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

use regexp to better discriminate From delimiter lines in mbox. Avoid reading mboxes twice

Browse Source
This commit is contained in:
dockes 2006-12-05 15:25:17 +00:00
parent 9a7d469e18
commit 290a7272be
1 changed files with 36 additions and 26 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

62
src/internfile/mh_mail.cpp
View File

@ -1,5 +1,5 @@
#ifndef lint #ifndef lint
static char rcsid[] = "@(#$Id: mh_mail.cpp,v 1.20 2006-09-23 07:39:18 dockes Exp $ (C) 2005 J.F.Dockes"; static char rcsid[] = "@(#$Id: mh_mail.cpp,v 1.21 2006-12-05 15:25:17 dockes Exp $ (C) 2005 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
@ -23,6 +23,7 @@ static char rcsid[] = "@(#$Id: mh_mail.cpp,v 1.20 2006-09-23 07:39:18 dockes Exp
#include <errno.h> #include <errno.h>
#include <unistd.h> #include <unistd.h>
#include <time.h> #include <time.h>
#include <regex.h>
#include <map> #include <map>
#include <sstream> #include <sstream>
@ -93,6 +94,10 @@ MimeHandlerMail::mkDoc(RclConfig *cnf, const string &fn,
return MimeHandler::MHError; return MimeHandler::MHError;
} }
static const char *frompat = "^From .* [1-2][0-9][0-9][0-9]\n$";
static regex_t fromregex;
static bool regcompiled;
MimeHandler::Status MimeHandler::Status
MimeHandlerMail::processmbox(const string &fn, Rcl::Doc &docout, string& ipath) MimeHandlerMail::processmbox(const string &fn, Rcl::Doc &docout, string& ipath)
{ {
@ -116,6 +121,10 @@ MimeHandlerMail::processmbox(const string &fn, Rcl::Doc &docout, string& ipath)
} else { } else {
fp = (FILE *)m_vfp; fp = (FILE *)m_vfp;
} }
if (!regcompiled) {
regcomp(&fromregex, frompat, REG_NOSUB);
regcompiled = true;
}
// If we are called to retrieve a specific message, seek to bof // If we are called to retrieve a specific message, seek to bof
// (then scan up to the message). This is for the case where the // (then scan up to the message). This is for the case where the
@ -132,54 +141,54 @@ MimeHandlerMail::processmbox(const string &fn, Rcl::Doc &docout, string& ipath)
off_t start, end; off_t start, end;
bool iseof = false; bool iseof = false;
bool hademptyline = true;
string msgtxt;
do { do {
// Look for next 'From ' Line, start of message. Set start to // Look for next 'From ' Line, start of message. Set start to
// line after this // line after this
char line[301]; char line[501];
for (;;) { for (;;) {
if (!fgets(line, 300, fp)) { if (!fgets(line, 500, fp)) {
// Eof hit while looking for 'From ' -> file done. We'd need // Eof hit while looking for 'From ' -> file done. We'd need
// another return code here // another return code here
return MimeHandler::MHError; return MimeHandler::MHError;
} }
if (line[0] == '\n') {
if (!strncmp("From ", line, 5)) { hademptyline = true;
continue;
}
if (hademptyline && !regexec(&fromregex, line, 0, 0, 0)) {
start = ftello(fp); start = ftello(fp);
m_msgnum++;
break; break;
} }
hademptyline = false;
} }
// Look for next 'From ' line or eof, end of message (we let a // Look for next 'From ' line or eof, end of message.
// spurious empty line in)
for (;;) { for (;;) {
end = ftello(fp); end = ftello(fp);
if (!fgets(line, 300, fp) || !strncmp("From ", line, 5)) { if (!fgets(line, 500, fp)) {
if (ferror(fp) || feof(fp)) if (ferror(fp) || feof(fp))
iseof = true; iseof = true;
break; break;
} }
if (line[0] == '\n') {
hademptyline = true;
continue;
}
if (hademptyline && !regexec(&fromregex, line, 0, 0, 0)) {
break;
}
if (mtarg <= 0 || m_msgnum == mtarg) {
msgtxt += line;
}
hademptyline = false;
} }
m_msgnum++;
fseek(fp, end, SEEK_SET); fseek(fp, end, SEEK_SET);
} while (mtarg > 0 && m_msgnum < mtarg); } while (mtarg > 0 && m_msgnum < mtarg);
stringstream s(msgtxt);
size_t size = end - start;
fseek(fp, start, SEEK_SET);
char *cp = (char *)malloc(size);
if (cp == 0) {
LOGERR(("MimeHandlerMail::processmbox: malloc(%d) failed\n", size));
return MimeHandler::MHError;
}
if (fread(cp, 1, size, fp) != size) {
LOGERR(("MimeHandlerMail::processmbox: fread failed (errno %d)\n",
errno));
free(cp);
return MimeHandler::MHError;
}
string msgbuf(cp, size);
free(cp);
stringstream s(msgbuf);
Binc::MimeDocument doc; Binc::MimeDocument doc;
doc.parseFull(s); doc.parseFull(s);
if (!doc.isHeaderParsed() && !doc.isAllParsed()) { if (!doc.isHeaderParsed() && !doc.isAllParsed()) {
@ -189,6 +198,7 @@ MimeHandlerMail::processmbox(const string &fn, Rcl::Doc &docout, string& ipath)
} }
LOGDEB2(("Calling processMsg with msgnum %d\n", m_msgnum)); LOGDEB2(("Calling processMsg with msgnum %d\n", m_msgnum));
MimeHandler::Status ret = processMsg(docout, doc, 0); MimeHandler::Status ret = processMsg(docout, doc, 0);
LOGDEB2(("msgnum %d: [%s]\n", m_msgnum, docout.text.c_str()));
if (ret == MimeHandler::MHError) if (ret == MimeHandler::MHError)
return ret; return ret;
char buf[20]; char buf[20];