[ From:	CRDGW2::CRDGW2::MRGATE::"SMTP::CRVAX.SRI.COM::RELAY-INFO-VAX"  1-NOV-1990 12:48:30.74  To:	MRGATE::"ARISIA::EVERHART" CC:	/ Subj:	re: Is it possible to talk without GRPNAM   * Received:  by crdgw1.ge.com (5.57/GE 1.76)( 	 id AA17217; Thu, 1 Nov 90 12:36:41 ESTX Received: From VENUS.YCC.YALE.EDU by CRVAX.SRI.COM with TCP; Thu,  1 NOV 90 08:39:45 PDTM Received: from BULLDOG.CS.YALE.EDU by Venus.YCC.Yale.Edu; Thu, 1 Nov 90 11:36   ESTO Received: from lrw.UUCP by BULLDOG.CS.YALE.EDU via UUCP; Thu, 1 Nov 90 11:27:25   ESTG Received: by lrw.UUCP (DECUS UUCP w/Smail); Thu,  1 Nov 90 10:52:34 EDT ! Date: Thu,  1 Nov 90 10:52:34 EDT ' From: Jerry Leichter <leichter@LRW.COM> 2 Subject: re: Is it possible to talk without GRPNAM To: INFO-VAX@KL.SRI.COM 4 Message-Id: <9011011627.AA20518@BULLDOG.CS.YALE.EDU>" X-Envelope-To: info-vax@kl.sri.COM  E 	If I log into two separate terminals with the same account, Is there G 	any way to communicate between them using a mailbox without having the 0 	GRPNAM (I think that is the spelling) priv set?  L Sure.  The ability to communicate through a mailbox is dependent only on theM access rights you need to get to that mailbox.  If both processes are running J under the same UIC, they will both gain access via the OWNER protection ofI the mailbox, which normally will be set to allow access.  You can set the M protection to allow or deny access to GROUP, WORLD, and SYSTEM, or you can be " even more selective an use an ACL.  N So, where does GRPNAM come in?  In order to communicate through a mailbox, youJ have to know WHICH mailbox to communicate through.  Normally, you define aN logical name in some appropriate logical table accessible to all the processesN that need to use the mailbox.  In general, the logical is put in the job tableL for temporary mailboxes (and is thus accessible throughout the job tree) andM in the system table for permanent ones.  That won't work here, since there is M normally no table shared by the two independently-started jobs.  There are at . least three alternatives you can use, however:  F 	- Have the creator of the mailbox store its number in a file.  Every-@ 		one else reads the file.  Since the file need only be accessed? 		when a process wants to open the mailbox, the overhead should  		not be excessive.   A 	- Create a system-wide table accessible to both processes.  This ; 		requires privileges, but need only be done once, probably < 		at system startup.  The table would be a system-wide table> 		accessible only to the particular UIC involved.  The logical= 		defining the mailbox can be placed directly into that table - 		by simply redefining LNM$TEMPORARY_MAILBOX.   F 	- Hack approach.  LNM$JOB translates to an a system-wide table, owned> 		by your UIC, of the form LNM$JOB_xxxxxxxx, where the x's are? 		hex digits.  (They actually are the hex representation of the ? 		address of the JIB - Job Information Block - for the job.  It = 		would have made things SO much simpler if they were the hex @ 		representation of the PID of the process at the top of the job@ 		tree....)  So, while you normally have no reason to do it, you? 		CAN access the job tables of any of your jobs on the system - > 		IF you can find them.  There's no easy way to find the JIB's? 		of your other jobs, nor can you do wildcard access to a logi- > 		cal name table from a program; but you CAN from DCL with the; 		SHOW LOGICAL command.  Try SHOW LOGICAL/STRUCTURE for the > 		quickest look - it shows you all tables that you have access@ 		to, so automatically sorts out which JOB_xxx tables are yours.   							-- Jerry   