From:	CSBVAX::MRGATE!@KL.SRI.Com:TOLLIVER%ORN.MFENET@NMFECC.ARPA@SMTP  9-AUG-1987 11:27
To:	EVERHART
Subj:	Problems defining MAIL$EDIT to be CALLABLE_TPU

Received: from NMFECC.ARPA ([128.115.3.7].#Internet) by KL.SRI.COM with TCP; Sat 8 Aug 87 14:57:24-PDT
Received: from orn.mfenet by ccx.mfenet with Tell via MfeNet ;
	Sat, 8 Aug 87 14:54:17 PDT
Date:	Sat, 8 Aug 87 14:54:17 PDT
From:     TOLLIVER%ORN.MFENET@NMFECC.ARPA
Message-Id: <870808145417.0fh@NMFECC.ARPA>
Subject: Problems defining MAIL$EDIT to be CALLABLE_TPU
To:       INFO-VAX@KL.SRI.COM
Comment: From TOLLIVER@ORN.MFENET on  8-AUG-1987 17:57:15.55 EDT

I recall seeing a couple of messages a few months ago about how to define
MAIL$EDIT to be CALLABLE_TPU in order to invoke TPU as the editor of choice
from within MAIL.  Well, that works fine.  Now suppose that one is using
a private TPU$SECTION file by having properly defined the TPUSECINI logical
name to be that private file.  That still works fine.

Now for the problem:  Suppose one's private section file uses the TPU
built-in CALL_USER to invoke some external routine whose identity is
defined by the TPU$CALLUSER logical name.  That works fine normally, i.e.,
when invoking TPU directly from DCL via EDIT/TPU.  But when trying MAIL/EDIT,
the proper section file is picked up (i.e., properly translating TPUSECINI)
but "it" doesn't seem to know about the TPU$CALLUSER definition and
complains of (approximately) "error activating SYS$LIBRARY:TPU$CALLUSER.EXE"
and then says something like "Could not find a routine for CALL_USER to
invoke."

In summary, "it" (i.e., callable_TPU (or somebody)), can translate TPUSECINI
from the process table, but it seems unable to translate TPU$CALLUSER from
the same table.

Has anyone else seen this problem?  More importantly, does anyone
understand it and/or have a solution?

Thanks,

John Tolliver (TOLLIVER%ORN.MFENET@NMFECC.ARPA)