yetangitu/recoll
1
0
Fork 0
Code Issues Pull requests Releases Wiki Activity

catch Xapian exception caused by bad stemming language. Check stemdb creation status and have recollindex exit in error if it failed

Browse source
This commit is contained in:
Jean-Francois Dockes 2016-09-16 12:21:18 +02:00
parent c0892d2327
commit 589c59f6a4
2 changed files with 40 additions and 36 deletions
Download patch file Download diff file Expand all files Collapse all files

14
src/index/indexer.cpp
View file

@ -140,14 +140,17 @@ bool ConfIndexer::index(bool resetbefore, ixType typestorun, int flags)
if (m_updater && !m_updater->update(DbIxStatus::DBIXS_CLOSING, string()))
return false;
createStemmingDatabases();
bool ret = true;
if (!createStemmingDatabases()) {
ret = false;
}
if (m_updater && !m_updater->update(DbIxStatus::DBIXS_CLOSING, string()))
return false;
createAspellDict();
ret = ret && createAspellDict();
clearMimeHandlerCache();
if (m_updater)
m_updater->update(DbIxStatus::DBIXS_DONE, string());
return true;
return ret;
}
bool ConfIndexer::indexFiles(list<string>& ifiles, int flag)
@ -276,6 +279,7 @@ bool ConfIndexer::purgeFiles(std::list<string> &files, int flag)
bool ConfIndexer::createStemmingDatabases()
{
string slangs;
bool ret = true;
if (m_config->getConfParam("indexstemminglanguages", slangs)) {
if (!m_db.open(Rcl::Db::DbUpd)) {
LOGERR("ConfIndexer::createStemmingDb: could not open db\n" );
@ -292,10 +296,10 @@ bool ConfIndexer::createStemmingDatabases()
if (find(langs.begin(), langs.end(), *it) == langs.end())
m_db.deleteStemDb(*it);
}
m_db.createStemDbs(langs);
ret = ret && m_db.createStemDbs(langs);
}
m_db.close();
return true;
return ret;
}
bool ConfIndexer::createStemDb(const string &lang)

62
src/rcldb/expansiondbs.cpp
View file

@ -53,40 +53,40 @@ bool createExpansionDbs(Xapian::WritableDatabase& wdb,
return true;
}
// Stem dbs
vector<XapWritableComputableSynFamMember> stemdbs;
// Note: tried to make this to work with stack-allocated objects, couldn't.
// Looks like a bug in copy constructors somewhere, can't guess where
vector<std::shared_ptr<SynTermTransStem> > stemmers;
for (unsigned int i = 0; i < langs.size(); i++) {
stemmers.push_back(std::shared_ptr<SynTermTransStem>
(new SynTermTransStem(langs[i])));
stemdbs.push_back(
XapWritableComputableSynFamMember(wdb, synFamStem, langs[i],
stemmers.back().get()));
stemdbs.back().recreate();
}
// Unaccented stem dbs
vector<XapWritableComputableSynFamMember> unacstemdbs;
// We can reuse the same stemmer pointers, the objects are stateless.
if (!o_index_stripchars) {
for (unsigned int i = 0; i < langs.size(); i++) {
unacstemdbs.push_back(
XapWritableComputableSynFamMember(wdb, synFamStemUnac, langs[i],
stemmers.back().get()));
unacstemdbs.back().recreate();
}
}
SynTermTransUnac transunac(UNACOP_UNACFOLD);
XapWritableComputableSynFamMember
diacasedb(wdb, synFamDiCa, "all", &transunac);
if (!o_index_stripchars)
diacasedb.recreate();
// Walk the list of all terms, and stem/unac each.
string ermsg;
try {
// Stem dbs
vector<XapWritableComputableSynFamMember> stemdbs;
// Note: tried to make this to work with stack-allocated objects, couldn't.
// Looks like a bug in copy constructors somewhere, can't guess where
vector<std::shared_ptr<SynTermTransStem> > stemmers;
for (unsigned int i = 0; i < langs.size(); i++) {
stemmers.push_back(std::shared_ptr<SynTermTransStem>
(new SynTermTransStem(langs[i])));
stemdbs.push_back(
XapWritableComputableSynFamMember(wdb, synFamStem, langs[i],
stemmers.back().get()));
stemdbs.back().recreate();
}
// Unaccented stem dbs
vector<XapWritableComputableSynFamMember> unacstemdbs;
// We can reuse the same stemmer pointers, the objects are stateless.
if (!o_index_stripchars) {
for (unsigned int i = 0; i < langs.size(); i++) {
unacstemdbs.push_back(
XapWritableComputableSynFamMember(wdb, synFamStemUnac, langs[i],
stemmers.back().get()));
unacstemdbs.back().recreate();
}
}
SynTermTransUnac transunac(UNACOP_UNACFOLD);
XapWritableComputableSynFamMember
diacasedb(wdb, synFamDiCa, "all", &transunac);
if (!o_index_stripchars)
diacasedb.recreate();
Xapian::TermIterator it = wdb.allterms_begin();
// We'd want to skip to the first non-prefixed term, but this is a bit
// complicated, so we just jump over most of the prefixed term and then