Hi all! My OS: Redhat 7.1 on Pentium with 128Mb Software: MnoGoSearch 3.2.3, mysql 3.23.46
During testing MnoGoSearch I got some problem. As I understand this due protocol mis-correspondence between searchd and search.cgi. My problem is: the search.cgi don't respond (freezes forever) then query should give a non-empty reply having searchd switched on. Then the query give a zero result, it displays correctly. search.htm: DBAddr mysql://login:passowrd@localhost/mnoGoSearch/ DBMode cache LocalCharset koi8-r BrowserCharset koi8-r StoredAddr localhost:7003 (others are defaults) searchd.htm: Listen 7003 DBAddr mysql://login:passowrd@localhost/mnoGoSearch/ DBMode cache LocalCharset koi8-r (other commented out) The bug looks this. I submit a search form, get the first match and wait forever. The last lines I receives from search.cgi: --- # ./search.cgi america ... <b>1.</b> <a href="http://ahome/America1/outside.htm"; TARGET="_blank"><b>������ �������������� ���� �� ��������� ���</b></a> [<b>302055424</b>]<DD> [ ������� ] [ ������������ ] [ ������������ ] [ ���.������� ] [ ������ ������� ] [ ���.���� ��� ��� ] [ ���.������ � ��� ] [ ��������� ] [ "���� ����" ] [ ���������� ���.������� ] [ ������� ] [ ���������� ] [ ����� ���������� ] [ "�������� ��...<BR> <UL> <li><A HREF="http://ahome/America1/outside.htm"; TARGET="_blank">http://ahome/America1/outside.h...</A> (text/html) Mon, 11 Jun 2001 23:50:40 GMT, 53106 bytes <BR> </UL> </DL> --- The searchd writes: --- searchd[2369]: Connect 127.0.0.1 searchd[2369]: Allow by default searchd[2372]: Received header cmd=3 len=35 searchd[2372]: Received words len=35 words='q=america&BrowserCharset=koi8-r&IP='searchd[2372]: Query: america [Charset: koi8-r] Start UdmFindCache('america') Reading limits... Done (0.00) Reading .wrd files (1 words)... Done (0.00) Merging (1 groups, 47 urls)... Done (0.00) nwrd=47 nwrd1=47 Grouping by url_id... Done (0.00) Stop UdmFindCache() - 0.01 searchd[2372]: Sent total_found packet 22 bytes buf='Total_found=45' searchd[2372]: Sent withoffset packet 8 bytes searchd[2372]: Sent wordinfo packet 21 bytes searchd[2372]: Sent WWL packet 79 bytes cmd=8 len=71 nwords=1 searchd[2372]: Sent words packet 368 bytes cmd=3 len=360 nwords=45 searchd[2372]: Waiting for command header searchd[2372]: Received header cmd=5 len=80 searchd[2372]: Received DOCINFO command len=80 ndocs=20 searchd[2372]: Sent doc_info packet 10820 bytes searchd[2369]: Connect 127.0.0.1 searchd[2369]: Allow by default searchd[2372]: Waiting for command header searchd[2372]: Received header cmd=6 len=4 searchd[2372]: Received Clones request for crc32=-1875626605 searchd[2372]: Sent clones packet 16 bytes cmd=3 len=8 nclones=1 searchd[2372]: Waiting for command header searchd[2374]: Waiting for command header searchd[2372]: Received header cmd=5 len=4 searchd[2372]: Received DOCINFO command len=4 ndocs=1 searchd[2372]: Sent doc_info packet 637 bytes searchd[2372]: Waiting for command header --- a here it waiting, waiting, waiting... Here also some tracing. That the search.cgi want to: --- ... 2397 socket(PF_INET, SOCK_STREAM, IPPROTO_IP) = 4 2397 connect(4, {sin_family=AF_INET, sin_port=htons(7003), sin_addr=inet_addr("127.0.0.1")}}, 16) = 0 2397 send(4, "F", 1, 0) = 1 ... (1) 2397 send(4, "icy\242", 4, 0) = 4 2397 recv(4, (2) --- (1) - here we already go http output (2) - here we hang until timeout At this time the searchd: --- 2409 send(3, "<190>Jan 4 02:48:32 searchd[240"..., 63, 0) = 63 2409 send(3, "<190>Jan 4 02:48:32 searchd[240"..., 74, 0) = 74 2409 recv(5, "]\0\0\0", 4, 0) = 4 2409 send(5, "\5\0\0\0u\2\0\0", 8, 0) = 8 2409 send(5, "DI=93\tDU=http://ahome/A";..., 629, 0) = 629 2409 send(3, "<190>Jan 4 02:48:32 searchd[240"..., 66, 0) = 66 2409 send(3, "<190>Jan 4 02:48:32 searchd[240"..., 62, 0) = 62 2409 recv(5, <unfinished ...> 2411 <... recv resumed> "icy\242", 8, 0) = 4 (1) 2411 recv(5, "", 8, 0) = 0 (2) --- (1) - we've got the shit, the search.cgi sent us (2) - here we'll wait infinitely So, that we got? Two processes are reading from the same socket. Not nice. Then I switch off the searchd in search.htm all works fine, but the performance of the search with ispell things is very bad. Could anyone help me to find out, what is wrong? May be this is a bug? --Anton ___________________________________________ If you want to unsubscribe send "unsubscribe general" to [EMAIL PROTECTED]
