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

ensure that the names of files with filter errors get indexed anyway

Browse Source
This commit is contained in:
dockes 2007-12-20 09:08:04 +00:00
parent 46a7f05cbc
commit e1584c2e21
2 changed files with 27 additions and 9 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

30
src/index/indexer.cpp
View File

@ -1,5 +1,5 @@
#ifndef lint #ifndef lint
static char rcsid[] = "@(#$Id: indexer.cpp,v 1.64 2007-12-13 06:58:21 dockes Exp $ (C) 2004 J.F.Dockes"; static char rcsid[] = "@(#$Id: indexer.cpp,v 1.65 2007-12-20 09:08:04 dockes Exp $ (C) 2004 J.F.Dockes";
#endif #endif
/* /*
* This program is free software; you can redistribute it and/or modify * This program is free software; you can redistribute it and/or modify
@ -418,24 +418,40 @@ DbIndexer::processone(const std::string &fn, const struct stat *stp,
FileInterner::Status fis = FileInterner::FIAgain; FileInterner::Status fis = FileInterner::FIAgain;
bool hadNullIpath = false; bool hadNullIpath = false;
Rcl::Doc doc; Rcl::Doc doc;
const string plus = "+";
char ascdate[20]; char ascdate[20];
sprintf(ascdate, "%ld", long(stp->st_mtime)); sprintf(ascdate, "%ld", long(stp->st_mtime));
while (fis == FileInterner::FIAgain) { while (fis == FileInterner::FIAgain) {
doc.erase(); doc.erase();
string ipath; string ipath;
fis = interner.internfile(doc, ipath); fis = interner.internfile(doc, ipath);
if (fis == FileInterner::FIError) { if (fis == FileInterner::FIError) {
list<string> ext = interner.getMissingExternal(); list<string> ext = interner.getMissingExternal();
m_missingExternal.merge(ext); m_missingExternal.merge(ext);
m_missingExternal.unique(); m_missingExternal.unique();
// We dont stop indexing for one bad doc // We used to return at this point.
return FsTreeWalker::FtwOk; //
} // The nice side was that if a filter failed because of a
// lacking supporting app, the file would be indexed once
// the app was installed.
//
// The not so nice point was that the file name was not
// indexed.
//
// We now index at least the file name. We use a dirty
// hack to ensure that the indexing will be retried each
// time: the stored number as decimal ascii mtime is
// prefixed with a '+', which doesnt change its value for
// atoll() but is tested by rcldb::needUpdate()
// Reset the date as set by the handler if any
doc.fmtime.erase();
// Go through:
}
// Set the date if this was not done in the document handler
if (doc.fmtime.empty()) { if (doc.fmtime.empty()) {
doc.fmtime = ascdate; // Set the date if this was not done in the document handler
doc.fmtime = (fis == FileInterner::FIError) ? plus + ascdate :
ascdate;
} }
// Internal access path for multi-document files // Internal access path for multi-document files

6
src/rcldb/rcldb.cpp
View File

@ -1,5 +1,5 @@
#ifndef lint #ifndef lint
static char rcsid[] = "@(#$Id: rcldb.cpp,v 1.129 2007-12-13 06:58:21 dockes Exp $ (C) 2004 J.F.Dockes"; static char rcsid[] = "@(#$Id: rcldb.cpp,v 1.130 2007-12-20 09:08:04 dockes Exp $ (C) 2004 J.F.Dockes";
#endif #endif
/* /*
* This program is free software; you can redistribute it and/or modify * This program is free software; you can redistribute it and/or modify
@ -1237,7 +1237,9 @@ bool Db::needUpdate(const string &filename, const struct stat *stp)
cp+= 6; cp+= 6;
} }
#endif #endif
time_t mtime = cp ? atoll(cp) : 0; // If the time string begins with a "+", force an update. Happens
// after a filter error, see indexer.cpp, processone()
time_t mtime = (!cp || *cp == '+') ? 0 : atoll(cp);
// Retrieve file size as stored in db data // Retrieve file size as stored in db data
cp = strstr(data.c_str(), "fbytes="); cp = strstr(data.c_str(), "fbytes=");