Re: Configuring GoldEd+ Linux
  By: Matt Bedynek to Eric Renfro on Tue Jul 21 2015 01:07 am

 ER>> I have so many issues it's not even funny though. GoldEd+ doesn't
 ER>> understand properly the TERM=screen-256color, so I manually have to
 ER>> set it to TERM=screen for that execution. I'm fine with this, but,
 ER>> not find with hittnig F1 and getting "Can't find ~/golded.gel" for
 ER>> help. I'm also not fine trying to put in a name and it saying it
 ER>> can't read any nodelist, yet I've configured it in..

 MB> Do you store your golded configuration in a different location than the 
 MB> directory you execute it from?

No I do not. I actually utilized, and hacked up a little bit, the golded.spec
to build a CentOS 7 binary rpm package. Thus golded itself is installed to
/usr/bin/genlnx, and similar /usr-prefixed directories of the system like any
package. 
I use export GOLDED="$HOME/.golded.cfg" for my configuration path. 

 MB> I have the golded binaries in my search path and my environment setup as
 MB> such: 

 MB> export PATH=$PATH:/opt/fido/bin
 MB> export FIDOCONFIG=/opt/ftn/fido/husky/husky.conf
 MB> export LD_LIBRARY_PATH="/opt/fido/lib"
 MB> export GED=/opt/fido/etc/golded
 MB> alias vi=vim
 MB> alias ged=gedlnx

 MB> This configuration is stored in .profile which is executed when I change 
 MB> identity to the fido user (su - fido). I cannot run any aspect of my node
 MB> as any other user (this prevents screwing up permissions) because the
 MB> binaries are not in PATH.

 MB> Most paths defined in golded.cfg are relative. So I assume that the 
 MB> environment variable GED has precedence when absolute paths are not
 MB> explicitied defined in config.

I setup my golded.cfg to be absolute paths pretty much anywhere.

 ER>> If anyone can provide me some help on this all to at least try to
 ER>> get GoldEd+ LNX working for Netmail (since Synchronet message base
 ER>> support appears to be broken entirely), I'd be... So very exstatic.
 ER>> :D 

 MB> I would turn on logging for golded to see if it capture more specifically
 MB> any misconfiguration. It has been a while but I do recall testing earlier
 MB> versions with SMB. It worked. Have you tried to statically define an ara
 MB> using AREADEF?

I have not tried explicitely using AREADEF itself, I don't actually know what
the format would be specifically for that in GED+ for SMB. What I can say
however, when I use AreaFile Synchronet /sbbs/ctrl/msgs.cnf. what happends when
I actually attempt to open a message base in Synchronet, GOLDED writes an empty
file with a very VERY peculiar permission set. 
-r----x--t    1 bss bbs      0 Jul 21 09:08 agn_bbs.sdt
-r----x--t    1 bbs bbs     32 Jul 21 09:08 agn_bbs.shd
-r----x--t    1 bbs bbs      0 Jul 21 09:08 agn_bbs.sid

Properly, those files are actually named agora-agn_bbs.* which have the actual
files. It's missing the group. Then of course once those files mentioned above
exist, the next time to try to open the message base, it claims it's locked.
First time it's just an empty list. Unless I have AreaScan * on which then it
makes a whole horde of those sticky-bit files. 

 ER>> Any help would be greatly appreciated. So far I've had better luck
 ER>> (though still not 100%) with Husky Project's msged, it at least will
 ER>> let me search nodelists and send a netmail out. :)

 MB> I use husky with JMB. Both are working here for me.

JAM message bases? hehe. 

What was totally missed is why the nodelist portion of golded just completely
fails. I provide the NodePath and Nodelist nodelist.999, and it can't use them.
I mean, I'm fine IF I cannot use GED+ with SMB message bases. I really only
/need/ it for netmail, mostly because Synchronet does not actually let you do
specific things, right now anwyay. like specify WHICH exact AKA you want to use
when sending netmail, it just picks the first one matching the zone you're
writing to. As an RC for another net, I have 3 AKA in the same zone. :)

On another note, regarding Synchronet SMB message bases, I have tried all three
message base formats (two of which are technically compatible), Self-Packing
and Fast Allocation, and Hyper Allocation. Neither seems to function as they
should.

)))[Psi-Jack -//- Decker]
--- SBBSecho 2.27-Linux