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

html files would still be indexed even when filtered out by indexedmimetypes because the handler cache was searched before using the config data

Browse Source
This commit is contained in:
Jean-Francois Dockes 2010-07-09 09:21:07 +02:00
parent a15d6c0540
commit c7c3f2e269
1 changed files with 23 additions and 16 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

39
src/internfile/mimehandler.cpp
View File

@ -140,24 +140,31 @@ void returnMimeHandler(Dijon::Filter *handler)
Dijon::Filter *getMimeHandler(const string &mtype, RclConfig *cfg, Dijon::Filter *getMimeHandler(const string &mtype, RclConfig *cfg,
bool filtertypes) bool filtertypes)
{ {
if (mtype.empty()) LOGDEB2(("getMimeHandler: mtype [%s] filtertypes %d\n",
return false; mtype.c_str(), filtertypes));
// Do we already have one ? // Get handler definition for mime type. We do this even if an
map<string, Dijon::Filter *>::iterator it = o_handlers.find(mtype); // appropriate handler object may be in the cache (indexed by mime
if (it != o_handlers.end()) { // type). This is fast, and necessary to conform to the
Dijon::Filter *h = it->second; // configuration, (ie: text/html might be filtered out by
o_handlers.erase(it); // indexedmimetypes but an html handler could still be in the
LOGDEB2(("getMimeHandler: found in cache\n")); // cache because it was needed by some other interning stack).
return h;
}
// Get handler definition for mime type
string hs; string hs;
hs = cfg->getMimeHandlerDef(mtype, filtertypes); hs = cfg->getMimeHandlerDef(mtype, filtertypes);
if (!hs.empty()) { if (!hs.empty()) { // Got a handler definition line
// Break definition into type and name/command string
// Do we already have a handler object in the cache ?
map<string, Dijon::Filter *>::iterator it = o_handlers.find(mtype);
if (it != o_handlers.end()) {
Dijon::Filter *h = it->second;
o_handlers.erase(it);
LOGDEB2(("getMimeHandler: found in cache\n"));
return h;
}
// Not in cache. Break definition into type and name/command
// string and instanciate handler object
string::size_type b1 = hs.find_first_of(" \t"); string::size_type b1 = hs.find_first_of(" \t");
string handlertype = hs.substr(0, b1); string handlertype = hs.substr(0, b1);
if (!stringlowercmp("internal", handlertype)) { if (!stringlowercmp("internal", handlertype)) {
@ -194,8 +201,8 @@ Dijon::Filter *getMimeHandler(const string &mtype, RclConfig *cfg,
} }
#endif #endif
// Finally, unhandled files are either ignored or their name is // Finally, unhandled files are either ignored or their name and
// indexed, depending on configuration // generic metadata is indexed, depending on configuration
bool indexunknown = false; bool indexunknown = false;
cfg->getConfParam("indexallfilenames", &indexunknown); cfg->getConfParam("indexallfilenames", &indexunknown);
if (indexunknown) { if (indexunknown) {