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

rationalize stopsuffix list usage

Browse Source
This commit is contained in:
dockes 2006-12-11 14:50:53 +00:00
parent 0bf54fe7d6
commit 02c14a6281
5 changed files with 52 additions and 21 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

10
src/common/rclconfig.cpp
View File

@ -1,5 +1,5 @@
#ifndef lint
static char rcsid[] = "@(#$Id: rclconfig.cpp,v 1.33 2006-11-20 15:28:14 dockes Exp $ (C) 2004 J.F.Dockes";
static char rcsid[] = "@(#$Id: rclconfig.cpp,v 1.34 2006-12-11 14:50:53 dockes Exp $ (C) 2004 J.F.Dockes";
#endif
/*
* This program is free software; you can redistribute it and/or modify
@ -230,7 +230,7 @@ std::list<string> RclConfig::getAllMimeTypes()
return lst;
}
bool RclConfig::getStopSuffixes(list<string>& sufflist)
const list<string>* RclConfig::getStopSuffixes()
{
if (stopsuffixes == 0 && (stopsuffixes = new list<string>) != 0) {
string stp;
@ -239,11 +239,7 @@ bool RclConfig::getStopSuffixes(list<string>& sufflist)
}
}
if (stopsuffixes) {
sufflist = *stopsuffixes;
return true;
}
return false;
return stopsuffixes;
}
string RclConfig::getMimeTypeFromSuffix(const string &suff)

4
src/common/rclconfig.h
View File

@ -16,7 +16,7 @@
*/
#ifndef _RCLCONFIG_H_INCLUDED_
#define _RCLCONFIG_H_INCLUDED_
/* @(#$Id: rclconfig.h,v 1.25 2006-11-20 15:28:04 dockes Exp $ (C) 2004 J.F.Dockes */
/* @(#$Id: rclconfig.h,v 1.26 2006-12-11 14:50:53 dockes Exp $ (C) 2004 J.F.Dockes */
#include <list>
#include <string>
@ -84,7 +84,7 @@ class RclConfig {
* The list is initialized on first call, and not changed for subsequent
* setKeydirs.
*/
bool getStopSuffixes(list<string>& sufflist);
const list<string>* getStopSuffixes();
/**
* Check in mimeconf if input mime type is a compressed one, and

16
src/index/mimetype.cpp
View File

@ -1,5 +1,5 @@
#ifndef lint
static char rcsid[] = "@(#$Id: mimetype.cpp,v 1.17 2006-03-21 11:04:39 dockes Exp $ (C) 2004 J.F.Dockes";
static char rcsid[] = "@(#$Id: mimetype.cpp,v 1.18 2006-12-11 14:50:53 dockes Exp $ (C) 2004 J.F.Dockes";
#endif
/*
* This program is free software; you can redistribute it and/or modify
@ -109,15 +109,11 @@ string mimetype(const string &fn, RclConfig *cfg, bool usfc)
if (cfg == 0)
return "";
list<string> stoplist;
cfg->getStopSuffixes(stoplist);
if (!stoplist.empty()) {
for (list<string>::const_iterator it = stoplist.begin();
it != stoplist.end(); it++) {
if (it->length() > fn.length())
continue;
if (!stringicmp(fn.substr(fn.length() - it->length(),
string::npos), *it)) {
const list<string>* stoplist = cfg->getStopSuffixes();
if (stoplist && !stoplist->empty()) {
for (list<string>::const_iterator it = stoplist->begin();
it != stoplist->end(); it++) {
if (!stringisuffcmp(fn, *it)) {
LOGDEB(("mimetype: fn %s in stoplist (%s)\n", fn.c_str(),
it->c_str()));
return "";

38
src/utils/smallut.cpp
View File

@ -1,5 +1,5 @@
#ifndef lint
static char rcsid[] = "@(#$Id: smallut.cpp,v 1.20 2006-12-07 07:07:18 dockes Exp $ (C) 2004 J.F.Dockes";
static char rcsid[] = "@(#$Id: smallut.cpp,v 1.21 2006-12-11 14:50:53 dockes Exp $ (C) 2004 J.F.Dockes";
#endif
/*
* This program is free software; you can redistribute it and/or modify
@ -120,6 +120,21 @@ int stringicmp(const string & s1, const string& s2)
}
}
extern int stringisuffcmp(const string& s1, const string& s2)
{
string::const_reverse_iterator r1 = s1.rbegin(), re1 = s1.rend(),
r2 = s2.rbegin(), re2 = s2.rend();
while (r1 != re1 && r2 != re2) {
char c1 = ::toupper(*r1);
char c2 = ::toupper(*r2);
if (c1 != c2) {
return c1 > c2 ? 1 : -1;
}
++r1; ++r2;
}
return 0;
}
// s1 is already lowercase
int stringlowercmp(const string & s1, const string& s2)
{
@ -570,9 +585,22 @@ struct spair pairs[] = {
{"a", "Ab"},
};
int npairs = sizeof(pairs) / sizeof(struct spair);
struct spair suffpairs[] = {
{"", ""},
{"", "a"},
{"a", ""},
{"a", "a"},
{"toto.txt", ".txt"},
{"TXT", "toto.txt"},
{"toto.txt", ".txt1"},
{"toto.txt1", ".txt"},
};
int nsuffpairs = sizeof(suffpairs) / sizeof(struct spair);
int main(int argc, char **argv)
{
#if 0
for (int i = 0; i < npairs; i++) {
{
int c = stringicmp(pairs[i].s1, pairs[i].s2);
@ -591,6 +619,14 @@ int main(int argc, char **argv)
}
printf("\n");
}
#else
for (int i = 0; i < nsuffpairs; i++) {
int c = stringisuffcmp(suffpairs[i].s1, suffpairs[i].s2);
printf("[%s] %s [%s] \n", suffpairs[i].s1,
c == 0 ? "matches" : c < 0 ? "<" : ">", suffpairs[i].s2);
}
#endif
}
#endif

5
src/utils/smallut.h
View File

@ -16,7 +16,7 @@
*/
#ifndef _SMALLUT_H_INCLUDED_
#define _SMALLUT_H_INCLUDED_
/* @(#$Id: smallut.h,v 1.20 2006-12-07 07:07:18 dockes Exp $ (C) 2004 J.F.Dockes */
/* @(#$Id: smallut.h,v 1.21 2006-12-11 14:50:53 dockes Exp $ (C) 2004 J.F.Dockes */
#include <string>
#include <list>
#include <map>
@ -31,6 +31,9 @@ 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);
// Is one string the end part of the other ?
extern int stringisuffcmp(const string& s1, const string& s2);
// Compare charset names, removing the more common spelling variations
extern bool samecharset(const string &cs1, const string &cs2);