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

29
src/internfile/mimehandler.cpp
View File

@ -140,10 +140,21 @@ 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
// appropriate handler object may be in the cache (indexed by mime
// type). This is fast, and necessary to conform to the
// configuration, (ie: text/html might be filtered out by
// indexedmimetypes but an html handler could still be in the
// cache because it was needed by some other interning stack).
string hs;
hs = cfg->getMimeHandlerDef(mtype, filtertypes);
if (!hs.empty()) { // Got a handler definition line
// Do we already have a handler object in the cache ?
map<string, Dijon::Filter *>::iterator it = o_handlers.find(mtype); map<string, Dijon::Filter *>::iterator it = o_handlers.find(mtype);
if (it != o_handlers.end()) { if (it != o_handlers.end()) {
Dijon::Filter *h = it->second; Dijon::Filter *h = it->second;
@ -152,12 +163,8 @@ Dijon::Filter *getMimeHandler(const string &mtype, RclConfig *cfg,
return h; return h;
} }
// Get handler definition for mime type // Not in cache. Break definition into type and name/command
string hs; // string and instanciate handler object
hs = cfg->getMimeHandlerDef(mtype, filtertypes);
if (!hs.empty()) {
// Break definition into type and name/command string
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) {