Possible cause? Berkeley-DB v1.85 btree failure?

Possible cause? Berkeley-DB v1.85 btree failure?

Post by dk smi » Thu, 02 Oct 1997 04:00:00



Hello,

I hope someone has seen this when building Berkeley-DB 1.85 on Solaris
2.5.1. I plan to use the resulting libdb.a with sendmail 8.8.*.

The library failed test 12 (btree). Is this normal? I can use the lib to
build database files (makemap) in the hash format but obviously the btree
functionality should not be trusted.

Does anyone have a recommendation of how to correct this? Compiler is
newly installed gcc 2.7.3.2.

Thanks,
dk smith

--
dk smith                                     MediaWeb, Inc.

 
 
 

Possible cause? Berkeley-DB v1.85 btree failure?

Post by dk smi » Thu, 02 Oct 1997 04:00:00




> I hope someone has seen this when building Berkeley-DB 1.85 on Solaris
> 2.5.1. I plan to use the resulting libdb.a with sendmail 8.8.*.

> The library failed test 12 (btree). Is this normal? I can use the lib to
> build database files (makemap) in the hash format but obviously the btree
> functionality should not be trusted.

I have one more piece of information. The test program is complaining
because the program called "rev" cannot be found. Is there a substitute
that I could use  in order to test the makemap program?

thanks,
dk

--
dk smith                                     MediaWeb, Inc.


 
 
 

Possible cause? Berkeley-DB v1.85 btree failure?

Post by Dave Carriga » Thu, 02 Oct 1997 04:00:00


 dk> Hello, I hope someone has seen this when building Berkeley-DB
 dk> 1.85 on Solaris 2.5.1. I plan to use the resulting libdb.a with
 dk> sendmail 8.8.*.

Don't forget to remove the ndbm functionality out of libdb (remove the
ndbm.o from the OBJ1 macro in the Makefile).

 dk> The library failed test 12 (btree). Is this normal? I can use the
 dk> lib to build database files (makemap) in the hash format but
 dk> obviously the btree functionality should not be trusted.

The failed test is due to there being no rev program on Solaris. I
don't know what rev does, but I think it's supposed to print a file in
reverse order. GNU fileutils provides a program called tac which does
this; you may want to try it. FWIW, we've been using sendmail and
libdb btrees for a couple of years now with no problems. I've also
done a lot of programming with Perl's DB_File module and creating
btree databases with no data loss, so I wouldn't worry too much about
this.

--
Dave Carrigan                            | Yow! HUMAN REPLICAS are inserted
Unix Specialist, Technology Services     | into VATS of NUTRITIONAL YEAST...
Interprovincial Pipe Line Inc.           |
Edmonton, Alberta, Canada                |

 
 
 

Possible cause? Berkeley-DB v1.85 btree failure?

Post by Dan Strombe » Thu, 02 Oct 1997 04:00:00





<
< dk> Hello, I hope someone has seen this when building Berkeley-DB
< dk> 1.85 on Solaris 2.5.1. I plan to use the resulting libdb.a with
< dk> sendmail 8.8.*.
<
<Don't forget to remove the ndbm functionality out of libdb (remove the
<ndbm.o from the OBJ1 macro in the Makefile).
<
< dk> The library failed test 12 (btree). Is this normal? I can use the
< dk> lib to build database files (makemap) in the hash format but
< dk> obviously the btree functionality should not be trusted.
<
<The failed test is due to there being no rev program on Solaris. I
<don't know what rev does, but I think it's supposed to print a file in
<reverse order. GNU fileutils provides a program called tac which does
<this; you may want to try it.

rev reverses the characters in each line of input.

tac reverses the order of the lines in a file.

EG, rev on:
        123
        456

is
        321
        654

while tac gives
        456
        123

 
 
 

1. Berkeley db v1.85 compile problem.

I have been trying to compile Berkeley db v1.85 on my Solaris x86 v2.5
box for use with Sendmail v8.8.5 virtual user tables and am running into
troubles.

I've done this on SunOS v4 boxes, but never on Solaris.

Anyway, when I try to compile the 'test' program (not the lib itself) I
get complaints about missing include files.  I didn't get any such
complaints when I compiled the library itself.

First it was db.h, which I obviously had just forgotten to copy into
/usr/include like the README says.

Then is was sys/cdefs.h, which I found in the includ directory of the db
directory structure that I copied into /usr/include/sys.

Now, I am getting complaints about useful.h and conf.h, and I have no
idea what they are or where they come from.

Any ideas?


responses)

-Michael F. Oryl Jr.
 President, Intersphere Communications Ltd.

 http://www.intersphere.com/    Corporate
 http://www.burn.com/          Personal

2. Windows www Browsers reports "transmission interrupt" errors when run against Linux httpd web site.

3. Question about Berkeley DB (btree's)

4. setting path for superuser

5. question on using db-1.85 with freebsd

6. Replacing No Access Page in Apache

7. solaris 7(x86) and db-1.85

8. converting csh script to ksh

9. Apache compile - Error: None of Berkeley-DB 1.x, 2.x or 3.x

10. Problems compiling libdb.a v1.85 with 2.1.5?

11. possible to setup a 85 key keyboard in linux?

12. db for sendmail to make a btree: where?

13. install - updatt_all failures 4.3.1 - 4.3.3 possible causes