httpstats-final

00010                                         {
00011 /* A parse succeeds if all components can be parsed.
00012    This should be modified so that we only parse what is necessary,
00013    according to the configuration. But for now, we are too lazy.
00014 */
00015 return parse_date(line) && parse_path(line) && parse_domain(line);
00016 }

00017 { return date_; }

00018 { return path_; }

00019 { return domain_; }

00020                                         {
00021 /* Select date part: just after first '[' up to first folloing ' '
00022    e.g.
00023 
00024         spider1.tiscalinet.be - - [02/Nov/2000:10:20:36 +0100] ..
00025 
00026    will select
00027 
00028         02/Nov/2000:10:20:36
00029 */
00030 string::size_type date_start(line.find_first_of('['));
00031 string::size_type date_end(line.find_first_of(' ',date_start));
00032 string            date_string(line,date_start+1,date_end - date_start-1);
00033 
00034 /* Make date_string acceptable to the Date::Date(const char*) parser.
00035 
00036    1. replace '/' by '-'. In the example, this will yield 
00037 
00038         02-Nov-2000:10:20:36
00039 */
00040 for (string::size_type i=0; i<date_string.size(); ++i)
00041   if (date_string[i]=='/')
00042     date_string[i] = '-';
00043 /*
00044    2. Replace first ':' by space. In the example, this will yield 
00045 
00046         02-Nov-2000 10:20:36
00047 */
00048 date_string.replace(date_string.find_first_of(':'),1," ");
00049 
00050 /* Now parse using Date::Date(const string&). It will throw an exception
00051    if the date cannot be parsed. In that case, we catch the exception
00052    and return false.
00053 */
00054 try {
00055   Date  d(date_string);
00056   date_.set(d.year(),d.month(),d.day(),d.hours());
00057   }
00058 catch (exception& e) {
00059   cerr << "LogRecord::parse_date failed: " << e.what() << endl;
00060   return false;
00061   }
00062 return true;
00063 }

00103                                         {
00104 /* The path part can be found between the first and second occurrences
00105    of `"' (double quote). An example is shown below.
00106 
00107     "GET /ssl/kiesoos.cgi?rolnr=58769\&stjcode=5L10021 HTTP/1.0" 
00108 
00109    This will result in 
00110 
00111      _path = <"","ssl","kiesoos.cgi">
00112 */
00113 string::size_type path_start(line.find_first_of('\"'));
00114 if (path_start==string::npos)
00115   return false;
00116 /* We are not interested in the verb (e.g. GET), so we skip until after the first ' '.
00117    In the example, path_start would then point to.
00118 
00119     /ssl/kiesoos.cgi?rolnr=58769\&stjcode=5L10021 HTTP/1.0" 
00120 */
00121 path_start = line.find_first_of(' ',path_start);
00122 if (path_start==string::npos)
00123   return false;
00124 ++path_start; // After ' '.
00125 /*  We only want the path until the first '?', '#' or ' '. In the example
00126     this should result in.
00127 
00128     path_string = /ssl/kiesoos.cgi
00129 */
00130 
00131 string::size_type path_end = line.find_first_of("#? ",path_start);
00132 if (path_end==string::npos)
00133   return false;
00134 
00135 if (path_end==path_start) // cannot have an empty path
00136   return false;
00137 
00138 string path_string(line,path_start,path_end - path_start);
00139 
00140 // Decode to translate "%7E" back to '~' etc.
00141 www_decode(path_string);
00142 
00143 return path_.parse(path_string);
00144 }

00148                                           {
00149 /* The domain informatin is easy to find: it's the first part of
00150    a line, followed by a ' ', as illustrated in the following example.
00151 
00152         igwe.vub.ac.be - - [12/Feb/2000:13:18:49 +0100] ...
00153         
00154 */
00155 string::size_type domain_start(0);
00156 string::size_type domain_end(line.find_first_of(' '));
00157 if (domain_end==string::npos)
00158   return false;
00159 string domain_string(line,domain_start,domain_end);
00160 
00161 return domain_.parse(domain_string);
00162 }

Public Methods
	LogRecord ()
bool	parse_line (const string &line)
	fill in date_, domain_, path_ from line.
const DatePattern&	date () const
const Path&	path () const
const Domain&	domain () const
Private Methods
bool	parse_date (const string &line)
	Auxiliary function for parse_line().
bool	parse_path (const string &line)
	Auxiliary function for parse_line().
bool	parse_domain (const string &line)
	Auxiliary function for parse_line().
Private Attributes
DatePattern	date_
Path	path_
Domain	domain_

LogRecord Class Reference

Public Methods

Private Methods

Private Attributes

Constructor & Destructor Documentation

Member Function Documentation

Member Data Documentation