[9fans] fossil consoles

[9fans] fossil consoles

[Fco.J.Ballesteros]

Is there any plan to split the output queue into several ones so that each
console has its own one?

I was trying to get a script done to advance the epoch and, in the
process, trying to get something like kfscmd done.  I'd just like to
send a command to a fossil srv file and receive the command output
(without the output from other consoles merged).

I was about going to do it but thought that perhaps it'd be better to ask
here. Maybe there's a better way to do this in fossil. I still think in
kfs terms.

thanks a lot

Re: [9fans] fossil consoles

[Russ Cox]

I'm torn about the console.  I like the fact that
everyone sees everyone else, but I also understand
that it makes programmatic access to console commands
difficult.  At the same time, I really really really
want to see the programmatic access on the main console.

Re: [9fans] fossil consoles

[Fco.J.Ballesteros]

It was the output in `broadcast mode' to all consoles what was
bothering me. But since it seems to be a feature and not a bug,
I just tried this script which words well enough for controlling
purposes. The script depends on a /sys/log/fossil file that gets
a copy of the fossil console (eg. you have a
cat /srv/fscmd >/sys/log/fossil).

#!/bin/rc
# usage: fscmd `fossilcons(8) command'

if (! test -e /srv/fscmd){
	echo '/srv/fscmd does not exist' >[1=2]
	exit 'srv'
}

oldsz=`{ls -l /sys/log/fossil | awk -F' ' '{print $6}'}
oldsz=`{echo $oldsz '+' `{echo $* | wc -c} | hoc}

echo $* >>/srv/fscmd

{ dd -if /sys/log/fossil -bs 1 -skip $oldsz >[2] /dev/null ;
  echo
}  | grep -v '^[a-z]*:'




>> I'm torn about the console.  I like the fact that
>> everyone sees everyone else, but I also understand
>> that it makes programmatic access to console commands
>> difficult.  At the same time, I really really really
>> want to see the programmatic access on the main console.
> 
> what about letting fossil provide a 'ctl' tree, e.g.
> 
> 	mount /srv/fossil /n/fossilctl ctl
> 
> which would provide access to fossil metainformation
> (e.g. usage info) and process control requests?
> it could also provide controlled access to the fossil console.

Re: [9fans] fossil consoles

[Fco.J.Ballesteros]

It was the output in `broadcast mode' to all consoles what was
bothering me. But since it seems to be a feature and not a bug,
I just tried this script which words well enough for controlling
purposes. The script depends on a /sys/log/fossil file that gets
a copy of the fossil console (eg. you have a
cat /srv/fscmd >/sys/log/fossil).

#!/bin/rc
# usage: fscmd `fossilcons(8) command'

if (! test -e /srv/fscmd){
	echo '/srv/fscmd does not exist' >[1=2]
	exit 'srv'
}

oldsz=`{ls -l /sys/log/fossil | awk -F' ' '{print $6}'}
oldsz=`{echo $oldsz '+' `{echo $* | wc -c} | hoc}

echo $* >>/srv/fscmd

{ dd -if /sys/log/fossil -bs 1 -skip $oldsz >[2] /dev/null ;
  echo
}  | grep -v '^[a-z]*:'




>> I'm torn about the console.  I like the fact that
>> everyone sees everyone else, but I also understand
>> that it makes programmatic access to console commands
>> difficult.  At the same time, I really really really
>> want to see the programmatic access on the main console.
> 
> what about letting fossil provide a 'ctl' tree, e.g.
> 
> 	mount /srv/fossil /n/fossilctl ctl
> 
> which would provide access to fossil metainformation
> (e.g. usage info) and process control requests?
> it could also provide controlled access to the fossil console.

Re: [9fans] fossil consoles

[rog]

> I'm torn about the console.  I like the fact that
> everyone sees everyone else, but I also understand
> that it makes programmatic access to console commands
> difficult.  At the same time, I really really really
> want to see the programmatic access on the main console.

what about letting fossil provide a 'ctl' tree, e.g.

	mount /srv/fossil /n/fossilctl ctl

which would provide access to fossil metainformation
(e.g. usage info) and process control requests?
it could also provide controlled access to the fossil console.