Hi,

when starting rcapid via xinetd, rcapid searches for /etc/capi20.conf which doesn't exist.
'ln -s /etc/isdn/capi.conf /etc/capi20.conf' solves this issue.

For rcapid to work the undocumented option "TRACELEVEL 0" needs to be added in /etc/isdn/capi.conf.
'capiinfo' without "TRACELEVEL 0" gives:
         capi20.c: 164              CapiDebug():[capi20_isinstalled]: standard loop - module: rcapi
         capi20.c: 164              CapiDebug():[capi20_isinstalled]: standard loop - module: standard
         capi20.c: 164              CapiDebug():[capi20_isinstalled]: capi_fd: 4
This will also be sent to the RCAPI client. (wireshark trace can be provided)

Adding "TRACELEVEL" gives a working rcapid but leads to an other (minor) problem:
'capiinit' doesn't recognize this option, but work as intended:
         FATAL: Module TRACELEVEL not found.
         ERROR: failed to load driver TRACELEVEL

after looking to the source code, i found that
/etc/capi20.conf is the config for libcapi Options (REMOTE, TRACELEVEL, TRACEFILE)
and
/etc/isdn/capi.conf is the config for capiinit, defining the ISDN hardware

my config:
/etc/capi20.conf:TRACELEVEL 0
/etc/isdn/capi.conf:b1pci           b1.t4           DSS1    -       -       -       -       P2P
works without any error messages.

The need for "TRACELEVEL 0" comes from a changed default debug level:
see /usr/share/doc/libcapi20-3/README.Debian

Hello Christian,

thank you for your bug report.  As you found out already you were at
least partially going down the wrong route until you realised that
/etc/capi20.conf and /etc/isdn/capi.conf deal with completely different
things.  After that was cleared up you now claim that "The need for
"TRACELEVEL 0" comes from a changed default debug level:
see /usr/share/doc/libcapi20-3/README.Debian" which is definitely wrong.
 I wrote that README and I wrote it merely as a reminder to myself how
to recompile with debug options activated when necessary.  In the code
itself nothing was changed.  If that is not enough proof for you, then
I'll point out that I don't even have an /etc/capi20.conf file.

So, what is the essence of your bug ticket?  I can't see any substance
being left here and thus would like to close as invalid.

Regards

Rolf

Hello Christian,

thank you for your bug report.  As you found out already you were at
least partially going down the wrong route until you realised that
/etc/capi20.conf and /etc/isdn/capi.conf deal with completely different
things.  After that was cleared up you now claim that "The need for
"TRACELEVEL 0" comes from a changed default debug level:
see /usr/share/doc/libcapi20-3/README.Debian" which is definitely wrong.
 I wrote that README and I wrote it merely as a reminder to myself how
to recompile with debug options activated when necessary.  In the code
itself nothing was changed.  If that is not enough proof for you, then
I'll point out that I don't even have an /etc/capi20.conf file.

So, what is the essence of your bug ticket?  I can't see any substance
being left here and thus would like to close as invalid.

Regards

Rolf

Hello Rolf,

after reading my bug report again, i have to admit it's confusing.

I can't even find nDebugLevel in the sources. The debug/trace level is set
in capi20/capi20.c line 103:
static int tracelevel=0xff;

The first host provides RCAPI via xinetd. xinetd is configured according
to the rcapid man page. On this host i have to set "TRACELEVEL 0" in
/etc/capi20.conf.
The second host connects to this host. With "TRACELEVEL 0" capiinfo and
other applications works fine. Without, it hangs. This is because xinetd
sends all rcapid output (stdout/stderr - including debug info) through the
network. (network trace can be provided)

The need for "TRACELEVEL 0" persists, no matter if the code has ever changed.


Regards

Christian