I've been trying out the latest snapshot and I indexed a small
part of our web to try it out and have found that htsearch crashed
about half the time, particularly if you give it multiple words
in the search. The traceback is

#0  0x80dc1d1 in ?? ()
#1  0x400d2cb3 in DictionaryEntry::~DictionaryEntry (this=0x80dcba8,
__in_chrg=3) at Dictionary.cc:39
#2  0x400d3348 in Dictionary::Remove (this=0x80d3ba8, name=0x8099860
"114") at Dictionary.cc:240
#3  0x8050e56 in ResultList::remove (this=0x80d3ba8, id=114) at
ResultList.cc:76
#4  0x8055a89 in Parser::score (this=0x80c59a8, wordList=0x80c5ec0,
weight=1) at parser.cc:360
#5  0x8055387 in Parser::perform_push (this=0x80c59a8) at parser.cc:260
#6  0x8054e17 in Parser::factor (this=0x80c59a8, output=1) at
parser.cc:155
#7  0x8054d73 in Parser::term (this=0x80c59a8, output=1) at
parser.cc:117
#8  0x8054d02 in Parser::expr (this=0x80c59a8, output=1) at parser.cc:91
#9  0x8054de8 in Parser::factor (this=0x80c59a8, output=1) at
parser.cc:141
#10 0x8054d73 in Parser::term (this=0x80c59a8, output=1) at
parser.cc:117
#11 0x8054d02 in Parser::expr (this=0x80c59a8, output=1) at parser.cc:91
#12 0x8054b9a in Parser::fullexpr (this=0x80c59a8, output=1) at
parser.cc:54
#13 0x8056266 in Parser::parse (this=0x80c59a8, tokenList=0xbffff9c8,
resultMatches=@0x80d38a0) at parser.cc:563
#14 0x80546b0 in htsearch (wordfile=0x80d3878
"/opt/www/var/htdig/db.words.db", searchWords=@0xbffff9c8,
parser=0x80c59a8) at htsearch.cc:665
#15 0x8052af1 in main (ac=1, av=0xbffff9fc) at htsearch.cc:257


The offending entry seems to be

(gdb) p *next
$18 = {hash = 13, key = 0x80dc200 "��\032@��\032@@", value = 0x80dc1d0,
next = 0x0}

-- 
 Toivo Pedaste                        Email:  [EMAIL PROTECTED]
 University Computing Services,       Phone:  +61 8 9 380 2605
 University of Western Australia      Fax:    +61 8 9 380 1109
"The time has come", the Walrus said, "to talk of many things"...

------------------------------------
To unsubscribe from the htdig3-dev mailing list, send a message to
[EMAIL PROTECTED] containing the single word "unsubscribe" in
the SUBJECT of the message.

Reply via email to