Has anyone else seen this error in the current snapshots?
What executable does it happen on?
I get it on htdig (the libhtdig version)
This error could happen on htmerge & htpurge.. any executable that is
changing the wordDB could corrupt the DBfile (if that's the problem).
Thanks
--
Neal Richter
Knowledgebase Developer
RightNow Technologies, Inc.
Customer Service for Every Web Site
Office: 406-522-1485
---------- Forwarded message ----------
Date: Thu, 8 Aug 2002 13:47:58 -0500
From: [EMAIL PROTECTED]
To: Neal Richter <[EMAIL PROTECTED]>
Subject: Re: [htdig-dev] Compressor::get_vals
You must have more debug in there, I don't see Exporting Document, nor can
I find it in the code 07/28, 08/04 or mifluz 08/04 snapshots. I did add
the CDB debug and now it doesn't die ...
a
Andy Bach, Sys. Mangler
Internet: [EMAIL PROTECTED]
VOICE: (608) 261-5738 FAX 264-5030
"To understand recursion, we must first understand recursion."
Neal Richter <[EMAIL PROTECTED]>
08/08/02 01:11 PM
To: [EMAIL PROTECTED]
cc: [EMAIL PROTECTED]
Subject: Re: [htdig-dev] Compressor::get_vals
Andy,
Please try this for me:
Add this line to db/mp_cmpr.c before the call to cmpr_info->uncompress
printf("CDB___memp_cmpr_read: uncompress page [%d] \n", first_pgno);
Add this line to db/mp_cmpr.c before the call to cmpr_info->compress
printf("CDB___memp_cmpr_write: compress page [%d] \n", db_io->pgno);
Then run your htdig again and post the output.
Here's the output I get:
Exporting Document [010802-002096] [99955]
Exporting Document [010802-002106] [99956]
Exporting Document [010802-002109] [99957]
Exporting Document [010802-002119] [99958]
Exporting Document [010802-002122] [99959]
CDB___memp_cmpr_write: compress page [236291]
Exporting Document [010802-002145] [99960]
Exporting Document [010802-002190] [99961]
Exporting Document [010802-002213] [99962]
Exporting Document [010802-002223] [99963]
Exporting Document [010802-002229] [99964]
--
Exporting Document [010803-000314] [100195]
Exporting Document [010803-000332] [100196]
Exporting Document [010803-000353] [100197]
Exporting Document [010803-000360] [100198]
Exporting Document [010803-000361] [100199]
CDB___memp_cmpr_read: uncompress page [236291]
Exporting Document [010803-000362] [100200]
Exporting Document [010803-000363] [100201]
Exporting Document [010803-000364] [100202]
Exporting Document [010803-000365] [100203]
Exporting Document [010803-000367] [100204]
--
Exporting Document [010804-001475] [101212]
Exporting Document [010804-001476] [101213]
Exporting Document [010804-001477] [101214]
Exporting Document [010804-001478] [101215]
Exporting Document [010804-001479] [101216]
CDB___memp_cmpr_read: uncompress page [236291]
Note that the page is writen, read, and read again (it dumps on the
second read). This is a bit unexpected, I would have expected a write
(with an error), then a read that dumps.
I am also noticing that it is somewhat inconsistent in when it dumps..
This leads me to believe its a memory corruption issue... I'll run
Insure++ on it, and wait for it to finish ;-)
Thanks!
On Wed, 7 Aug 2002 [EMAIL PROTECTED] wrote:
> Hey.
>
> I was working w/ the 3.2.0b4 7/28 snapshot and hit:
> FATAL ERROR:Compressor::get_vals invalid comptype
> FATAL ERROR at file:WordBitCompress.cc line:827 !!!
> Segmentation Fault - core dumped
>
> not sure where it was from (can't reproduce it) but I went to
> htword/WordBitCompress.cc and noticed that, though the run was w/ -vvvvv
> the:
> if(verbose)printf("get_vals:comptype:%d\n",comptype);
>
> stmts didn't show up in the output. I can only find 2 places where new
> Compressors are created, in htword/WordDBPage.cc and neithor of them
seems
> to pass along the verbose flag. I'm out of my depth here, but the
> constructor for Compressors:
> Compressor():BitStream()
> {
> verbose=0;
> }
> Compressor(int size0):BitStream(size0)
> {
> verbose=0;
> }
>
> seem to be setting always to zero.
>
> a
>
> Andy Bach, Sys. Mangler
> Internet: [EMAIL PROTECTED]
> VOICE: (608) 261-5738 FAX 264-5030
>
> "To understand recursion, we must first understand recursion."
>
>
> -------------------------------------------------------
> This sf.net email is sponsored by:ThinkGeek
> Welcome to geek heaven.
> http://thinkgeek.com/sf
> _______________________________________________
> htdig-dev mailing list
> [EMAIL PROTECTED]
> https://lists.sourceforge.net/lists/listinfo/htdig-dev
>
--
Neal Richter
Knowledgebase Developer
RightNow Technologies, Inc.
Customer Service for Every Web Site
Office: 406-522-1485
-------------------------------------------------------
This sf.net email is sponsored by:ThinkGeek
Welcome to geek heaven.
http://thinkgeek.com/sf
_______________________________________________
htdig-dev mailing list
[EMAIL PROTECTED]
https://lists.sourceforge.net/lists/listinfo/htdig-dev
