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

Support From "bla bla" (quoted) From lines

Browse Source
This commit is contained in:
dockes 2009-11-27 07:07:38 +00:00
parent ef7c432841
commit d5c80cdf16
1 changed files with 16 additions and 7 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

23
src/internfile/mh_mbox.cpp
View File

@ -129,22 +129,22 @@ static inline void stripendnl(line_type& line, int& ll)
// in "Date: " header format, like: From - Mon, 8 May 2006 10:57:32 // in "Date: " header format, like: From - Mon, 8 May 2006 10:57:32
// This was added as an alternative format. By the way it also fools "mail" and // This was added as an alternative format. By the way it also fools "mail" and
// emacs-vm, Recoll is not alone // emacs-vm, Recoll is not alone
// // Update: 2009-11-27: word after From may be quoted string: From "john bull"
static const char *frompat = static const char *frompat =
#if 0 //1.9.0 #if 0 //1.9.0
"^From .* [1-2][0-9][0-9][0-9]$"; "^From .* [1-2][0-9][0-9][0-9]$";
#endif #endif
#if 1 #if 1
"^From[ ]+[^ ]+[ ]+" // From whatever "^From[ ]+([^ ]+|\"[^\"]+\")[ ]+" // 'From (toto@tutu|"john bull") '
"[[:alpha:]]{3}[ ]+[[:alpha:]]{3}[ ]+[0-3 ][0-9][ ]+" // Date "[[:alpha:]]{3}[ ]+[[:alpha:]]{3}[ ]+[0-3 ][0-9][ ]+" // Fri Oct 26
"[0-2][0-9]:[0-5][0-9](:[0-5][0-9])?[ ]+" // Time, seconds optional "[0-2][0-9]:[0-5][0-9](:[0-5][0-9])?[ ]+" // Time, seconds optional
"([^ ]+[ ]+)?" // Optional tz "([^ ]+[ ]+)?" // Optional tz
"[12][0-9][0-9][0-9]" // Year, unanchored, more data may follow "[12][0-9][0-9][0-9]" // Year, unanchored, more data may follow
"|" // Or standard mail Date: header format "|" // Or standard mail Date: header format
"^From[ ]+[^ ]+[ ]+" // From toto@tutu "^From[ ]+[^ ]+[ ]+" // From toto@tutu
"[[:alpha:]]{3},[ ]+[0-3]?[0-9][ ]+[[:alpha:]]{3}[ ]+" // Date Mon, 8 May "[[:alpha:]]{3},[ ]+[0-3]?[0-9][ ]+[[:alpha:]]{3}[ ]+" // Mon, 8 May
"[12][0-9][0-9][0-9][ ]+" // Year "[12][0-9][0-9][0-9][ ]+" // Year
"[0-2][0-9]:[0-5][0-9](:[0-5][0-9])?" // Time, secs optional: 10:57(:32)? "[0-2][0-9]:[0-5][0-9](:[0-5][0-9])?" // Time, secs optional
; ;
#endif #endif
// "([ ]+[-+][0-9]{4})?$" // "([ ]+[-+][0-9]{4})?$"
@ -335,6 +335,15 @@ int main(int argc, char **argv)
cerr << "next_document failed" << endl; cerr << "next_document failed" << endl;
exit(1); exit(1);
} }
map<string, string>::const_iterator it =
mh.get_meta_data().find("content");
int size;
if (it == mh.get_meta_data().end()) {
size = -1;
} else {
size = it->second.length();
}
cout << "Doc " << docnt << " size " << size << endl;
docnt++; docnt++;
} }
cout << docnt << " documents found in " << filename << endl; cout << docnt << " documents found in " << filename << endl;