escaping.cc

Go to the documentation of this file.
00001 
00002 // -*- mode: c++; c-basic-offset:4 -*-
00003 
00004 // This file is part of libdap, A C++ implementation of the OPeNDAP Data
00005 // Access Protocol.
00006 
00007 // Copyright (c) 2002,2003 OPeNDAP, Inc.
00008 // Author: James Gallagher <jgallagher@opendap.org>
00009 //
00010 // This library is free software; you can redistribute it and/or
00011 // modify it under the terms of the GNU Lesser General Public
00012 // License as published by the Free Software Foundation; either
00013 // version 2.1 of the License, or (at your option) any later version.
00014 //
00015 // This library is distributed in the hope that it will be useful,
00016 // but WITHOUT ANY WARRANTY; without even the implied warranty of
00017 // MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
00018 // Lesser General Public License for more details.
00019 //
00020 // You should have received a copy of the GNU Lesser General Public
00021 // License along with this library; if not, write to the Free Software
00022 // Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA  02111-1307  USA
00023 //
00024 // You can contact OPeNDAP, Inc. at PO Box 112, Saunderstown, RI. 02874-0112.
00025 
00026 // Copyright (c) 1996, California Institute of Technology.
00027 // ALL RIGHTS RESERVED.   U.S. Government Sponsorship acknowledged.
00028 //
00029 // Please read the full copyright notice in the file COPYRIGHT_URI
00030 // in this directory.
00031 //
00032 // Author: Todd Karakashian, NASA/Jet Propulsion Laboratory
00033 //         Todd.K.Karakashian@jpl.nasa.gov
00034 //
00035 // $RCSfile: escaping.cc,v $ - Miscellaneous routines for OPeNDAP HDF server
00036 //
00037 // These two routines are for escaping/unescaping strings that are identifiers
00038 // in DAP2
00039 // id2www() -- escape (using WWW hex codes) non-allowable characters in a
00040 // DAP2 identifier
00041 // www2id() -- given an WWW hexcode escaped identifier, restore it
00042 //
00043 // These two routines are for escaping/unescaping strings storing attribute
00044 // values.  They use traditional octal escapes (\nnn) because they are
00045 // intended to be viewed by a user
00046 // escattr() -- escape (using traditional octal backslash) non-allowable
00047 // characters in the value of a DAP2 attribute
00048 // unescattr() -- given an octally escaped string, restore it
00049 //
00050 // These are routines used by the above, not intended to be called directly:
00051 //
00052 // hexstring()
00053 // unhexstring()
00054 // octstring()
00055 // unoctstring()
00056 //
00057 // -Todd
00058 
00059 #include <ctype.h>
00060 
00061 #include <iomanip>
00062 #include <string>
00063 #include <sstream>
00064 
00065 #include "GNURegex.h"
00066 #include "Error.h"
00067 #include "InternalErr.h"
00068 // #define DODS_DEBUG
00069 #include "debug.h"
00070 
00071 using namespace std;
00072 
00073 namespace libdap {
00074 
00075 // The next four functions were originally defined static, but I removed that
00076 // to make testing them (see generalUtilTest.cc) easier to write. 5/7/2001
00077 // jhrg
00078 
00079 string
00080 hexstring(unsigned char val)
00081 {
00082     ostringstream buf;
00083     buf << hex << setw(2) << setfill('0')
00084     << static_cast<unsigned int>(val);
00085 
00086     return buf.str();
00087 }
00088 
00089 string
00090 unhexstring(string s)
00091 {
00092     int val;
00093     istringstream ss(s);
00094     ss >> hex >> val;
00095     char tmp_str[2];
00096     tmp_str[0] = static_cast<char>(val);
00097     tmp_str[1] = '\0';
00098     return string(tmp_str);
00099 }
00100 
00101 string
00102 octstring(unsigned char val)
00103 {
00104     ostringstream buf;
00105     buf << oct << setw(3) << setfill('0')
00106     << static_cast<unsigned int>(val);
00107 
00108     return buf.str();
00109 }
00110 
00111 string
00112 unoctstring(string s)
00113 {
00114     int val;
00115 
00116     istringstream ss(s);
00117     ss >> oct >> val;
00118 
00119     DBG(cerr << "unoctstring: " << val << endl);
00120 
00121     char tmp_str[2];
00122     tmp_str[0] = static_cast<char>(val);
00123     tmp_str[1] = '\0';
00124     return string(tmp_str);
00125 }
00126 
00151 string
00152 id2www(string in, const string &allowable)
00153 {
00154     string::size_type i = 0;
00155 
00156     while ((i = in.find_first_not_of(allowable, i)) != string::npos) {
00157         in.replace(i, 1, "%" + hexstring(in[i]));
00158         i++;
00159     }
00160 
00161     return in;
00162 }
00163 
00174 string
00175 id2www_ce(string in, const string &allowable)
00176 {
00177     return id2www(in, allowable);
00178 }
00179 
00212 string
00213 www2id(const string &in, const string &escape, const string &except)
00214 {
00215     string::size_type i = 0;
00216     string res = in;
00217     while ((i = res.find_first_of(escape, i)) != string::npos) {
00218         if (except.find(res.substr(i, 3)) != string::npos) {
00219             i += 3;
00220             continue;
00221         }
00222         res.replace(i, 3, unhexstring(res.substr(i + 1, 2)));
00223     }
00224 
00225     return res;
00226 }
00227 
00228 static string
00229 entity(char c)
00230 {
00231     switch (c) {
00232     case '>': return "&gt;";
00233     case '<': return "&lt;";
00234     case '&': return "&amp;";
00235     case '\'': return "&apos;";
00236     case '\"': return "&quot;";
00237     default:
00238         throw InternalErr(__FILE__, __LINE__, "Unrecognized character.");
00239     }
00240 }
00241 
00242 // Assumption: There are always exactly two octal digits in the input
00243 // and two hex digits in the result.
00244 string
00245 octal_to_hex(const string &octal_digits)
00246 {
00247     int val;
00248 
00249     istringstream ss(octal_digits);
00250     ss >> oct >> val;
00251 
00252     ostringstream ds;
00253     ds << hex << setw(2) << setfill('0') << val;
00254     return ds.str();
00255 }
00256 
00263 string
00264 id2xml(string in, const string &not_allowed)
00265 {
00266     string::size_type i = 0;
00267 
00268     while ((i = in.find_first_of(not_allowed, i)) != string::npos) {
00269         in.replace(i, 1, entity(in[i]));
00270         ++i;
00271     }
00272 
00273     // OK, now scan for octal escape sequences like \\012 (where the '\'
00274     // is itself escaped). This type of attribute value comes from the netCDF
00275     // handler and maybe others. Assumption: The '\' will always appear as
00276     // in its escaped form: '\\'. NB: Both backslashes must be escaped in the
00277     // C++ string.
00278     string octal_escape = "\\\\";
00279     i = 0;
00280     string::size_type length = in.length();
00281     while ((i = in.find(octal_escape, i)) != string::npos) {
00282         // Get the three octal digits following the '\\0'
00283         string::size_type j = i + 2;
00284         if (j + 1 >= length)  // Check that we're not past the end
00285             break;
00286         string octal_digits = in.substr(j, 3);
00287         // convert to a &#xdd; XML escape
00288         string hex_escape = string("&#x");
00289         hex_escape.append(octal_to_hex(octal_digits));
00290         hex_escape.append(string(";"));
00291 
00292         // replace the octal escape with an XML/hex escape
00293         in.replace(i, 5, hex_escape);
00294 
00295         // increment i
00296         i += 6;
00297     }
00298 
00299     return in;
00300 }
00301 
00307 string
00308 xml2id(string in)
00309 {
00310     string::size_type i = 0;
00311 
00312     while ((i = in.find("&gt;", i)) != string::npos)
00313         in.replace(i, 4, ">");
00314 
00315     i = 0;
00316     while ((i = in.find("&lt;", i)) != string::npos)
00317         in.replace(i, 4, "<");
00318 
00319     i = 0;
00320     while ((i = in.find("&amp;", i)) != string::npos)
00321         in.replace(i, 5, "&");
00322 
00323     i = 0;
00324     while ((i = in.find("&apos;", i)) != string::npos)
00325         in.replace(i, 6, "'");
00326 
00327     i = 0;
00328     while ((i = in.find("&quot;", i)) != string::npos)
00329         in.replace(i, 6, "\"");
00330 
00331     return in;
00332 }
00333 
00339 string
00340 esc2underscore(string s)
00341 {
00342     string::size_type pos;
00343     while ((pos = s.find('%')) != string::npos)
00344         s.replace(pos, 3, "_");
00345 
00346     return s;
00347 }
00348 
00349 
00353 string
00354 escattr(string s)
00355 {
00356     const string printable = " ABCDEFGHIJKLMNOPQRSTUVWXYZabcdefghijklmnopqrstuvwxyz0123456789~`!@#$%^&*()_-+={[}]|\\:;<,>.?/'\"";
00357     const string ESC = "\\";
00358     const string DOUBLE_ESC = ESC + ESC;
00359     const string QUOTE = "\"";
00360     const string ESCQUOTE = ESC + QUOTE;
00361 
00362     // escape non-printing characters with octal escape
00363     string::size_type ind = 0;
00364     while ((ind = s.find_first_not_of(printable, ind)) != s.npos)
00365         s.replace(ind, 1, ESC + octstring(s[ind]));
00366 
00367     // escape \ with a second backslash
00368     ind = 0;
00369     while ((ind = s.find(ESC, ind)) != s.npos) {
00370         s.replace(ind, 1, DOUBLE_ESC);
00371         ind += DOUBLE_ESC.length();
00372     }
00373 
00374     // escape " with backslash
00375     ind = 0;
00376     while ((ind = s.find(QUOTE, ind)) != s.npos) {
00377         s.replace(ind, 1, ESCQUOTE);
00378         ind += ESCQUOTE.length();
00379     }
00380 
00381     return s;
00382 }
00383 
00392 string
00393 unescattr(string s)
00394 {
00395     Regex octal("\\\\[0-3][0-7][0-7]");  // matches 4 characters
00396     Regex esc_quote("\\\\\"");  // matches 3 characters
00397     Regex esc_esc("\\\\\\\\");      // matches 2 characters
00398     const string ESC = "\\";
00399     const string QUOTE = "\"";
00400     int matchlen;
00401     unsigned int index;
00402 
00403     DBG(cerr << "0XX" << s << "XXX" << endl);
00404     // unescape any escaped backslashes
00405     index = esc_esc.search(s.c_str(), s.length(), matchlen, 0);
00406     while (index < s.length()) {
00407         DBG(cerr << "1aXX" << s << "XXX index: " << index << endl);
00408         s.replace(index, 2, ESC);
00409         DBG(cerr << "1bXX" << s << "XXX index: " << index << endl);
00410         index = esc_esc.search(s.c_str(), s.length(), matchlen, 0);
00411     }
00412 
00413     // unescape any escaped double quote characters
00414     index = esc_quote.search(s.c_str(), s.length(), matchlen, 0);
00415     while (index < s.length()) {
00416         s.replace(index, 2, QUOTE);
00417         DBG(cerr << "2XX" << s << "XXX index: " << index << endl);
00418         index = esc_quote.search(s.c_str(), s.length(), matchlen, 0);
00419     }
00420 
00421     // unescape octal characters
00422     index = octal.search(s.c_str(), s.length(), matchlen, 0);
00423     while (index < s.length()) {
00424         s.replace(index, 4, unoctstring(s.substr(index + 1, 3)));
00425         DBG(cerr << "3XX" << s << "XXX index: " << index << endl);
00426         index = octal.search(s.c_str(), s.length(), matchlen, 0);
00427     }
00428 
00429     DBG(cerr << "4XX" << s << "XXX" << endl);
00430     return s;
00431 }
00432 
00433 string
00434 munge_error_message(string msg)
00435 {
00436     // First, add enclosing quotes if needed.
00437     if (*msg.begin() != '"')
00438         msg.insert(msg.begin(), '"');
00439     if (*(msg.end() - 1) != '"')
00440         msg += "\"";
00441 
00442     // Now escape any internal double quotes that aren't escaped.
00443     string::iterator miter;
00444     for (miter = msg.begin() + 1; miter != msg.end() - 1; miter++)
00445         if (*miter == '"' && *(miter - 1) != '\\')
00446             miter = msg.insert(miter, '\\');
00447 
00448     return msg;
00449 }
00450 
00455 string
00456 escape_double_quotes(string source)
00457 {
00458     string::size_type idx = 0;
00459     while((idx = source.find('\"', idx)) != string::npos) {
00460         source.replace(idx, 1, "\\\""); // a backslash and a double quote
00461         idx += 2;
00462     }
00463 
00464     return source;
00465 }
00466 
00472 string
00473 unescape_double_quotes(string source)
00474 {
00475     string::size_type idx = 0;
00476     while((idx = source.find("\\\"", idx)) != string::npos) {
00477         source.replace(idx, 2, "\""); // a backslash and a double quote
00478         ++idx;
00479     }
00480 
00481     return source;
00482 }
00483 
00484 } // namespace libdap
00485 

Generated on Wed Feb 10 16:08:02 2010 for libdap++ by  doxygen 1.4.7