.R", "#", "\n");
    add_codon("*.rss", "<!--", "-->");
    add_codon("*.sgml", "<!--", "-->");
    add_codon("*.sh", "#", "\n");
    add_codon("*.shtml", "<!--", "-->");
    add_codon("*.tex", "%", "\n");
    add_codon("*.xhtml", "<!--", "-->");
    add_codon("*.xml", "<!--", "-->");
    add_codon("*.xsd", "<!--", "-->");
    add_codon("*.xsl", "<!--", "-->");
    add_codon("*config", "#", "\n");
    add_codon("bootstrap", "#", "\n");
    add_codon("Makefile", "#", "\n");

    dictionary_ = VectorPair(1, std::make_pair("*.in", "$1"));
    image_format_ = "png";
    image_anchor_format_ = "png";
    output_blame_information_ = true;
    output_file_ = true;

  }


  bool Configuration::output_blame_information(void) const
  {
    return output_blame_information_;
  }


  bool Configuration::output_file(void) const
  {
    return output_file_;
  }


  const std::map<std::string, std::string>&
  Configuration::svn_properties(const std::string& filename) const
  {
    // reverse backwards as we prefer to to pick properties defined later 
    std::vector<props>::const_reverse_iterator iter = 
      find_fn(svn_props_.rbegin(), svn_props_.rend(), filename);
    if (iter==svn_props_.rend())
      return empty_str_map_;
    return iter->second;
  }


  size_t Configuration::tab_size(void) const
  {
    return tab_size_;
  }


  std::string Configuration::trac_root(void) const
  {
    return trac_root_;
  }


  void Configuration::validate_dictionary(void) const
  {
    VectorPair::const_iterator end(dictionary_.end());
    for (VectorPair::const_iterator iter(dictionary_.begin());iter!=end;++iter){
      std::string word(iter->first);
      replace(word, "*", "");
      replace(word, "?", "");
      // throws if dictionary is invalid
      translate(word, *iter);
    }
  }


  std::ostream& operator<<(std::ostream& os, const Configuration& conf)
  {
    os << "### This file configures various behaviors for svndigest\n"
       << "### The commented-out below are intended to demonstrate how to use\n"
       << "### this file.\n"
       << "\n";

    os << "### Section for setting output\n"
       << "[output]\n"
       << "# if true svndigest will output blame information for each file.\n"
       << "blame-information = ";
    if (conf.output_blame_information())
      os << "yes\n";
    else
      os << "no\n";
    os << "# if true report will have pages for files and not only "
       << "directories.\n"
       << "file = ";
    if (conf.output_file())
      os << "yes\n";
    else
      os << "no\n";
    os << "# svndigest uses this value to replace tabs "
       << "with spaces in blame output\n"
       << "tab-size = " << conf.tab_size() << "\n";


    os << "\n### Section for setting behaviour of copyright update\n"
       << "[copyright]\n"
       << "# if true svndigest will warn if file has no copyright statement.\n"
       << "missing-copyright-warning = ";
    if (conf.missing_copyright_warning())
      os << "yes\n";
    else
      os << "no\n";
    os << "copyright-string = " << conf.copyright_string_ << "\n"; 

    os << "\n"
       << "### Section for setting aliases used in copyright update\n"
       << "[copyright-alias]\n"
       << "# jdoe = John Doe\n";

    typedef std::vector<std::pair<std::string, Alias> > vector;
    vector vec;
    std::back_insert_iterator<vector> back_insert_iterator(vec);
    vec.reserve(conf.copyright_alias().size());
    std::copy(conf.copyright_alias().begin(), conf.copyright_alias().end(),
              back_insert_iterator);
    // sort with respect to Alias.id
    IdCompare id;
    PairSecondCompare<const std::string, Alias, IdCompare> comp(id);
    std::sort(vec.begin(),vec.end(), comp);
              

    for (vector::const_iterator i(vec.begin()); i!=vec.end(); ++i) {
      os << i->first << " = " << i->second.name() << "\n";
    }

    os << "\n"
       << "### Section for images\n"
       << "[image]\n"
       << "format = " << conf.image_format() << "\n";
    os << "anchor_format = " << conf.image_anchor_format() << "\n";


    os << "\n"
       << "### Section for author color in plots and blame output.\n"
       << "[author-color]\n"
       << "# jdoe = 000000\n";
    typedef Configuration::str_map str_map;
    for (str_map::const_iterator i(conf.author_color_.begin());
         i!=conf.author_color_.end(); ++i) {
      os << i->first << " = " << i->second << "\n";
    }

    typedef Configuration::props props;
    os << "\n"
       << "### Section for overriding svn properties.\n"
       << "### The format is the same as for section auto-props in subversion\n"
       << "### config file\n"
       << "[svn-props]\n";
    os << "# COPYING = svndigest:ignore\n";
    std::vector<props>::const_iterator p=conf.svn_props_.begin(); 
    for ( ; p!=conf.svn_props_.end(); ++p) {
      os << p->first << " = ";
      const str_map& map = p->second;
      str_map::const_iterator end = map.end();
      for (str_map::const_iterator i=map.begin(); i!=end; ++i) {
        if (i != map.begin())
          os << ";";
        os << i->first << "=" << i->second;
      }
      os << "\n";
    }


    os << "\n"
       << "### Section for setting trac environment.\n"
       << "[trac]\n"
       << "# If trac-root is set, svndigest will create anchors to "
       << "the Trac page.\n"
       << "# trac-root = http://dev.thep.lu.se/svndigest/\n";
    if (!conf.trac_root().empty())
      os << "trac-root = " << conf.trac_root() << "\n";

    os << "\n"
       << "### Section for setting dictionary for file names.\n"
       << "### Prior looking for file name pattern in section " 
       << "[parsing-codons],\n"
       << "### the file name may be translated according to the rules \n"
       << "### in this section. In default setting there is, for example,\n"
       << "### a rule to translate `<FILENAME>.in' to `<FILENAME>'.\n"
       << "### The format of the entries is:\n"
       << "###    file-name-pattern = new-name\n"
       << "### Left hand side may contain wildcards (such as '*' and '?').\n"
       << "### Right hand side may contain \"$i\", which will be replaced \n"
       << "### with the ith wild card in lhs string.\n";

    if (!conf.dictionary_.empty()) {
      os << "\n"
         << "### Section for setting dictionary for file names.\n"
         << "### Prior looking for file name pattern in section " 
         << "[parsing-codons],\n"
         << "### the file name may be translated according to the rules \n"
         << "### in this section. In default setting there is, for example,\n"
         << "### a rule to translate `<FILENAME>.in' to `<FILENAME>'.\n"
         << "### The format of the entries is:\n"
         << "###    file-name-pattern = new-name\n"
         << "### Left hand side may contain wildcards (such as '*' and '?').\n"
         << "### Right hand side may contain \"$i\", which will be replaced \n"
         << "### with the ith wild card in lhs string.\n"
         << "[file-name-dictionary]\n";
      for (size_t i=0; i<conf.dictionary_.size(); ++i)
        os << conf.dictionary_[i].first << " = " 
           << conf.dictionary_[i].second << "\n"; 
    }
    if (!conf.string2codons_.empty()) {
      os << "\n"
         << "### Section for setting parsing modes\n"
         << "### The format of the entries is:\n"
         << "###   file-name-pattern = \"start-code\" : \"end-code\"\n"
         << "### The file-name-pattern may contain wildcards (such as '*' "
         << "and '?').\n"
         << "### String \"\\n\" can be used for codons containing newline"
         << "\n### character.\n"
         << "[parsing-codons]\n";
      for (size_t i=0; i<conf.string2codons_.size(); ++i) {
        os << conf.string2codons_[i].first << " = "; 
        for (size_t j=0; j<conf.string2codons_[i].second.size(); ++j) {
          if (j)
            os << "  ;  ";
          os << "\"" << trans_beg_code(conf.string2codons_[i].second[j].first) 
             << "\":\"" 
             << trans_end_code(conf.string2codons_[i].second[j].second) 
             << "\""; 
        }
        os << "\n";
      }
    }
    return os;
  }

  
  Config_error::Config_error(const std::string& line,const std::string& message)
    : std::runtime_error(std::string("line: `") + line + 
                         std::string("' is invalid.\n") + message)
  {}

}} // end of namespace svndigest and namespace theplu