Hello!
If I index the http://niche.elte.hu/doc/packages/qt/html/annotated.html
page, my indexer core dump.
(gdb) bt
#0 0x3ff8148a6a8 in strcasecmp () from /usr/shlib/libc.so
#1 0x12004c9a0 in UdmParseHtml (Indexer=0x14000ee00,
CurSrv=0x140013d48,
CurURL=0x11fffa290, Doc=0x140048b80) at
Hi!
NE I tried to index ludens.elte.hu, but it answer: Can't connect to
NE ludens.elte.hu:80. If I see with netscape or debug with telnet, it works
NE well. I have debuged it for several hours, but nothing found.
NE I can index another site of corse.
NE What is (non)blocking mode, in
Hello!
I couldn't compile the source:
cc: Error: udmutils.c, line 271: Invalid statement.
I put a ';' char at the end of prev line (after `default:') and so I can
compile it.
Bye,
Ned
P.s.: I use nativ DEC compiler, and ther are a several new warning due
signed and unsigned char* missing.
Hello!
What about DeleteNoServer? I set to no, but indexer said:
[1] http://www.chem.elte.hu/General/UPS/_vti_cnf/LACI.HTM
[1] No 'Server' command for url... deleted.
[1] Done (2 seconds)
Of course I not set neither Server nor Realm. Do it means that it is an
unusable option?
Thanks,
Ned
Hi!
I tried to index ludens.elte.hu, but it answer: Can't connect to
ludens.elte.hu:80. If I see with netscape or debug with telnet, it works
well. I have debuged it for several hours, but nothing found.
I can index another site of corse.
What is (non)blocking mode, in connection part?
Bye,
Hi!
The current is 3.1.9pre12 and it is working well.
I wait for the next (pre)release, because it needs a several patch, but
Alex said to me that their version is as different as he can't patch the
source easy with my patch.
I hope the pre13 will be released soon!
Bye,
Ned
__
I have some question about indexer.conf in version 3.1.9-pre12
# Examples
# Allow everything:
#Allow .*
It looks like not fixed from previous version.
# Note that '?' has special meaning in "String" match type. Please use
# "Regex" to describe documents with '?' sign in URL.
Only the '?' has
What do mnogosearch if a url match to a CheckOnly rule? If I understand
well, it will be inserted into the url table only, but nothing else.
What is its sense? Will it change the search or indexing results in any
way?
indexer will send "HEAD" instead of "GET" request to take things like
When I idexing only http://hali.elte.hu/index.html the last log is:
[1] Error in URL: 'mailto:[EMAIL PROTECTED]'
when the source has:
a href="mailto:[EMAIL PROTECTED]"
It is not downloadable URL type.
Of course, I understood it, but it looks like if it would be an error,
not only a
Hello!
I tried the 3.1.9-pre12. I have a rule:
Allow */ *.htm *.html *.shtml *.phtml *.php *.txt *.ps *.eps *.pdf *.TXT
And when I indexing .../1.TXT in the log:
[1] Allow NoCase *.TXT
When I delete the *.TXT from rule it's skipped by indexer. If I Start
the rule with 'Case' then it works
Hello!
I couldn't compile the 3.1.9.pre12 version, because there is a snprintf
instead of udm_snprintf in ftp.c.
Please fix it in next (pre)release.
Thanks,
Ned
__
If you want to unsubscribe send "unsubscribe udmsearch"
to [EMAIL PROTECTED]
Hello!
My indexer(3.1.8) started indexing some dial-up site in my domain. I
found thats' referrer and I saw these are only robots.txt files:
| http://jkindla.dial.elte.hu/robots.txt |
http://ludens.elte.hu/robots.txt |
| http://denes.dial.elte.hu/robots.txt |
Hi!
I don't understand and I can't find documentation about realm, so I ask
about it:
If I use the realm option, it mean that indexer will fetch the zone
from the DNS and index all the servers in that domain (or zone)?
Or, what's the different from allow/server compinations and the realm
option?
Hello!
I must run the indexer in parralel, but I can't with threads. My problem
is this way is very slow, so I view the mysql logs, and I found
interesting logs. It looks like if the mnogosearch 3.1.8 reindex a page
first delete the words from database. It would be right, but it do word
by word:
Hi!
How can i search though database only by url and not using content? I
need mp3 search for example.
See the sample/ftpsearch.conf file!
Bye,
Ned
__
If you want to unsubscribe send "unsubscribe udmsearch"
to [EMAIL PROTECTED]
Hello!
I use the nativ C compiler of Tru64 UNIX to compile the 3.1.6-pre1
version. I got some errors:
cc -DHAVE_CONFIG_H -I. -I. -I../include -I../include
-I/usr/local/mysql/include/mysql -O5 -g3
-DUDM_CONF_DIR=\"/usr/local/udmsearch-3.1.6-pre1/etc\"
Hello!
I use the 3.1.6-pre1 version. It use gs to parse ps doc. There is a
document which can froze it. If I use nativ the gs (over pstotext) it
works fine, but if call it the indexer it works a lot then it gets into
idle state. When I atached to gs and indexer with gdb, I can get the
following
Has somebody any idea what may be wrong?
Thanks,
Ned
ned:~$ /usr/local/udmsearch/sbin/indexer
indexer from UdmSearch v.3.1.5-pre4/MySQL started with
'/usr/local/udmsearch-3.1.5-pre4//etc/indexer.conf'
[1] ftp://ftp.inf.elte.hu/pub/MI_nyelvek/
[1] Error: '#1100: Table 'dict2' was not locked
Hello!
I'm debugging the latest release (because a memory handling problem) and
I found two UdmCloneList function (in sql.c and files.c) with same
header but diferent content. It's correct or it was left in the code at
a venture?
Bye,
Ned
__
If you want to unsubscribe send
Hello!
checking whether make sets ${MAKE}... ./configure: make: not found
What is the value of MAKE shell variable? There's that file? Where is
the make file? ("whereis make")
Bye,
Ned
__
If you want to unsubscribe send "unsubscribe udmsearch"
to [EMAIL PROTECTED]
Hello!
I've installed udmsearch in Budapest, Hungary. My problem is that the
search.cgi doesn't find the words with comma characters like "hz". I
have tried many character set in the indexer.conf file, but it did'
work. What else should i set to let it work?
There isn't enough to set
Hello!
Question 3: I can't find any logic which compares the last_modified date
with anything else and forces a skip if the file is un-modified. I found
reference to last_modified in files.c, indexer.c, sql.c, and search.c.
Am I missing something?
This value is sent to HTTP server in
Hi!
If I turn on FollowOutside option in the indexer.conf the indexer core
dump after some seconds. If I turn off it the indexer work properly.
How can I set if I want to index only comp.* groups from news.elte.hu?
(Server news://)
Thanks,
Ned
__
If you want to unsubscribe
Hello!
If I turn on FollowOutside option in the indexer.conf the indexer core
dump after some seconds. If I turn off it the indexer work properly.
Can you send gdb "backtrace" output?
If you tell me how I can do it. (Unfortunetly I'm too beginner in code
debugging; at the moment I can
Hi!
I not compiled any sqlserver-support into udmsearch, I like to use the
built-in database-manager. Where can I set which directory are used to
save database? How must I set DBAddr variable or any else?
Thanks,
Ned
__
If you want to unsubscribe send "unsubscribe udmsearch"
to
25 matches
Mail list logo