Forks/open-source-search-engine
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity

Merge branch 'testing' of github.com:gigablast/open-source-search-engine into testing

Browse Source
This commit is contained in:
mwells
2014-06-19 05:42:15 -07:00
parent d4202cb49d fdf9d51280
commit 8cc8e19d4d
13 changed files with 395 additions and 70 deletions

6
Collectiondb.cpp

@ -240,7 +240,7 @@ bool Collectiondb::addExistingColl ( char *coll, collnum_t collnum ) {
//cr->m_collnum = oldCollnum;
// get the default.conf from working dir if there
g_parms.setToDefault( (char *)cr , OBJ_COLL );
g_parms.setToDefault( (char *)cr , OBJ_COLL , cr );
strcpy ( cr->m_coll , coll );
cr->m_collLen = gbstrlen ( coll );
@ -383,7 +383,7 @@ bool Collectiondb::addNewColl ( char *coll ,
// . get the default.conf from working dir if there
// . i think this calls CollectionRec::reset() which resets all of its
// rdbbase classes for its collnum so m_collnum needs to be right
g_parms.setToDefault( (char *)cr , OBJ_COLL );
g_parms.setToDefault( (char *)cr , OBJ_COLL , cr );
/*
// the default conf file
@ -1657,7 +1657,7 @@ bool CollectionRec::load ( char *coll , long i ) {
// also reset some counts not included in parms list
reset();
// before we load, set to defaults in case some are not in xml file
g_parms.setToDefault ( (char *)this , OBJ_COLL );
g_parms.setToDefault ( (char *)this , OBJ_COLL , this );
// get the filename with that id
File f;
char tmp2[1024];

4
Conf.cpp

@ -165,12 +165,12 @@ bool Conf::isConnectIp ( unsigned long ip ) {
// . set all member vars to their default values
void Conf::reset ( ) {
g_parms.setToDefault ( (char *)this , OBJ_CONF );
g_parms.setToDefault ( (char *)this , OBJ_CONF ,NULL);
m_save = true;
}
bool Conf::init ( char *dir ) { // , long hostId ) {
g_parms.setToDefault ( (char *)this , OBJ_CONF );
g_parms.setToDefault ( (char *)this , OBJ_CONF ,NULL);
m_save = true;
char fname[1024];
if ( dir ) sprintf ( fname , "%slocalgb.conf", dir );

5
DiskPageCache.cpp

@ -28,7 +28,10 @@ DiskPageCache::DiskPageCache () {
// sometimes db may pass an unitialized DiskPageCache to a BigFile
// so make sure when BigFile::close calls DiskPageCache::rmVfd() our
// m_memOff vector is all NULLed out, otherwise it will core
memset ( m_memOff , 0 , 4 * MAX_NUM_VFDS2 );
//memset ( m_memOff , 0 , sizeof(long *) * MAX_NUM_VFDS2 );
for ( long i = 0 ; i < MAX_NUM_VFDS2 ; i++ )
m_memOff[i] = NULL;
m_availMemOff = NULL;
//m_isOverriden = false;
reset();

4
Makefile

@ -546,8 +546,8 @@ master-rpm:
# DEBIAN PACKAGE SECTION BEGIN
# need to do 'apt-get intall dh-make'
testing-deb:
git archive --format=tar --prefix=gb-1.0/ testing > ../gb_1.0.orig.tar
master-deb:
git archive --format=tar --prefix=gb-1.0/ master > ../gb_1.0.orig.tar
rm -rf debian
# change "-p gb_1.0" to "-p gb_1.1" to update version for example
dh_make -e gigablast@mail.com -p gb_1.0 -f ../gb_1.0.orig.tar

27
Msg39.cpp

@ -37,7 +37,7 @@ Msg39::Msg39 () {
void Msg39::reset() {
if ( m_inUse ) { char *xx=NULL;*xx=0; }
m_allocedTree = false;
m_numDocIdSplits = 1;
//m_numDocIdSplits = 1;
m_tmpq.reset();
m_numTotalHits = 0;
m_gotClusterRecs = 0;
@ -241,29 +241,6 @@ void Msg39::getDocIds2 ( Msg39Request *req ) {
QUICKPOLL ( m_r->m_niceness );
// assume not doing special docid splitting
m_numDocIdSplits = 1;
// . do not do splits if caller is already specifying a docid range
// like for gbdocid: queries i guess.
// . make sure m_msg2 is non-NULL, because if it is NULL we are
// evaluating a query for a single docid for seo tools
if ( m_r->m_minDocId == -1 ) { // && m_msg2 ) {
long nt = m_tmpq.getNumTerms();
m_numDocIdSplits = nt / 2;
if ( m_numDocIdSplits == 0 ) m_numDocIdSplits = 1;
}
// for testing
//m_numDocIdSplits = 3;
//if ( ! g_conf.m_doDocIdRangeSplitting )
// m_numDocIdSplits = 1;
// limit to 10
if ( m_numDocIdSplits > 10 )
m_numDocIdSplits = 10;
// . if caller already specified a docid range, then be loyal to that!
// . or if we do not have enough query terms to warrant splitting
//if ( m_numDocIdSplits == 1 ) {
@ -351,7 +328,7 @@ bool Msg39::controlLoop ( ) {
// the starting docid...
long long d0 = m_ddd;
// shortcut
long long delta = MAX_DOCID / (long long)m_numDocIdSplits;
long long delta = MAX_DOCID / (long long)m_r->m_numDocIdSplits;
// advance to point to the exclusive endpoint
m_ddd += delta;
// ensure this is exclusive of ddd since it will be

9
Msg39.h

@ -69,8 +69,10 @@ class Msg39Request {
m_getDocIdScoringInfo = 1;
// -1 means to not to docid range restriction
m_minDocId = -1;
m_maxDocId = -1;
m_minDocId = -1LL;
m_maxDocId = -1LL;
m_numDocIdSplits = 1;
// for widget, to only get results to append to last docid
m_maxSerpScore = 0.0;
@ -92,6 +94,7 @@ class Msg39Request {
char m_niceness;
long m_maxAge;
long m_maxQueryTerms;
long m_numDocIdSplits;
//long m_compoundListMaxSize;
char m_boolFlag;
uint8_t m_language;
@ -238,7 +241,7 @@ class Msg39 {
char m_debug;
long m_numDocIdSplits;
//long m_numDocIdSplits;
bool m_allocedTree;
long long m_ddd;
long long m_dddEnd;

66
Msg40.cpp

@ -245,10 +245,10 @@ bool Msg40::getResults ( SearchInput *si ,
if ( /*m_si->m_firstResultNum == 0 && */get < m_docsToScanForTopics )
get = m_docsToScanForTopics;
// for alden's reranking. often this is 50!
if ( get < m_si->m_docsToScanForReranking )
get = m_si->m_docsToScanForReranking;
//if ( get < m_si->m_docsToScanForReranking )
// get = m_si->m_docsToScanForReranking;
// for zak's reference pages
if ( get < m_si->m_refs_numToGenerate ) get=m_si->m_refs_numToGenerate;
// if(get<m_si->m_refs_numToGenerate ) get=m_si->m_refs_numToGenerate;
// limit to this ceiling though for peformance reasons
//if ( get > m_maxDocIdsToCompute ) get = m_maxDocIdsToCompute;
// ok, need some sane limit though to prevent malloc from
@ -562,6 +562,43 @@ bool Msg40::getDocIds ( bool recall ) {
mr.m_minSerpDocId = m_si->m_minSerpDocId;
mr.m_maxSerpScore = m_si->m_maxSerpScore;
//
// how many docid splits should we do to avoid going OOM?
//
CollectionRec *cr = g_collectiondb.getRec(m_firstCollnum);
RdbBase *base = NULL;
if ( cr ) g_titledb.getRdb()->getBase(cr->m_collnum);
long long numDocs = 0;
if ( base ) numDocs = base->getNumTotalRecs();
// for every 5M docids per host, lets split up the docid range
// to avoid going OOM
long mult = numDocs / 5000000;
if ( mult <= 0 ) mult = 1;
// . do not do splits if caller is already specifying a docid range
// like for gbdocid: queries i guess.
// . make sure m_msg2 is non-NULL, because if it is NULL we are
// evaluating a query for a single docid for seo tools
//if ( m_r->m_minDocId == -1 ) { // && m_msg2 ) {
long nt = m_si->m_q.getNumTerms();
long numDocIdSplits = nt / 2; // ;/// 2;
if ( numDocIdSplits <= 0 ) numDocIdSplits = 1;
// and mult based on index size
numDocIdSplits *= mult;
//if ( numDocIdSplits < 5 ) numDocIdSplits = 5;
//}
// for testing
//m_numDocIdSplits = 3;
//if ( ! g_conf.m_doDocIdRangeSplitting )
// m_numDocIdSplits = 1;
// limit to 10
if ( numDocIdSplits > 15 )
numDocIdSplits = 15;
// store it in the reuquest now
mr.m_numDocIdSplits = numDocIdSplits;
// . get the docIds
// . this sets m_msg3a.m_clusterLevels[] for us
//if(! m_msg3a.getDocIds ( &m_r, m_si->m_q, this , gotDocIdsWrapper))
@ -1240,6 +1277,18 @@ bool Msg40::launchMsg20s ( bool recalled ) {
if ( m_si->m_streamResults &&
i >= m_printi + MAX_OUTSTANDING_MSG20S - 1 )
break;
// if we have printed enough summaries then do not launch
// any more, wait for them to come back in
if ( m_printi >= m_docsToGetVisible ) {
logf(LOG_DEBUG,"query: got %li summaries. done. "
"waiting on remaining "
"%li to return."
, m_printi
, m_numRequests-m_numReplies);
break;
}
// do not double count!
//if ( i <= m_lastProcessedi ) continue;
// do not repeat for this i
@ -1552,6 +1601,9 @@ bool Msg40::gotSummary ( ) {
State0 *st = (State0 *)m_state;
// keep socket alive if not streaming. like downloading csv...
// this fucks up HTTP replies by inserting a space before the "HTTP"
// it does not render properly on the browser...
/*
long now2 = getTimeLocal();
if ( now2 - m_lastHeartbeat >= 10 && ! m_si->m_streamResults &&
// incase socket is closed and recycled for another connection
@ -1561,7 +1613,7 @@ bool Msg40::gotSummary ( ) {
log("msg40: sent heartbeat of %li bytes on sd=%li",
(long)n,(long)st->m_socket->m_sd);
}
*/
/*
@ -4979,12 +5031,14 @@ static int factCmp ( const void *a, const void *b ) {
return 0;
}
// . aka NUGGABITS
// . now make the fast facts from the gigabits and the samples.
// . these are sentences containing the query and a gigabit.
// . sets m_factBuf
bool Msg40::computeFastFacts ( ) {
// skip for now
return true;
//return true;
bool debugGigabits = m_si->m_debugGigabits;
@ -5037,7 +5091,7 @@ bool Msg40::computeFastFacts ( ) {
//
// store Facts (sentences) into this safebuf (nuggets)
// store Facts (sentences) into this safebuf (nuggets)(nuggabits)
//
char ftmp[100000];
SafeBuf factBuf(ftmp,100000);

261
PageResults.cpp

@ -718,6 +718,219 @@ void gotState ( void *state ){
}
// print all sentences containing this gigabit (fast facts) (nuggabits)
static bool printGigabitContainingSentences ( State0 *st,
SafeBuf *sb ,
Msg40 *msg40 ,
Gigabit *gi ,
SearchInput *si ,
Query *gigabitQuery ) {
static long s_gigabitCount = 0;
sb->safePrintf("<nobr><b>");
//"<img src=http://search.yippy.com/"
//"images/new/button-closed.gif><b>");
HttpRequest *hr = &st->m_hr;
// make a new query
sb->safePrintf("<a href=\"/search?gigabits=1&q=");
sb->urlEncode(gi->m_term,gi->m_termLen);
sb->safeMemcpy("+|+",3);
char *q = hr->getString("q",NULL,"");
sb->urlEncode(q);
sb->safePrintf("\">");
sb->safeMemcpy(gi->m_term,gi->m_termLen);
sb->safePrintf("</a></b>");
sb->safePrintf(" <font color=gray size=-1>");
long numOff = sb->m_length;
sb->safePrintf(" ");//,gi->m_numPages);
sb->safePrintf("</font>");
sb->safePrintf("</b>");
if ( si->m_isAdmin )
sb->safePrintf("[%.0f]{%li}",
gi->m_gbscore,
gi->m_minPop);
long revert = sb->length();
sb->safePrintf("<font color=blue style=align:right;>"
"<a onclick=ccc(%li);>"
, s_gigabitCount
);
long spaceOutOff = sb->length();
sb->safePrintf( "%c%c%c",
0xe2,
0x87,
0x93);
sb->safePrintf(//"[more]"
"</a></font>");
sb->safePrintf("</nobr>"); // <br>
// get facts
long numNuggets = 0;
long numFacts = msg40->m_factBuf.length() / sizeof(Fact);
Fact *facts = (Fact *)msg40->m_factBuf.getBufStart();
bool first = true;
bool second = false;
bool printedSecond = false;
//long long lastDocId = -1LL;
long saveOffset = 0;
for ( long i = 0 ; i < numFacts ; i++ ) {
Fact *fi = &facts[i];
// if printed for a higher scoring gigabit, skip
if ( fi->m_printed ) continue;
// check gigabit match
long k; for ( k = 0 ; k < fi->m_numGigabits ; k++ )
if ( fi->m_gigabitPtrs[k] == gi ) break;
// skip this fact/sentence if does not contain gigabit
if ( k >= fi->m_numGigabits ) continue;
// do not print if no period at end
char *s = fi->m_fact;
char *e = s + fi->m_factLen;
if ( e[-1] != '*' ) continue;
e--;
again:
// first time, print in the single fact div
if ( first ) {
sb->safePrintf("<div "
"style=\"border:1px lightgray solid;\" "
"id=fd%li>",s_gigabitCount);
}
if ( second ) {
sb->safePrintf("<div style=\"max-height:300px;"
"display:none;"
"overflow-x:hidden;"
"overflow-y:auto;"//scroll;"
"border:1px lightgray solid;\" "
"id=sd%li>",s_gigabitCount);
printedSecond = true;
}
Msg20Reply *reply = fi->m_reply;
// ok, print it out
if ( ! first && ! second ) {
//if ( reply->m_docId != lastDocId )
sb->safePrintf("<br><br>\n");
//else {
// sb->setLength ( saveOffset );
// sb->safePrintf("<br><br>\n");
//}
}
else {
sb->safePrintf("<br>");
}
numNuggets++;
// print the fast fact (sentence)
//sb->safeMemcpy ( s , e-s );
// let's highlight with gigabits and query terms
SafeBuf tmpBuf;
Highlight h;
h.set ( &tmpBuf , // print it out here
s , // content
e - s , // len
si->m_queryLangId , // from m_defaultSortLang
gigabitQuery , // the gigabit "query" in quotes
true , // stemming? -- unused
false , // use anchors?
NULL , // baseurl
"<u>", // front tag
"</u>", // back tag
0 , // fieldCode
0 ); // niceness
// now highlight the original query as well but in black bold
h.set ( sb , // print it out here
tmpBuf.getBufStart() , // content
tmpBuf.length() , // len
si->m_queryLangId , // from m_defaultSortLang
&si->m_q , // the regular query
true , // stemming? -- unused
false , // use anchors?
NULL , // baseurl
"<b>" , // front tag
"</b>", // back tag
0 , // fieldCode
0 ); // niceness
fi->m_printed = 1;
saveOffset = sb->length();
sb->safePrintf(" <a href=/get?cnsp=0&"
"strip=1&d=%lli>",reply->m_docId);
long dlen; char *dom = getDomFast(reply->ptr_ubuf,&dlen);
sb->safeMemcpy(dom,dlen);
sb->safePrintf("</a>\n");
//lastDocId = reply->m_docId;
if ( first ) {
sb->safePrintf("</div>");
}
if ( second ) {
second = false;
}
if ( first ) {
first = false;
second = true;
// print first gigabit all over again but in 2nd div
goto again;
}
}
// we counted the first one twice since we had to throw it into
// the hidden div too!
if ( numNuggets > 1 ) numNuggets--;
// do not print the double down arrow if no nuggets printed
if ( numNuggets <= 0 ) {
sb->m_length = revert;
sb->safePrintf("</nobr>");
}
// just remove down arrow if only 1...
else if ( numNuggets == 1 ) {
char *dst = sb->getBufStart()+spaceOutOff;
dst[0] = ' ';
dst[1] = ' ';
dst[2] = ' ';
}
// store the # of nuggets in ()'s like (10 )
else {
char tmp[10];
sprintf(tmp,"(%li)",numNuggets);
char *src = tmp;
// starting storing digits after "( "
char *dst = sb->getBufStart()+numOff;
long srcLen = gbstrlen(tmp);
if ( srcLen > 5 ) srcLen = 5;
for ( long k = 0 ; k < srcLen ; k++ )
dst[k] = src[k];
}
s_gigabitCount++;
if ( printedSecond ) {
sb->safePrintf("</div>");
}
return true;
}
/*
// print all sentences containing this gigabit
static bool printGigabit ( State0 *st,
SafeBuf *sb ,
@ -757,6 +970,7 @@ static bool printGigabit ( State0 *st,
return true;
}
*/
class StateAU {
public:
@ -1883,9 +2097,49 @@ bool printSearchResultsHeader ( State0 *st ) {
Gigabit *gigabits = (Gigabit *)gbuf->getBufStart();
//long numCols = 5;
//long perRow = numGigabits / numCols;
if ( numGigabits && si->m_format == FORMAT_HTML )
// gigabit unhide function
sb->safePrintf (
"<script>"
"function ccc ( gn ) {\n"
"var e = document.getElementById('fd'+gn);\n"
"var f = document.getElementById('sd'+gn);\n"
"if ( e.style.display == 'none' ){\n"
"e.style.display = '';\n"
"f.style.display = 'none';\n"
"}\n"
"else {\n"
"e.style.display = 'none';\n"
"f.style.display = '';\n"
"}\n"
"}\n"
"</script>\n"
);
if ( numGigabits && si->m_format == FORMAT_HTML )
sb->safePrintf("<table cellspacing=7 bgcolor=lightgray>"
"<tr><td width=200px; valign=top>");
Query gigabitQuery;
SafeBuf ttt;
// limit it to 40 gigabits for now
for ( long i = 0 ; i < numGigabits && i < 40 ; i++ ) {
Gigabit *gi = &gigabits[i];
ttt.pushChar('\"');
ttt.safeMemcpy(gi->m_term,gi->m_termLen);
ttt.pushChar('\"');
ttt.pushChar(' ');
}
if ( numGigabits > 0 )
gigabitQuery.set2 ( ttt.getBufStart() ,
si->m_queryLangId ,
true , // queryexpansion?
true ); // usestopwords?
for ( long i = 0 ; i < numGigabits ; i++ ) {
if ( i > 0 && si->m_format == FORMAT_HTML )
sb->safePrintf("<hr>");
@ -1893,8 +2147,11 @@ bool printSearchResultsHeader ( State0 *st ) {
// sb->safePrintf("</td><td valign=top>");
// print all sentences containing this gigabit
Gigabit *gi = &gigabits[i];
printGigabit ( st,sb , msg40 , gi , si );
sb->safePrintf("<br>");
//printGigabit ( st,sb , msg40 , gi , si );
//sb->safePrintf("<br>");
printGigabitContainingSentences(st,sb,msg40,gi,si,
&gigabitQuery);
sb->safePrintf("<br><br>");
}
if ( numGigabits && si->m_format == FORMAT_HTML )
sb->safePrintf("</td></tr></table>");

41
Parms.cpp

@ -991,9 +991,9 @@ bool Parms::setGigablastRequest ( TcpSocket *socket ,
long obj = OBJ_GBREQUEST;
//
// reset THIS to defaults
// reset THIS to defaults. use NULL for cr since mostly for SearchInput
//
setToDefault ( THIS , obj );
setToDefault ( THIS , obj , NULL);
// loop through cgi parms
for ( long i = 0 ; i < hr->getNumFields() ; i++ ) {
@ -1533,7 +1533,7 @@ bool Parms::printParms2 ( SafeBuf* sb ,
page = PAGE_SECURITY;
GigablastRequest gr;
g_parms.setToDefault ( (char *)&gr , OBJ_GBREQUEST );
g_parms.setToDefault ( (char *)&gr , OBJ_GBREQUEST , NULL);
// find in parms list
for ( long i = 0 ; i < m_numParms ; i++ ) {
@ -2628,11 +2628,15 @@ void Parms::setParm ( char *THIS , Parm *m , long mm , long j , char *s ,
float oldVal = 0;
float newVal = 0;
if ( ! s && m->m_type != TYPE_CHARPTR) {
if ( ! s &&
m->m_type != TYPE_CHARPTR &&
m->m_type != TYPE_FILEUPLOADBUTTON &&
m->m_defOff==-1) {
s = "0";
char *tit = m->m_title;
if ( ! tit || ! tit[0] ) tit = m->m_xml;
log(LOG_LOGIC,"admin: Parm \"%s\" had NULL default value.",
log(LOG_LOGIC,"admin: Parm \"%s\" had NULL default value. "
"Forcing to 0.",
tit);
//char *xx = NULL; *xx = 0;
}
@ -2916,8 +2920,9 @@ void Parms::setToDefault ( char *THIS , char objType , CollectionRec *argcr ) {
if ( m->m_obj != objType ) continue;
if ( m->m_obj == OBJ_NONE ) continue;
if ( m->m_type == TYPE_COMMENT ) continue;
if ( m->m_type == TYPE_FILEUPLOADBUTTON )
continue;
// no, we gotta set GigablastRequest::m_contentFile to NULL
//if ( m->m_type == TYPE_FILEUPLOADBUTTON )
// continue;
if ( m->m_type == TYPE_MONOD2 ) continue;
if ( m->m_type == TYPE_MONOM2 ) continue;
if ( m->m_type == TYPE_CMD ) continue;
@ -2958,6 +2963,7 @@ void Parms::setToDefault ( char *THIS , char objType , CollectionRec *argcr ) {
if ( m->m_max <= 1 ) {
//if ( i == 282 ) // "query" parm
// log("hey");
//if ( ! m->m_def ) { char *xx=NULL;*xx=0; }
setParm ( THIS , m, i, 0, m->m_def, false/*not enc.*/,
false );
//((CollectionRec *)THIS)->m_orig[i] = 1;
@ -5150,6 +5156,7 @@ void Parms::init ( ) {
m->m_page = PAGE_BASIC_SETTINGS;
m->m_obj = OBJ_COLL;
m->m_off = 0;
m->m_def = NULL;
m->m_type = TYPE_FILEUPLOADBUTTON;
m->m_flags = PF_NOSAVE | PF_DUP;
m++;
@ -6987,6 +6994,21 @@ void Parms::init ( ) {
m->m_obj = OBJ_COLL;
m++;
m->m_title = "number of reference pages to generate";
m->m_desc = "What is the number of "
"reference pages to generate per query? Set to 0 to save "
"CPU time.";
m->m_cgi = "snrp";
m->m_off = (char *)&si.m_refs_numToGenerate - y;
m->m_type = TYPE_LONG;
m->m_defOff =(char *)&cr.m_refs_numToGenerate - x;
m->m_priv = 0;
m->m_smin = 0;
m->m_flags = PF_HIDDEN | PF_NOSAVE;
m->m_page = PAGE_NONE;
m->m_obj = OBJ_SI;
m++;
m->m_title = "number of reference pages to display";
m->m_desc = "What is the number of "
"reference pages to display per query?";
@ -8312,6 +8334,7 @@ void Parms::init ( ) {
m->m_cgi = "iw";
m->m_page = PAGE_NONE;
m->m_obj = OBJ_SI;
m->m_def = "200";
m++;
m->m_title = "image height";
@ -8322,6 +8345,7 @@ void Parms::init ( ) {
m->m_cgi = "ih";
m->m_page = PAGE_NONE;
m->m_obj = OBJ_SI;
m->m_def = "200";
m++;
// m->m_title = "password";
@ -13352,6 +13376,7 @@ void Parms::init ( ) {
m->m_cgi = "urls";
m->m_page = PAGE_ADDURL2;
m->m_obj = OBJ_NONE;
m->m_def = NULL;
m->m_type = TYPE_FILEUPLOADBUTTON;
m++;
*/
@ -14152,7 +14177,7 @@ void Parms::init ( ) {
m->m_cgi = "dsrt";
m->m_off = (char *)&cr.m_docsToScanForTopics - x;
m->m_type = TYPE_LONG;
m->m_def = "300";
m->m_def = "30";
m->m_flags = PF_API;
m->m_page = PAGE_SEARCH;
m->m_obj = OBJ_COLL;

2
Parms.h

@ -368,7 +368,7 @@ class Parms {
bool isHtmlEncoded , bool fromRequest ) ;
void setToDefault ( char *THIS , char objType ,
CollectionRec *argcr = NULL ) ;
CollectionRec *argcr );//= NULL ) ;
bool setFromFile ( void *THIS ,
char *filename ,

2
SearchInput.cpp

@ -287,7 +287,7 @@ bool SearchInput::set ( TcpSocket *sock , HttpRequest *r ) { //, Query *q ) {
}
// and set from the http request. will set m_coll, etc.
g_parms.setToDefault ( (char *)this , OBJ_SI );
g_parms.setToDefault ( (char *)this , OBJ_SI , cr );
///////

5
Sections.cpp

@ -17162,6 +17162,11 @@ bool Sectiondb::init ( ) {
// . just hard-code 5MB for now
long pcmem = 5000000; // = g_conf.m_sectiondbMaxDiskPageCacheMem;
// do not use for now i think we use posdb and store the 32bit
// val in the key for facet type stuff
pcmem = 0;
maxTreeMem = 100000;
maxTreeNodes = 1000;
/*
key128_t k;

33
XmlDoc.cpp

@ -2090,8 +2090,8 @@ void XmlDoc::getRevisedSpiderRequest ( SpiderRequest *revisedReq ) {
// this must be valid for us of course
if ( ! m_firstIpValid ) { char *xx=NULL;*xx=0; }
// wtf?
if ( m_firstIp == 0 || m_firstIp == -1 ) { char *xx=NULL;*xx=0; }
// wtf? it might be invalid!!! parent caller will handle it...
//if ( m_firstIp == 0 || m_firstIp == -1 ) { char *xx=NULL;*xx=0; }
// store the real ip in there now
revisedReq->m_firstIp = m_firstIp;
@ -22122,8 +22122,6 @@ char *XmlDoc::getMetaList ( bool forDelete ) {
setStatus("adding spider request");
// checkpoint
saved = m_p;
// copy it
*m_p++ = RDB_SPIDERDB;
// store it here
SpiderRequest revisedReq;
// this fills it in
@ -22141,6 +22139,8 @@ char *XmlDoc::getMetaList ( bool forDelete ) {
goto skipNewAdd2;
}
// copy it
*m_p++ = RDB_SPIDERDB;
// store it back
memcpy ( m_p , &revisedReq , revisedReq.getRecSize() );
// skip over it
@ -28530,15 +28530,15 @@ Msg20Reply *XmlDoc::getMsg20Reply ( ) {
reply->size_content = size_utf8Content;
}
if ( m_req->m_getSectionVotingInfo && m_tmpBuf3.getCapacity() <= 0 ) {
Sections *ss = getSections();
if ( ! ss || ss == (void *)-1) return (Msg20Reply *)ss;
// will at least store a \0 in there, but will not count
// as part of the m_tmpBuf.length()
ss->printVotingInfoInJSON ( &m_tmpBuf3 );
reply-> ptr_sectionVotingInfo = m_tmpBuf3.getBufStart();
reply->size_sectionVotingInfo = m_tmpBuf3.length() + 1;
}
// if ( m_req->m_getSectionVotingInfo && m_tmpBuf3.getCapacity() <=0) {
// Sections *ss = getSections();
// if ( ! ss || ss == (void *)-1) return (Msg20Reply *)ss;
// // will at least store a \0 in there, but will not count
// // as part of the m_tmpBuf.length()
// ss->printVotingInfoInJSON ( &m_tmpBuf3 );
// reply-> ptr_sectionVotingInfo = m_tmpBuf3.getBufStart();
// reply->size_sectionVotingInfo = m_tmpBuf3.length() + 1;
// }
// breathe
QUICKPOLL ( m_niceness );
@ -29697,7 +29697,7 @@ SafeBuf *XmlDoc::getSampleForGigabits ( ) {
e[-2] == '!' ) )
endsInPeriod = true;
long off = reply.length();
//long off = reply.length();
if ( ! reply.safePrintFilterTagsAndLines ( p , e-p ,false ) )
return NULL;
@ -29734,8 +29734,9 @@ SafeBuf *XmlDoc::getSampleForGigabits ( ) {
*pc = '\0';
// debug
char *x = reply.getBufStart() + off;
log("fastfact: %s",x);
//char *x = reply.getBufStart() + off;
// turn off fast fact debug for now
//log("fastfact: fastfact: %s",x);
// revert back to |
*pc = '|';