[Dxspider-support] negative strlen

[charlie carroll]

Dirk:
I, and I'm sure a few others, am interested in any reasons for moving
from Perl ver. 5.6.1 to ver. 5.8.x.  You also mentioned even going from
5.8.2 to 5.8.3.  Is there some significant functionality improvement? 
Why change?

Today I ran a little experiment.  I know others have had problems
installing ver. 5.8.x of Perl and then running with Spider.  I took one
of my test PCs that was running Perl ver. 5.6.1 and Spider 57.25x and
changed to Perl ver. 5.8.2.  This PC has W2K, SP4 installed.

With very little difficulty I was able to get this up and running.  I
used:
Perl 5.8.2.808
Time-Hires-1.49
TimeDate-1.16
Net-Telnet-3.03
DB_File-1.806
Data-Dumper is now part of the 5.8.2 package.
Digest::SHA is also part of 5.8.2.

Please note, everything I've described is for a Windows installation. 
The CPAN modules all came from the Windows 8xx series and not the 6xx
series which is pointed to by the link in the Installation manual.  I
also took great care to uninstall ver. 5.6.1 and then manually delete
the c:\perl directory which is called for in the 5.8.x installation
notes.  Using the new 5.8.x version and the old 6xx CPAN modules is
probably the source of a lot of installation problems. 

The one issue is Storable.  I was not able to find an 8xx version of
Storable.  However, Spider will run, but with a small problem at the
beginning.  I used a clean install of Spider and did a create_sysop.pl. 
On startup, Spider halts with the "RTFM error" telling me to run
create_sysop.  After running create_sysop.pl the second time, operation
seems normal.  I believe this may have something to do with detecting
Storable in DXUser.pm

One last thing, I did note that the usual messages that show when
cluster.pl is first started...

"EINPROGRESS" is not exported by the Errno module at /spider/perl/Msg.pm
line 56
"EWOULDBLOCK" is not exported by the Errno module at /spider/perl/Msg.pm
line 56

have now changed to...

Constant subroutine Msg::EINPROGRESS redefined at (eval 7) line 1.
Prototype mismatch: sub Msg::EINPROGRESS () vs none at (eval 7) line 1.
Constant subroutine Msg::EAGAIN redefined at (eval 8) line 1.
Prototype mismatch: sub Msg::EAGAIN () vs none at (eval 8) line 1.

Probably not a big deal, but a difference.

73 charlie, k1xx