#ifndef lint
static char rcsid[] = "@(#$Id: mh_exec.cpp,v 1.14 2008-10-09 09:19:37 dockes Exp $ (C) 2005 J.F.Dockes";
#endif
/*
* This program is free software; you can redistribute it and/or modify
* it under the terms of the GNU General Public License as published by
* the Free Software Foundation; either version 2 of the License, or
* (at your option) any later version.
*
* This program is distributed in the hope that it will be useful,
* but WITHOUT ANY WARRANTY; without even the implied warranty of
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
* GNU General Public License for more details.
*
* You should have received a copy of the GNU General Public License
* along with this program; if not, write to the
* Free Software Foundation, Inc.,
* 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA.
*/
#include <stdio.h>
#include <iostream>
#include <sstream>
#include "mh_execm.h"
#include "mh_html.h"
#include "debuglog.h"
#include "cancelcheck.h"
#include "smallut.h"
#include "transcode.h"
#include "md5.h"
#include "rclconfig.h"
#include "mimetype.h"
#include "idfile.h"
#include <sys/types.h>
#include <sys/wait.h>
#ifndef NO_NAMESPACES
using namespace std;
#endif /* NO_NAMESPACES */
bool MimeHandlerExecMultiple::startCmd()
{
LOGDEB(("MimeHandlerExecMultiple::startCmd\n"));
if (params.empty()) {
// Hu ho
LOGERR(("MHExecMultiple::mkDoc: empty params\n"));
m_reason = "RECFILTERROR BADCONFIG";
return false;
}
// Command name
string cmd = params.front();
// Build parameter list: delete cmd name
list<string>::iterator it = params.begin();
list<string>myparams(++it, params.end());
// Start filter
m_cmd.putenv(m_forPreview ? "RECOLL_FILTER_FORPREVIEW=yes" :
"RECOLL_FILTER_FORPREVIEW=no");
if (m_cmd.startExec(cmd, myparams, 1, 1) < 0) {
m_reason = string("RECFILTERROR HELPERNOTFOUND ") + cmd;
missingHelper = true;
return false;
}
return true;
}
// Note: data is not used if this is the "document:" field: it goes
// directly to m_metaData["content"] to avoid an extra copy
//
// Messages are made of data elements. Each element is like:
// name: len\ndata
// An empty line signals the end of the message, so the whole thing
// would look like:
// Name1: Len1\nData1Name2: Len2\nData2\n
bool MimeHandlerExecMultiple::readDataElement(string& name, string &data)
{
string ibuf;
// Read name and length
if (m_cmd.getline(ibuf) <= 0) {
LOGERR(("MHExecMultiple: getline error\n"));
return false;
}
// Empty line (end of message) ?
if (!ibuf.compare("\n")) {
LOGDEB(("MHExecMultiple: Got empty line\n"));
name = "";
return true;
}
// Filters will sometimes abort before entering the real protocol, ie if
// a module can't be loaded. Check the special filter error first word:
if (ibuf.find("RECFILTERROR ") == 0) {
m_reason = ibuf;
if (ibuf.find("HELPERNOTFOUND") != string::npos)
missingHelper = true;
return false;
}
// We're expecting something like Name: len\n
vector<string> tokens;
stringToTokens(ibuf, tokens);
if (tokens.size() != 2) {
LOGERR(("MHExecMultiple: bad line in filter output: [%s]\n",
ibuf.c_str()));
return false;
}
vector<string>::iterator it = tokens.begin();
name = *it++;
string& slen = *it;
int len;
if (sscanf(slen.c_str(), "%d", &len) != 1) {
LOGERR(("MHExecMultiple: bad line in filter output: [%s]\n",
ibuf.c_str()));
return false;
}
LOGDEB1(("MHExecMultiple: got name [%s] len: %d\n", name.c_str(), len));
// Hack: check for 'Document:' and read directly the document data
// to m_metaData["content"] to avoid an extra copy of the bulky
// piece
string *datap = &data;
if (!stringlowercmp("document:", name)) {
datap = &m_metaData["content"];
} else {
datap = &data;
}
// Read element data
datap->erase();
if (len > 0 && m_cmd.receive(*datap, len) != len) {
LOGERR(("MHExecMultiple: expected %d bytes of data, got %d\n",
len, datap->length()));
return false;
}
return true;
}
bool MimeHandlerExecMultiple::next_document()
{
LOGDEB(("MimeHandlerExecMultiple::next_document(): [%s]\n", m_fn.c_str()));
if (m_havedoc == false)
return false;
if (missingHelper) {
LOGDEB(("MHExecMultiple::next_document(): helper known missing\n"));
return false;
}
if (m_cmd.getChildPid() < 0 && !startCmd()) {
return false;
}
// Send request to child process. This maybe the first/only
// request for a given file, or a continuation request. We send an
// empty file name in the latter case.
ostringstream obuf;
if (m_filefirst) {
obuf << "FileName: " << m_fn.length() << "\n" << m_fn;
// m_filefirst is set to true by set_document_file()
m_filefirst = false;
} else {
obuf << "Filename: " << 0 << "\n";
}
if (m_ipath.length()) {
obuf << "Ipath: " << m_ipath.length() << "\n" << m_ipath;
}
obuf << "Mimetype: " << m_mimeType.length() << "\n" << m_mimeType;
obuf << "\n";
if (m_cmd.send(obuf.str()) < 0) {
m_cmd.zapChild();
LOGERR(("MHExecMultiple: send error\n"));
return false;
}
// Read answer (multiple elements)
LOGDEB1(("MHExecMultiple: reading answer\n"));
bool eofnext_received = false;
bool eofnow_received = false;
string ipath;
string mtype;
for (int loop=0;;loop++) {
string name, data;
if (!readDataElement(name, data)) {
m_cmd.zapChild();
return false;
}
if (name.empty())
break;
if (!stringlowercmp("eofnext:", name)) {
LOGDEB(("MHExecMultiple: got EOFNEXT\n"));
eofnext_received = true;
}
if (!stringlowercmp("eofnow:", name)) {
LOGDEB(("MHExecMultiple: got EOFNOW\n"));
eofnow_received = true;
}
if (!stringlowercmp("ipath:", name)) {
ipath = data;
LOGDEB(("MHExecMultiple: got ipath [%s]\n", data.c_str()));
}
if (!stringlowercmp("mimetype:", name)) {
mtype = data;
LOGDEB(("MHExecMultiple: got mimetype [%s]\n", data.c_str()));
}
if (loop == 10) {
// ??
LOGERR(("MHExecMultiple: filter sent too many parameters\n"));
return false;
}
}
if (eofnow_received) {
// No more docs
m_havedoc = false;
return false;
}
// It used to be that eof could be signalled just by an empty document, but
// this was wrong. Empty documents can be found ie in zip files and should
// not be interpreted as eof.
if (m_metaData["content"].length() == 0) {
LOGDEB0(("MHExecMultiple: got empty document inside [%s]: [%s]\n",
m_fn.c_str(), ipath.c_str()));
}
// If this has an ipath, it is an internal doc from a
// multi-document file. In this case, either the filter supplies the
// mimetype, or the ipath MUST be a filename-like string which we can use
// to compute a mime type
if (!ipath.empty()) {
m_metaData["ipath"] = ipath;
if (mtype.empty()) {
mtype = mimetype(ipath, 0, RclConfig::getMainConfig(), false);
if (mtype.empty()) {
// mimetype() won't call idFile when there is no file. Do it
mtype = idFileMem(m_metaData["content"]);
if (mtype.empty()) {
// Note this happens for example for directory zip members
// We could recognize them by the end /, but wouldn't know
// what to do with them anyway.
LOGINFO(("MHExecMultiple: cant guess mime type\n"));
mtype = "application/octet-stream";
}
}
}
m_metaData["mimetype"] = mtype;
string md5, xmd5;
MD5String(m_metaData["content"], md5);
m_metaData["md5"] = MD5HexPrint(md5, xmd5);
} else {
m_metaData["mimetype"] = mtype.empty() ? "text/html" : mtype;
m_metaData.erase("ipath");
string md5, xmd5, reason;
if (MD5File(m_fn, md5, &reason)) {
m_metaData["md5"] = MD5HexPrint(md5, xmd5);
} else {
LOGERR(("MimeHandlerExecM: cant compute md5 for [%s]: %s\n",
m_fn.c_str(), reason.c_str()));
}
}
if (eofnext_received)
m_havedoc = false;
return true;
}