1 INFO-VAX	Thu, 11 Nov 2004	Volume 2004 : Issue 627       Contents: Re: ANN: VMS Mosaic 3.8  Re: ANN: VMS Mosaic 3.8 # COBOL PAK for the Hobbyist License? ' Re: COBOL PAK for the Hobbyist License? ' Re: COBOL PAK for the Hobbyist License?  DCL suggestion for f$verify() ! Re: DCL suggestion for f$verify()  Re: DCPS Version Support! Dynamic FTP site and file locking > Re: Help with OPCOM and OPA0: and console vs. DW/motif display- Re: how to implement an append only log file? - Re: how to implement an append only log file? - Re: how to implement an append only log file? - Re: how to implement an append only log file? " Re: Need help with a PeeCee again. Oracle DST Setting.  Problem to Partition a disk 0 Real documentation on SYSGEN memory parameters ? Re: Use of FTP in .Com files3 [Q]Shutting down TCPIP Sockets with AST read posted   F ----------------------------------------------------------------------    Date: 11 Nov 2004 07:46:28 -0600; From: koehler@eisner.nospam.encompasserve.org (Bob Koehler)   Subject: Re: ANN: VMS Mosaic 3.83 Message-ID: <A15q7dnZ4tEj@eisner.encompasserve.org>   \ In article <41928DEF.7AE7EE00@teksavvy.com>, JF Mezei <jfmezei.spamnot@teksavvy.com> writes: > 
 > 		default :  > 		}  <- end of switch >   -    That's perfectly valid in C, C++, or Java.    ------------------------------  % Date: Thu, 11 Nov 2004 14:43:59 -0000 * From: "Richard Brodie" <R.Brodie@rl.ac.uk>  Subject: Re: ANN: VMS Mosaic 3.82 Message-ID: <cmvuof$ujg$1@blackmamba.itd.rl.ac.uk>  H "Bob Koehler" <koehler@eisner.nospam.encompasserve.org> wrote in message- news:A15q7dnZ4tEj@eisner.encompasserve.org...   
 > > default :  > > }  <- end of switch  > >  > / >    That's perfectly valid in C, C++, or Java.   ? Current versions of DECC, CXX and gcc give warnings if pressed. F A statement is required, even if a null one. It's a bit harsh, though,5 to call it a major bug when it often passes silently.    ------------------------------    Date: 11 Nov 2004 13:22:51 +0100( From: Andreas Davour <ante@update.uu.se>, Subject: COBOL PAK for the Hobbyist License?4 Message-ID: <cs9u0rwfuck.fsf@Psilocybe.Update.UU.SE>  F I booted by simh/VAX yesterday and since my licenses had expired I got: the new ones from Montagar and registered and loaded them.  G Now I notice that I no longer have any COBOL license anymore. Do anyone F know why it have dissapeared, and if somehow HP could be smooth-talked: into letting us have a COBOL PAK as hobbyists once again?   1 I actually had fun learning some COBOL on my VAX.    /andreas   --  A A: Because it fouls the order in which people normally read text. ' Q: Why is top-posting such a bad thing?  A: Top-posting. ; Q: What is the most annoying thing on usenet and in e-mail?    ------------------------------  % Date: Thu, 11 Nov 2004 06:05:17 -0800 # From: "Tom Linden" <tom@kednos.com> 0 Subject: Re: COBOL PAK for the Hobbyist License?( Message-ID: <opshaz23k4zgicya@hyrrokkin>  H On 11 Nov 2004 13:22:51 +0100, Andreas Davour <ante@update.uu.se> wrote:  H > I booted by simh/VAX yesterday and since my licenses had expired I got< > the new ones from Montagar and registered and loaded them.I > Now I notice that I no longer have any COBOL license anymore. Do anyone H > know why it have dissapeared, and if somehow HP could be smooth-talked; > into letting us have a COBOL PAK as hobbyists once again? 3 > I actually had fun learning some COBOL on my VAX.   3 Cobol fun? Try PL/I if you really want to have fun.    --  C Using Opera's revolutionary e-mail client: http://www.opera.com/m2/    ------------------------------  # Date: Thu, 11 Nov 2004 14:26:10 GMT & From: John Reagan <john.reagan@hp.com>0 Subject: Re: COBOL PAK for the Hobbyist License?/ Message-ID: <6EKkd.2782$4o2.6@news.cpqcorp.net>    Andreas Davour wrote: H > I booted by simh/VAX yesterday and since my licenses had expired I got< > the new ones from Montagar and registered and loaded them. > I > Now I notice that I no longer have any COBOL license anymore. Do anyone H > know why it have dissapeared, and if somehow HP could be smooth-talked< > into letting us have a COBOL PAK as hobbyists once again?  > 3 > I actually had fun learning some COBOL on my VAX.  > 
 > /andreas > H I doubt that it was requested by anybody here at HP.  I'll forward your . post to the COBOL PL to let him track it down.   --   John Reagan / HP Pascal/{A|I}MACRO for OpenVMS Project Leader  Hewlett-Packard Company    ------------------------------  % Date: Thu, 11 Nov 2004 05:39:34 -0500 - From: JF Mezei <jfmezei.spamnot@teksavvy.com> & Subject: DCL suggestion for f$verify(), Message-ID: <41934138.CAFA8DF9@teksavvy.com>  I Trying to debug a dcl procedure that has a loop that iterates a gazillion + times before aborting can be quite tedious.   J Therefore I would suggest an addition to the SET VERIFY and F$VERIFY which+ would create a second mode of verification:   U During execution, nothing is logged, but when an error occurs, 2 lines are displayed: G The original line in the file (and possibly its line number ?), and the ' expanded line with symbol substitution.   M I just spent a long time trying to debug AUTOGEN.COM to find out why it would K abort midway (turns out I had a space in modparams.dat at the wrong place).    ------------------------------  # Date: Thu, 11 Nov 2004 14:31:43 GMT & From: John Reagan <john.reagan@hp.com>* Subject: Re: DCL suggestion for f$verify()1 Message-ID: <jJKkd.2783$4o2.753@news.cpqcorp.net>    JF Mezei wrote: K > Trying to debug a dcl procedure that has a loop that iterates a gazillion - > times before aborting can be quite tedious.  > L > Therefore I would suggest an addition to the SET VERIFY and F$VERIFY which- > would create a second mode of verification:  > W > During execution, nothing is logged, but when an error occurs, 2 lines are displayed: I > The original line in the file (and possibly its line number ?), and the ) > expanded line with symbol substitution.  > O > I just spent a long time trying to debug AUTOGEN.COM to find out why it would M > abort midway (turns out I had a space in modparams.dat at the wrong place).   - Not exactly what you want, but have you tried    $ on error then set verify   --   John Reagan / HP Pascal/{A|I}MACRO for OpenVMS Project Leader  Hewlett-Packard Company    ------------------------------  # Date: Thu, 11 Nov 2004 15:18:09 GMT * From: Paul Anderson <paul.anderson@hp.com>! Subject: Re: DCPS Version Support 5 Message-ID: <111120041018081784%paul.anderson@hp.com>   5 In article <41929845.D1112F32@teksavvy.com>, JF Mezei % <jfmezei.spamnot@teksavvy.com> wrote:    > Paul Anderson wrote: > C > > HP, DCPS is not allowed to say we support those versions.  DCPS D > > does nothing to prevent someone from running on older versions. C > > DCPS is fairly insensitive to operating system version changes,  > B > Generally speaking, would doing an ANA/IMAGE to find the list ofD > system services called give one a clue of whether a new version ofG > DCPS makes use of any new services not available on the current (old)  > version of VMS ?  E I suppose so, although doing so sounds sort of ugly.  And not all new D operating system features used by a product would show up that way. F Take, for example, the longer string used by the queue manager to holdD a printer device name.  DCPS does not use it now, but if we did, its* use would not show up with ANALYZE /IMAGE.   Paul   --    Paul Anderson   OpenVMS Engineering    Hewlett-Packard Company    ------------------------------    Date: 11 Nov 2004 10:16:55 -0800& From: jordan@ccs4vms.com (Rich Jordan)* Subject: Dynamic FTP site and file locking= Message-ID: <cc5619f2.0411111016.5c3cb6f2@posting.google.com>   A We're going to be setting up an FTP site, hopefully on an OpenVMS F Alpha system.  The site will hold data files placed there by dozens ofF sources, updated every 20-30 minutes, via authenticated FTP sessions. > The connection the data sources use is low bandwidth, and eachA transfer will take about 45-60 seconds; access by the downloading D clients (using anonymous FTP) will be much faster.  Older (VMS file)C versions of the data files will be retained, but users coming in to C get the files should normally only retrieve the most 'current' one.   F The system will be running VMS V7.3-2, TCPIP V5.4 ECO2, and _might_ be* running Hunter's FTP server package (TBD).  D From some testing with a network camera that FTPs files to a VMS FTPC server, we've noted that if a request to d/l a file comes in during F the period a data source is creating a new one, the downloading clientA will get an access error.  Given the 45-60 second potential delay 6 before a retry works, that isn't going to work for us.  > One obvious fix is to have the data sources FTP to a differentB directory or directory tree and use a process on the VMS system toD watch for and move the data files to the outbound FTP directory; theC interlock time would be reduced to a second or two per file.  Or we F could move the file into the anonymous FTP directory under a differentC name, then do a rename, which should reduce the interlock time to a  fraction of a second (I think).   ; That will work, but is there a better solution available?      Thanks   Rich CCS    ------------------------------  % Date: Thu, 11 Nov 2004 08:37:20 -0500  From: norm.raphael@metso.comG Subject: Re: Help with OPCOM and OPA0: and console vs. DW/motif display Q Message-ID: <OF567EA413.C3219E68-ON85256F48.0079A5FF-85256F49.004AD613@metso.com>   
 [New Text]  : More detail on the current state (Short uptimes due to UPS reconfiguration):   1 %SYSMAN-I-OUTPUT, command execution on node NODEB  (  (LNM$SYSTEM_TABLE)   "OPC$LOGFILE_ENABLE" = "TRUE"    "OPC$OPA0_ENABLE" = "TRUE"  1 %SYSMAN-I-OUTPUT, command execution on node NODEV   4 %SHOW-S-NOTRAN, no translation for logical name OPC*1 %SYSMAN-I-OUTPUT, command execution on node NODEA    (LNM$SYSTEM_TABLE)   "OPC$LOGFILE_ENABLE" = "TRUE"    "OPC$OPA0_ENABLE" = "TRUE"   SYSMAN> do sho sys/proc=opcom 1 %SYSMAN-I-OUTPUT, command execution on node NODEB J OpenVMS V7.3-1  on node NODEB  11-NOV-2004 08:21:13.10  Uptime  4 22:05:18F   Pid    Process Name    State  Pri      I/O       CPU       Page flts Pages F 2020020D OPCOM           HIB      8    14941   0 00:00:09.25       179 521 %SYSMAN-I-OUTPUT, command execution on node NODEV H OpenVMS V7.2  on node NODEV  11-NOV-2004 08:21:14.29  Uptime  4 21:58:10F   Pid    Process Name    State  Pri      I/O       CPU       Page flts Pages F 2040010E OPCOM           HIB      9    30185   0 00:00:37.67       590 366 1 %SYSMAN-I-OUTPUT, command execution on node NODEA J OpenVMS V7.3-1  on node NODEA  11-NOV-2004 08:21:14.67  Uptime  4 21:51:21F   Pid    Process Name    State  Pri      I/O       CPU       Page flts Pages F 20600210 OPCOM           HIB      8    15652   0 00:00:09.63       181 53  1 %SYSMAN-I-OUTPUT, command execution on node NODEB C Terminal OPA0:, device type unknown, is online, enabled as operator 	 terminal, -     record-oriented device, carriage control. <     Error count                    0    Operations completed 497 1     Owner process                 ""    Owner UIC  [SYSTEM]0     Owner process ID        00000000    Dev Prot S:RWPL,O:RWPL,G,W ;     Reference count                0    Default buffer size  801 %SYSMAN-I-OUTPUT, command execution on node NODEV D Terminal OPA0:, device type LA36, is online, record-oriented device, carriage     control.<     Error count                    0    Operations completed 16186 1     Owner process                 ""    Owner UIC  [SYSTEM]0     Owner process ID        00000000    Dev Prot S:RWPL,O:RWPL,G,W ;     Reference count                0    Default buffer size  132 1 %SYSMAN-I-OUTPUT, command execution on node NODEA C Terminal OPA0:, device type unknown, is online, enabled as operator 	 terminal, -     record-oriented device, carriage control. <     Error count                    0    Operations completed 437 1     Owner process                 ""    Owner UIC  [SYSTEM]0     Owner process ID        00000000    Dev Prot S:RWPL,O:RWPL,G,W ;     Reference count                0    Default buffer size  80  + This executes on the Alphas during startup:   ! $   DEFINE/USER SYS$COMMAND OPA0:  $   REPLY/ENABLE! $   DEFINE/USER SYS$COMMAND OPA0:  $   REPLY/ENABLE=(OPER1,-                OPER2,-                OPER3,-                OPER4,-                OPER5,-                OPER6,-                OPER7,-                OPER8,-                OPER9,-                OPER10,-               OPER11,-               OPER12)    but if I do this with  $   REPLY/ENABLE=(<etc>)1 on the VAX, then the application has the problem.    [End New Text]   Thanks for the response.  > hoff@hp.nospam (Hoff Hoffman) wrote on 11/10/2004 03:56:56 PM:  ? > In article <OFED6B7330.889D871E-ON85256F48.006D3821-85256F48. 5 > 006EFEEA@metso.com>, norm.raphael@metso.com writes:  > : = > :Help with OPCOM and OPA0: and console vs. DW/motif display  > ..? > :I have read the doc on how to make the graphics console into = > :a display console, but without understanding.  What I need 6 > :is to be able to replace the VAX and still have the> > :application believe the messages were written successfully. > K >   Disable OPA0: as an operator per the OpenVMS Frequently Asked Questions G >   (FAQ), and find some other terminal to enable as an operator if the E >   application actually requires it, or create a pseudo-terminal and C >   enable it as an operator if you lack a spare physical terminal.   F I find no match for pseudo-terminal in the FAQ.  Is there something on how to create one?   > > > :Is there a way to keep the DW on the graphics terminal, butB > :direct the messages either to a VT on a TTA port or a blackhole& > :and if so, how do I configure that? > % >   Sure; please see the OpenVMS FAQ.  >   ; I see it in there, but I am uncomfortable with what I read.   D >   I'd almost wonder if the application was causing the stall, whatB >   with an ever-accumulating pile of requests.  For grins, do tryC >   a REPLY/STATUS from an terminal enabled as an operator, and see A >   how many requests have stacked up from this application.  (Is A >   it the lack of the terminal and an output I/O stall, or is it A >   the lack of a response and a backlog of operator requests, in  >   other words.)  > D >   I will assume you are aware that the application appears to have+ >   at least one coding problem, of course.  >   > Yes, indeed.  I made a vain attempt to communicate this to the: maintainers back when they still had some resources in the> OpenVMS space, and they seemed unable to comprehend that fact.8 It is now a lost cause, and I must needs work around it.   > 2 >  ---------------------------- #include <rtfaq.h> ----------------------------- 3 >     For additional, please see the OpenVMS FAQ --  www.hp.com/go/openvms/faq 4 >  --------------------------- pure personal opinion --------------------------- G >         Hoff (Stephen) Hoffman   OpenVMS Engineering   hoff[at]hp.com  >    ------------------------------  % Date: Thu, 11 Nov 2004 10:19:21 +0000 - From: Roy Omond <Roy.Omond@BlueBubble.UK.Com> 6 Subject: Re: how to implement an append only log file?+ Message-ID: <2vgsp3F2m7ociU1@uni-berlin.de>    Tom Linden wrote:    > [...snip...] > J > Came across this other day when googling on watchdog, to emphasize the  7 > point    http://www.wordserf.co.uk/mh/vaxhackpro.html   : Ah, the days of the Cuckoo's Egg, July 1987, and the Chaos Computer Club.  < I was the whistleblower ... it was an "interesting" exercise> in reverse enginnering to see what they had done, and how they had done it.   ------------------------------  + Date: Thu, 11 Nov 2004 10:42:54 +0000 (UTC)  From: david20@alpha2.mdx.ac.uk6 Subject: Re: how to implement an append only log file?) Message-ID: <cmvfne$gfs$1@news.mdx.ac.uk>   ] In article <cmtsnd$q5f$1@reader1.panix.com>, Dale Dellutri <ddelQQQlutr@panQQQix.com> writes: \ >On 10 Nov 2004 12:51:59 -0600, Bob Koehler <koehler@eisner.nospam.encompasserve.org> wrote:` >> In article <cmti40$mfv$1@reader1.panix.com>, Dale Dellutri <ddelQQQlutr@panQQQix.com> writes: >> >  I >> > So let me describe a possible system.  Tell me what's wrong with it:  > # >>    That software already exists.  > E >> > Assume a modification to VMS which sends all audit events from a I >> > "secured machine" to another machine (the "audit machine").  The log G >> > is maintained on the audit machine, so that it can't be modified / H >> > deleted / tampered with by the manager of the secured system.  The 7 >> > audit machine is solely controlled by the auditor.  > H >> > The manager of the secured system could modify the operating systemH >> > (patch it, change memory locations, whatever) to disable or, worse,K >> > spoof the logging.  However, this change would be an audit event which I >> > would show up in the log on the audit machine because it would occur E >> > before the change.  If the audit machine receives such an event,  >> > something is wrong. > F >>    If the manager knew what he was doing, he'd make the change in aI >>    non-auditable way, by first patching the audit software with a tool # >>    that doesn't generate audits.  > F >If you mean patching the software in memory while it's running, then G >at some point you'll have to make a system call to change kernel code  D >memory, and this system call would produce an audit event _before_  >making the change.  > K So you are going to audit everytime any program on the system does a CMKRNL L and then audit everything that program does whilst in Kernel mode (includingN all the running code which makes up the OS itself). You are then going to send9 all this output across a network link to a remote system. N Sounds like an awful lot of logging information in which to hide incriminatingH evidence - the incriminating evidence just being writing to a few memoryH locations (the location of which you won't know unless you know the full" details of every possible attack).    
 David Webb Security team leader CCSS Middlesex University    H >If you mean patching the software in the module file on disk, then, of = >course, changing certain files (or perhaps files in certain  B >directories or on certain disks) would have to be an audit event.' >And so would loading them into memory.  > E >Remember that I assumed that OpenVMS would have to be changed to do  E >this, and the audit events would have to be fine-grained, more than  & >what's now considered an audit event. >  >-- 8 >Dale Dellutri <ddelQQQlutr@panQQQix.com> (lose the Q's)   ------------------------------    Date: 11 Nov 2004 07:42:55 -0600; From: koehler@eisner.nospam.encompasserve.org (Bob Koehler) 6 Subject: Re: how to implement an append only log file?3 Message-ID: <yOeOgl2F$dDN@eisner.encompasserve.org>   ] In article <cmtsnd$q5f$1@reader1.panix.com>, Dale Dellutri <ddelQQQlutr@panQQQix.com> writes:  > G > If you mean patching the software in memory while it's running, then  H > at some point you'll have to make a system call to change kernel code E > memory, and this system call would produce an audit event _before_   > making the change.  D    I'm sure I'm not the only one who has patched the kernel from theF    console while the CPU was halted.  How are you going to audit that?E    I just had to break up my patch entry into small enought pieces to .    keep various sanity timers from triggering.  E    And there are more ideas I've seen floating around.  In the end at 0    some place you have to have people you trust.   ------------------------------  # Date: Thu, 11 Nov 2004 14:06:33 GMT 3 From: hammond@not@peek.ssr.hp.com (Charlie Hammond) 6 Subject: Re: how to implement an append only log file?2 Message-ID: <JlKkd.2781$%a2.1304@news.cpqcorp.net>  - In article <41929331.64F709D6@teksavvy.com>,  / JF Mezei <jfmezei.spamnot@teksavvy.com> writes:    ..7 >In the end, you have to trust your system manager. ...   5 So it is with all security measures, computer or not. & Utimately, you have to trust somebody.: This is fine -- if you cannot truse anybody, then there is little point in doing anything.    --  J       Charlie Hammond -- Hewlett-Packard Company -- Ft Lauderdale  FL  USAF           (hammond@not@peek.ssr.hp.com -- remove "@not" when replying)J       All opinions expressed are my own and not necessarily my employer's.   ------------------------------  % Date: Thu, 11 Nov 2004 12:23:12 -0600 / From: Chris Scheers <chris@applied-synergy.com> + Subject: Re: Need help with a PeeCee again. + Message-ID: <2vhp03F2kv01rU1@uni-berlin.de>    JF Mezei wrote:  > Chris Scheers wrote: > C >>It appears that your CD burning software is using the short names H >>instead of the long names.  It may have an option to specify which set >>of names to use. >  > H > Nop. Joliet extension has a similar encoding thing. It stores both 8.3O > filenames in the standard  ISO 9660 place, and stored the complete file names 3 > in a different place used by the Joliet software.  > P > So when you mount a CD that is Joliet encoded on a non Joliet machine, you seeO > the 8.3 names. Mount the same CD on a Joliet equipped machine and you see the 
 > full names.   H If I understand this correctly, the ISO9660 standard defines 8.3 names, 6 and then an extension is used to support longer names.  ' What extension(s) are supported by VMS?   G ----------------------------------------------------------------------- $ Chris Scheers, Applied Synergy, Inc.  B Voice: 817-237-3360            Internet: chris@applied-synergy.com    Fax: 817-237-3074   ------------------------------  % Date: Thu, 11 Nov 2004 18:41:29 -0000 = From: "Andoni" <no_spam_please@andoni-at-ireland-dot-com.com>  Subject: Oracle DST Setting.2 Message-ID: <snOkd.41628$Z14.16268@news.indigo.ie>   Hello,  I Does anybody use Oracle 9i for OpenVMS? Do you have to change the logical  name:   &   "SYS$TIMEZONE_DIFFERENTIAL" = "3600"  2 twice a year, when the clocks go back or forwards?  L We have found that Oracle does not seem to see the ordinary operating systemJ time but rather is adding this value to real time (winter time) so we have? to put it in place in Spring and set it to zero in Autumn/Fall.    Surely there is a better way?   7 Thanks in advance for any help you may be able to give.    Andoni.    ------------------------------  % Date: Thu, 11 Nov 2004 16:10:13 +0100 ( From: "Rudolf Wingert" <win@fom.fgan.de>$ Subject: Problem to Partition a disk3 Message-ID: <000001c4c800$9d370380$994614ac@wat153>    Hello,  A I would like to patition a 146GB disk with the LDdriver into four C partitions. The first problem (under OpenVMS 7.3-1 AXP) was, that I G could not initialyze this disk with clustersize = 1. With clustersize = A 4 there was no problem. I did also set the /INDEX to BEGIN. After > mounting I created a directory ([LD]) and tried to create fourG contiguous files of the size = 71000000  with the LDdriver. But I could @ create only three of them. The last one did not have enaugh freeE contiguous disk space. The DEFRAG shows, that the disk was fragmented G into a few extends (four). I don't understand why. What's wrong with my   initalyzation und file creation?   TIA and best Regards R. Wingert    ------------------------------  % Date: Thu, 11 Nov 2004 05:46:42 -0500 - From: JF Mezei <jfmezei.spamnot@teksavvy.com> 9 Subject: Real documentation on SYSGEN memory parameters ? , Message-ID: <419342E3.A1AEEB83@teksavvy.com>  L Where could I find some real documentation on all the various memory relatedI sysgen paramenters (for vax at least) and how to calculate them and their  inter-relationship ?  I For instance, if I add to parameter X, does that also require I add to Y  " because X consumes memory from Y ?  L The guide to performance really doesn't cover anything like SPTREQ, NPAGDYN, IRPCNT, SYSMWCNT etc etc etc.   L So it is pretty hard to start to tweak a system without really understanding* the interrelationship of those parameters.   ------------------------------    Date: 11 Nov 2004 08:47:33 -0800 From: raycook@f2s.com (Ray) % Subject: Re: Use of FTP in .Com files = Message-ID: <29bfccac.0411110847.338bc081@posting.google.com>   O Thanks to all who replyed. Its been a great help and the problem is now solved.    Cheers   ------------------------------    Date: 11 Nov 2004 01:12:31 -0800/ From: stuie_norris@yahoo.com.au (Stuart Norris) < Subject: [Q]Shutting down TCPIP Sockets with AST read posted= Message-ID: <51262235.0411110112.75166092@posting.google.com>   
 Dear Readers,   = I have a query about shutting-down a TCPIP socket that has an E outstanding AST read posted.  I am interested in the flow of events I  should see.   F My program accepts a connection via a listener socket and posts an AST< read on the socket which completes via a routine "readast". 
 (readvblk)  A In another part of my program I decide to shutdown  and close the ! socket connected on this channel.   ; I issue a QIOW on the socket channel with an IO$_DEACCESS + = IO$M_SHUTDOWN with TCP$C_DSC_ALL in P4 to shutdown the socket   9 I then issue a QIOW on the socket with a IO$_DEACCESS and ( TCP$C_DSC_ALL in P4 to close the socket.   I then de-assign the channel.    My questions are:   A Firstly should my AST  readast be called due to this and at what  point?  E How in my AST routine can I tell that it has been executed due to the  socket shutting down.   E Is it possible using my AST read routine to tell that my program shut   down the socket not the client ?  A If I disable ASTs before shutting down the socket will change the  above answers.  8 Are there any examples / diagrams that explain the flow?   Thanks   Stuart   ------------------------------   End of INFO-VAX 2004.627 ************************