tris/recoll
1
0
Fork 0
Code Issues Pull Requests Projects Releases Wiki Activity

improve charset name comparison

Browse Source
This commit is contained in:
dockes 2005-11-23 10:16:28 +00:00
parent 5b5be0c853
commit 44fb0eb359
3 changed files with 25 additions and 3 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

3
src/internfile/mh_html.cpp
View File

@ -33,6 +33,7 @@
#include "myhtmlparse.h"
#include "indextext.h"
#include "mh_html.h"
#include "smallut.h"
#include <iostream>
using namespace std;
@ -101,7 +102,7 @@ MimeHandlerHtml::mkDoc(RclConfig *conf, const string &,
} catch (bool) {
pres = p;
if (!pres.doccharset.empty() &&
pres.doccharset != pres.ocharset) {
!samecharset(pres.doccharset, pres.ocharset)) {
LOGDEB(("textHtmlToDoc: charset '%s' doc charset '%s',"
"reparse\n", charset.c_str(),pres.doccharset.c_str()));
charset = pres.doccharset;

20
src/utils/smallut.cpp
View File

@ -1,5 +1,5 @@
#ifndef lint
static char rcsid[] = "@(#$Id: smallut.cpp,v 1.6 2005-10-21 08:14:42 dockes Exp $ (C) 2004 J.F.Dockes";
static char rcsid[] = "@(#$Id: smallut.cpp,v 1.7 2005-11-23 10:16:28 dockes Exp $ (C) 2004 J.F.Dockes";
#endif
#ifndef TEST_SMALLUT
#include <string>
@ -156,6 +156,24 @@ int stringuppercmp(const string & s1, const string& s2)
}
}
// Compare charset names, removing the more common spelling variations
bool samecharset(const string &cs1, const string &cs2)
{
string mcs1, mcs2;
// Remove all - and _, turn to lowecase
for (int i = 0; i < cs1.length();i++) {
if (cs1[i] != '_' && cs1[i] != '-') {
mcs1 += ::tolower(cs1[i]);
}
}
for (int i = 0; i < cs2.length();i++) {
if (cs2[i] != '_' && cs2[i] != '-') {
mcs2 += ::tolower(cs2[i]);
}
}
return mcs1 == mcs2;
}
#else
#include <string>

5
src/utils/smallut.h
View File

@ -1,6 +1,6 @@
#ifndef _SMALLUT_H_INCLUDED_
#define _SMALLUT_H_INCLUDED_
/* @(#$Id: smallut.h,v 1.6 2005-04-06 10:20:11 dockes Exp $ (C) 2004 J.F.Dockes */
/* @(#$Id: smallut.h,v 1.7 2005-11-23 10:16:28 dockes Exp $ (C) 2004 J.F.Dockes */
#include <string>
#include <list>
#include <iostream>
@ -11,6 +11,9 @@ extern int stringicmp(const string& s1, const string& s2);
extern int stringlowercmp(const string& alreadylower, const string& s2);
extern int stringuppercmp(const string& alreadyupper, const string& s2);
// Compare charset names, removing the more common spelling variations
extern bool samecharset(const string &cs1, const string &cs2);
extern bool maketmpdir(string& tdir);
extern string stringlistdisp(const list<string>& strs);