1 INFO-VAX	Thu, 12 Jun 2003	Volume 2003 : Issue 324       Contents:& Re: CSWB 1.2.1 fails to display images& Re: CSWB 1.2.1 fails to display images" Re: Dell to make Itanium 2 systems Re: Firewall for VMS / TRU64
 Forcing ip Re: Forcing ip Re: Forcing ipD GNU-izing a Unix System V IPC, signal and socket system for OpenVMS? Re: Help listing SYSGEN params( Re: Help with 3100/98(was help with VMS)( Re: Help with 3100/98(was help with VMS)# Re: Here's a commercial for you all # Re: Here's a commercial for you all " Re: HP #1 in Servers and Notebooks- Re: Knowledge of VMS tests for job candidates - Re: Knowledge of VMS tests for job candidates - Re: Knowledge of VMS tests for job candidates - Re: Knowledge of VMS tests for job candidates - Re: Knowledge of VMS tests for job candidates - Re: Knowledge of VMS tests for job candidates ) Re: Motif 1.3, Euro sign not standard ?!? ) Re: Motif 1.3, Euro sign not standard ?!? ) Re: Motif 1.3, Euro sign not standard ?!? ) Re: Motif 1.3, Euro sign not standard ?!? ) Re: Motif 1.3, Euro sign not standard ?!? ) Re: Motif 1.3, Euro sign not standard ?!? ) Re: Motif 1.3, Euro sign not standard ?!?  Re: perl question  Re: perl question  Re: perl question # Re: PIPE output defined to a symbol # Re: PIPE output defined to a symbol # Re: PIPE output defined to a symbol  PLUG: txt2pdf 6.6  Re: Portents of Itanium death  Re: Portents of Itanium death  Re: Portents of Itanium death  Re: Portents of Itanium death  Re: Portents of Itanium death  Re: Portents of Itanium death  Re: Portents of Itanium death  Re: Portents of Itanium death  Re: Portents of Itanium death  Re: Portents of Itanium death ? Re: Printer recommendation, hobbyist/decwindows/mozilla and VMS 0 Problems running a .com file using a Java socket# Re: Rookie question for the experts # Re: Rookie question for the experts # Re: Rookie question for the experts # Re: Rookie question for the experts # RE: Rookie question for the experts # Re: Rookie question for the experts # Re: Rookie question for the experts  Re: TK70 tape drive problem  Re: TK70 tape drive problem % Re: TK70 tape drive problem - Updatye % Re: TK70 tape drive problem - Updatye % Re: TK70 tape drive problem - Updatye  Tracking VMS SMTP mail Re: Tracking VMS SMTP mail Re: uptime-alike Re: uptime-alike' Re: Wininet.dll (FindFirstFile) and VMS ' Re: Wininet.dll (FindFirstFile) and VMS ' Re: Wininet.dll (FindFirstFile) and VMS 5 [OT] SCO may revoke IBM's license to SysV tomorrow...   F ----------------------------------------------------------------------    Date: 11 Jun 2003 23:28:58 -0700% From: Bart.Zorn@xs4all.nl (Bart Zorn) / Subject: Re: CSWB 1.2.1 fails to display images = Message-ID: <a98cd882.0306112228.26dbba47@posting.google.com>   [ Colin Blake <colin@theblakes.com> wrote in message news:<3EE785CE.5090107@theblakes.com>...  > Bart Zorn wrote: > G > >Just after I posted my previous message (it has not even appeared on @ > >Google yet), I found what the problem was. At system startup,H > >MOZILLA$STARTUP.COM defines SYSTEM logical names for it's components. > > 4 > MOZILLA$STARTUP.COM?  Where did this come from????  F I traced my steps back. I wrote that my self to be able to use mozillaF in stead of netscape from BNU, the Book Navigation Utility. ApparentlyF it is not possible to have BNU invoke a commandprocedure when it wants to fire up the html browser.A So I wrote MOZILLA$STARTUP.COM to define all the logical names as E /SYSTEM, so that BNU could just run the mozilla image in stead of the  netscape image. B Of course this was a few months ago, so when the newer versions of, mozilla came up, I had forgotten about this.   Regards,  	 Bart Zorn    ------------------------------  # Date: Thu, 12 Jun 2003 11:26:03 GMT ' From: Colin Blake <colin@theblakes.com> / Subject: Re: CSWB 1.2.1 fails to display images * Message-ID: <3EE86342.30408@theblakes.com>   VAXman- wrote:  2 >I believe the URL would be http://www.hp.com/DSPP > E >You cannot register.  Walk through the registration process and you   >will see the problem(s).  > H I did just try it, using SWB 1.2.1. I was able to successfully register  on my first attempt.  F >I am not supplying my email address to the bugzilla site.  Therefore, >I will not enter a bug report.  >  That is your choice.   ------------------------------  + Date: Thu, 12 Jun 2003 11:05:42 +0000 (UTC) + From: david20@alpha2.mdx.ac.uk (David Webb) + Subject: Re: Dell to make Itanium 2 systems + Message-ID: <bc9mq6$nif$1@aquila.mdx.ac.uk>   N In article <3EE784C7.3776FAE6@istop.com>, JF Mezei <jfmezei@istop.com> writes: >David Harrold wrote: P >> I didn't say and didn't mean to imply that Dell didn't build "solid" servers.F >> I tried to say I doubt someone would by a Dell for their productionM >> environment because VMS on a Dell server is not a supported configuration.  > L >Right now, it isn't supported. This will be the real test to see whether HPK >will officially support VMS on other people's hardware. Or perhaps HP will  >expect DELL to support VMS. > L I was at the OpenVMS on Itanium Technical update in London earler this week.J We were asked whether we needed a workstation system running VMS - all theN slides in the presentations showed rack mounted systems even at the lower end.O It was good that we were asked but a bit suprising that it was necessary to ask G after all the past discussions with Compaq on the need for low end VMS   workstation systems.  
 David Webb VMS and Unix team leader CCSS Middlesex University        K >Hey, I just got this wild idea: imagine if we could convince Mikie Dell to I >sell VMS ? Imagine how popular VMS would become if Dell were to start to O >market VMS servers, clusters etc , all of the low end stuff that HP refuses to N >market ? (HP can retain the couple sales of wildfire-class machines each yearI >while Dell would be the growing number of low end and midrange systems).  > J >And with the majority of new VMS customers running on DELL, VMS engineers: >would be forced to ensure that VMS runs on DELL machines. >  >DUDE !, you're getting VMS !    ------------------------------  + Date: Thu, 12 Jun 2003 10:40:47 +0000 (UTC) + From: david20@alpha2.mdx.ac.uk (David Webb) % Subject: Re: Firewall for VMS / TRU64 + Message-ID: <bc9lbe$ngf$1@aquila.mdx.ac.uk>   N In article <3EE76615.D0B96283@istop.com>, JF Mezei <jfmezei@istop.com> writes: >David Webb wrote:P >> If you are providing a service to the rest of the world then the traffic just( >> passes straight through the firewall. >  > I >But with a firewall, you can provide many local services such as telnet, G >LPR/LPD etc whose intrinsic security isn't very high with the firewall N >blocking access to those services, at the same time as providing HTTP service >to the world. > K >However, in fairness, the TCPIP Services stack does allow you to define IP H >adresses to accept/block for each service. However, that feature is notK >applicable to applicatiosn that bind to a port to listen for calls without ( >being defined in the services database. > L Yes DEC TCPIP services is rather limited in it's access control. However theG same cannot be said for TCPWARE which provides better access control on H services and also provides packet filtering. The packet filtering can beL applied to all open ports (just as on a router) irrespective of whether they% are started by the services database.   L Hence since Bob was talking about VMS and TCPWARE then I'd say he has a fairO point about a firewall in front of such a secured VMS system not providing much  extra security. N Note this does not support his initial claim that VMS + TCPWARE is a firewall.O There is a vast difference between the facilities needed to protect connections K to your own system and screening connections through your system to lots of  other systems.  G However thats not to say a firewall doesn't provide any extra security. N As I stated a firewall in front of the system is an extra layer of protection.L The fact it is a different system protects you from any possibility (howeverJ unlikely) that their might be holes in the stack itself which would allow H connections to somehow bypass the packet filtering. It allows for easierL management of security rules by putting them in one place when you have lotsK of systems - though you should also if possible also harden all the systems M behind the firewall anyway to guard against penetration through the firewall. N It's multiple interfaces and physical positioning may also allow it to detect H certain types of ip spoofing - ie packets claiming to come from internal= systems entering the firewall through the external interface.       M >A firewall with NAT also has an advantage: by default, all incoming calls go N >either nowhere or to one default host. And you then "enable" connections on aM >service/port by port basis. So you don't have to worry about Satan trying to K >connect to port 666 of every host in your lan, or some porn star trying to  >download porn to port 69. >   N The NAT actually provides pretty much nothing on security - you can easily getE the same affect with a default deny policy on a firewall without NAT. N NAT's purpose is to solve an addressing problem. It's security usage is a sideM affect and is really just security by obscurity (which any security book will  tell you is no security). H NAT is like painting your house black (bricks, doors, windows) so that aJ burglar at night can't tell whether your house is a small cottage or a big mansion.N Any hacker scanning your network actually has an easier job. Instead of havingO to scan every machine on a class C or bigger network they just have to scan one L address and suddenly they see this (at first sight) really tempting target aO single machine with lots of open ports. Subsequent in depth probing will reveal  the deception.  N >But perhaps a greater advantage of a firewall is the denial of service issue.G >Ping of deaths mayh disable your router, but the rest of your intranet K >continues to work. Without a firewall, the hackers will touch your systems $ >directly and can slow/disable them. > L Bob explicitly excluded denial of service attacks from his discussion at theI start. He conceded that firewalls can help protect from denial of service  attacks.      J >And let's talk about buffer overflow. Consider that much of the VMS TCPIPO >Services stack comes from Unix. Consider that the POP server is installed with H >SYSPRV and one could wipe out any file on the system by starting it andO >specifiying any file to be created as a log file. Seems to me that this is the J >same type of quality control that results in unix or windows applications >being vulnerable. > O >An application that runs with sysprv enabled is just as dangerous on VMS as it  >it on windows.  > N >Compare those with the OSU web server whose security was designed with VMS inL >mind and where its privileges are disabled once port 80 has been bound, andG >the rest can run in unprivileged mode, with little risk to the system.  >   G As I said the presence or absense of a firewall has no bearing on this. L The pop server issue was really bad (compared with the usual sort of quality7 control we have come to expect from the VMS engineers). G Of course a badly written (or badly ported) piece of privileged code is 
 dangerous.  N >How does Compaq's proprietary version of Apache for VMS handle its privilegesO >? Does it do the same thing as OSU ? Or does it run with sysprv enabled at all  >times ?  O Sorry I've no idea since I've never used it on VMS. I hope after the pop server O issue a higher level of checking for possible security holes is being done with K ported code. I also trust that the work the VMS engineers are doing to make O porting Unix code to VMS (hard links, soft links and other file system changes, I guid and uid support, fork etc ) will be added in a manner which does not N compromise VMS security - as some of these features have in the past sometimes compromised Unix security.      
 David Webb VMS and Unix team leader CCSS Middlesex University   ------------------------------    Date: 12 Jun 2003 08:39:14 -0700$ From: mmabey@cenhud.com (Mike Mabey) Subject: Forcing ip = Message-ID: <c13d5929.0306120739.481c282f@posting.google.com>   E I have a program (Compaq Fortran V7.5) on my Alpha cluster that sends F data to another program on a VAX (not owned by my company, not part of( the cluster) by opening a file this way:  9 OPEN(1,NAME='VAXNODE"user password"::"TASK=LISTENER"' ...   E My program then performs periodic writes to the file and the LISTENER $ program reads the data from SYS$NET.A My question is this: Is there any way to force this all to happen @ using only tcpip? (not DECNET over ip) We are putting a firewallC between us and them and the decision was made not to spring for the . more expensive firewall that was DECNET aware. Any help appreciated.   
 Mike Mabey Central Hudson Gas & Electric    ------------------------------    Date: 12 Jun 2003 11:59:39 -0500; From: koehler@eisner.nospam.encompasserve.org (Bob Koehler)  Subject: Re: Forcing ip 3 Message-ID: <FzyY0KQfQrrv@eisner.encompasserve.org>   d In article <c13d5929.0306120739.481c282f@posting.google.com>, mmabey@cenhud.com (Mike Mabey) writes:G > I have a program (Compaq Fortran V7.5) on my Alpha cluster that sends H > data to another program on a VAX (not owned by my company, not part of* > the cluster) by opening a file this way: > ; > OPEN(1,NAME='VAXNODE"user password"::"TASK=LISTENER"' ...  > G > My program then performs periodic writes to the file and the LISTENER & > program reads the data from SYS$NET.C > My question is this: Is there any way to force this all to happen ) > using only tcpip? (not DECNET over ip)    B    Yes, you could completely rewrite the application, using TCP/IPE    sockets and supplying a custom server for the remote end that the  ?    other company owns, to replace LISTENER, which is probably a     simple DCL .COM file.  G    OBTW, unless you write it into the application and server, you won't D    need to be bothered with silly little security details like usersA    and passwords.  Of course, you may hear from the other company C    when some hacker discovers this interesting open port and stuffs     random data in it.   C    Or you could run DECnet over IP, which the routers and firewalls B    will see as ordinary IP traffic (port 700?).  So why do you say    "not DECNET over ip"?   ------------------------------  % Date: Thu, 12 Jun 2003 13:04:09 -0400 " From: JF Mezei <jfmezei@istop.com> Subject: Re: Forcing ip ) Message-ID: <3EE8B287.AA2CB908@istop.com>    Mike Mabey wrote: ; > OPEN(1,NAME='VAXNODE"user password"::"TASK=LISTENER"' ... C > My question is this: Is there any way to force this all to happen  > using only tcpip?   . Yes, but you'll have to rewrite both programs.  M Instead of the simple OPEN, you'll have to use eithet the "sockets" routines, L or the SYS$QIO interface to connect to the IP device, bind to a port, defineJ the TCP charactertistics, do the DNS request to convert the host name to a7 binary IP address, and then connect to the remote host.   M The remote host will have to do something similar except it binds to the port N it wants to listen to and then waits for your call. The big difference is thatK you won't have the user password authentication into the call request. (eg:  anyone can connect).  L Using decnet-over-IP would solve the router issue and not require changes to your programs.   ------------------------------    Date: 12 Jun 2003 09:54:08 -0700$ From: rcbryan@hotmail.com (RC Bryan)M Subject: GNU-izing a Unix System V IPC, signal and socket system for OpenVMS? = Message-ID: <fbcf38dc.0306120854.1d928349@posting.google.com>   B Is there any interest in GNU-izing a Unix System V IPC, signal and socket system for OpenVMS?   D Years ago, in porting a massive Unix based application (>3.5 millionC lines of code) to VMS, my company developed an integrated system of E System V IPC, signal and socket routines.  The reason the development C occurred is because our software depends heavily upon having SIGIOs B interrupt the msgrcv service.  Since the basic UCX socket routinesF don't issue a SIGIO on I/O completion, we were kind of stuck, but that was a long time ago.   E The actual development of these was before my time with the company.  C My personal feeling on developments like this is "you don't have to C write the operating system, the OS vendor has already taken care of E that for you."  While it is possible that there could have been a way D to make things work without writing a Unix subsystem, it is written,E it works and it is solid.  It is in use all over the world and I have 6 had to fix about two bugs in the last 18 months or so.    What is there:D 1. Full System V IPCs.  Every last bit in every service is accountedC for.  I was blown away when I first saw how complete this is.  This B comes with ipcs and ipcrm routines.  Semaphores and other internalB data structures are maintained in memory that can only be accessed? only in kernel mode so they are safe from misbehaving programs. E 2. TCP/IP Sockets.  This is not as complete as the IPCs.  We are kind B of subject to the underlying UCX stack.  This is a layer on top of $QIOs, parallel to UCX. @ 3. POSIX signal handling.  This is pretty good.  The signals areF modeled as ASTs.  It was news to me that one process could schedule anF AST for another similar to the Unix kill service (via SCH$QAST) but itD is in there.  The availability of alarm, ualarm and kill is a given.B 4. Integration between the IPCs and the sockets.  We evaluated allC kinds of products and nothing provided the ability to have an EINTR B return from a msgrcv; ie, breaking out of a IPC queue message read when network I/O is received. B 5. Process management.  It is kind of hard to quickly describe butD because we have to do interprocess ASTs, all processes must be known< to the subsystem.  It is possible to spawn independent childC processes, not subject to the quotas of the parent.  All background F processes can be run under the debugger and can use the "system" C-RTLC routine.  A facility to provide selective privileges and quotas for E background processes is included.  There is not a fork but there is a % functional replacement for fork/exec. 4 6. Child processes inherit process table logicals.  F 7. Test routines that can be used to test/demonstrate about 95% of theE functionality. (Not 100% because the functionality was extended after @ the initial development and the test routines were not amended.)E 8. A set of disk I/O routines that when the I/O is complete, the data D is on disk -- 100% guaranteed.  This is one of those things where weD tried all the RMS options from the C-RTL and nothing would work 100%C of the time.  Our QIO routines bypass RMS and they work every time. C 9. Active socket passing.  A construct that is common on Unix is to A open a TCP/IP socket from a client and pass it to a child process F after some initialization data has been exchanged.  We needed this and it is part of the package.    What is missing:E 1. UDP/IP. This is not used by our code so no effort was put into it. E 2. Threading.  This is one reason I want to make this open source.  I D have some ideas of how to extend the code for full threading support( but I don't have time to implement them.F 3. TCP/IP streams.  The new version of our product uses streams.  This) is another reason to make it open source.     Where we stand: E The code is currently has the bad old style closed source proprietary F restrictions but we don't make money on it as a separate product and I< don't see how it is going to do anything but help us and ourB customers, to make it generally available.   I think that if I canB convince management that there is interest on the outside and thatD people will be fixing bugs or adding enhancements, they will tell meB "take it away!"  Realistically, since I am the only one that knowsD this code inside and out these days (there are other people but I amF the only one for almost two years), this will mean more work for me inC the short term, answering questions and fixing bugs but in the long ? term, we will get a better product and the community will get a  valuable resource.    Comments anyone?    Regards,	 /RC Bryan    ------------------------------   Date: 12 Jun 03 09:17:45 +0200) From: p_sture@elias.decus.ch (Paul Sture) ' Subject: Re: Help listing SYSGEN params ) Message-ID: <e1qXKlRJ39SA@elias.decus.ch>   T In article <xSPd$jFfjYC+@eisner.encompasserve.org>, briggs@encompasserve.org writes:q > In article <bc22e6$ctj$1$8300dec7@news.demon.co.uk>, "Phil Sharpe" <fed_up_with_spam@nonexistant.co.uk> writes: B >> Anyone know how to get a complete listing of SYSGEN parameters.J >> I've tried copying the contents of my scrollback buffer after SHOW/ALL,A >> but there are gaps (presumably because of the screen heading).  > 9 > Annoying, isn't it?  Fancy behavior for no good reason.  >  > The simple solution is to: >  > $ define sys$output foo.bar  > $ mcr sysgen > SYSGEN>  SHOW /ALL > SYSGEN>  SHOW /SPECIAL > SYSGEN>   Exit   > $ type foo.bar  C > %TYPE-W-OPENIN, error opening DISK1420:[VAXS09]FOO.BAR;2 as input 3 > -RMS-E-FLK, file currently locked by another user ! > $ deassign sys$output	! *blush*  > $ type foo.bar >    SYSGEN>  HELP SET /OUTPUT    SET   	   /OUTPUT   C        Establishes a file to be used for output during the session. F        By default the output is written to SYS$OUTPUT, but you can use+        SET/OUTPUT to designate a disk file.   H        At any time you can direct the output back to SYS$OUTPUT by using)        the SET/OUTPUT=SYS$OUTPUT command.   
        Format   !          SET/OUTPUT[=]  file-spec     %     Additional information available:        Parameter  Example     --  
 Paul Sture   ------------------------------    Date: 12 Jun 2003 06:09:29 -0700) From: machinegunkelley@hotmail.com (Mark) 1 Subject: Re: Help with 3100/98(was help with VMS) = Message-ID: <b3d4f0be.0306120509.11f4aa72@posting.google.com>    Thank you for your replys!B Since the console reports the unit as a KA59-A is this a guarantee that the CPU is a 98        Mark    ------------------------------  % Date: Thu, 12 Jun 2003 15:38:06 +0100 + From: John Laird <john@laird-towers.org.uk> 1 Subject: Re: Help with 3100/98(was help with VMS) 8 Message-ID: <jn3hevc24h9s543ktcoq8aiu77oqcd4l3t@4ax.com>  I On 12 Jun 2003 06:09:29 -0700, machinegunkelley@hotmail.com (Mark) wrote:    >Thank you for your replys! C >Since the console reports the unit as a KA59-A is this a guarantee  >that the CPU is a 98   F Well, it is either telling the truth because it has probed the cpu, orH telling you merely what is hardcoded into its own eprom (I would not betI against the possibility that each motherboard has its own dedicated rom).   J I checked my own 6.2 system files - they do indeed only contain entries upI to the /96.  The fact that your system reports a /95 is a little odd, but I one can only guess what happens when the probed cpu/mb/rom returns values J "out of range".   There are entries for "Unknown Microvax 3100" however...  G What does the lexical function f$getsyi return for items XCPU and SID ? K There used to be some net-available tables of valid values for these, which @ indicate, amongst other things, processor series and generation.     	John    ------------------------------  % Date: Thu, 12 Jun 2003 11:00:14 -0400 < From: "Peter Weaver" <WeaverConsultingServices@sympatico.ca>, Subject: Re: Here's a commercial for you all5 Message-ID: <bca4id$hbc72$1@ID-141708.news.dfncis.de>    David J. Dachtera wrote: >...: > Experience shows that in matters VMS, hp rarely listens. > > > In deferance to Warren Sander, however, yes - at times there	 *IS* some  > positive response. > ; > Now, if only Carly and her ilk would take an example from 
 the Warren > Sanderses of the world...     5 Don't forget that we recently saw this announced too; 6         For VMS on Itanium [and OpenVMS Alpha V7.3-2], bit-to-text ;         translation (what ANALYZE/ERROR used to do) will be  provided,         by the operating system folks again.  9 Two examples of HP listening recently. The start of a new  trend?   -- Peter Weaver Weaver Consulting Services Inc. ) Serving Southern Ontario/Western New York    ------------------------------  # Date: Thu, 12 Jun 2003 16:18:22 GMT # From: "John Smith" <a@nonymous.com> , Subject: Re: Here's a commercial for you allG Message-ID: <iJ1Ga.65599$G_.33843@news02.bloor.is.net.cable.rogers.com>   ? "Peter Weaver" <WeaverConsultingServices@sympatico.ca> wrote in 7 message news:bca4id$hbc72$1@ID-141708.news.dfncis.de...  > David J. Dachtera wrote: > >...< > > Experience shows that in matters VMS, hp rarely listens. > > @ > > In deferance to Warren Sander, however, yes - at times there > *IS* some  > > positive response. > > = > > Now, if only Carly and her ilk would take an example from  > the Warren > > Sanderses of the world...  >  > 7 > Don't forget that we recently saw this announced too; 8 >         For VMS on Itanium [and OpenVMS Alpha V7.3-2],
 > bit-to-text = >         translation (what ANALYZE/ERROR used to do) will be 
 > provided. >         by the operating system folks again. > ; > Two examples of HP listening recently. The start of a new  > trend?  C No trend in listening there. Just OVMS Engineering giving what they  know is needed.   ? If it needed 'listening' to, the request would have had to been D directed to the hp Deparment of Listening, to be queued into the 'To> Be Opened' queue (with non-guaranteed delivery), which is thenD directed to the person in charge of opening up the to be listened toA envelopes (providing that person wasn't taking a sick day for the % paper cut they got the previous day).   D Once the envelope was opened it would have to be stamped with the hpB Department of Listening received stamp and dated and then filed inB random order in the "To be Listened To" queue (with non-guaranteed
 delivery).  > At random intervals ranging from 1 day to several years the hpA Department of Listening would invite the members of the Listening E Committee to review one of the items in the "To be Listened To" queue E and decide whether to place the listening request back into the To Be F Listened To queue for futher listening at a later date, or to send theB listening request on to the appropriate hp Department  for further queueing in other queues.    ------------------------------    Date: 12 Jun 2003 09:05:00 -07001 From: keithparris_NOSPAM@yahoo.com (Keith Parris) + Subject: Re: HP #1 in Servers and Notebooks = Message-ID: <cf15391e.0306120804.62ec355c@posting.google.com>    Andrew Harrison SUNUK Consultancy <Andrew_No.Harrison_No@nospamn.sun.com> wrote in message news:<bc23nf$t25$1@new-usenet.uk.sun.com>... ? > According to Gartner IBM is the largest Server vendor not HP.   E Gartner's and IDC's figures vary, but both agree: Times have changed; @ it's a two-horse race between HP and IBM now; Sun has fallen way4 behind.  And the trend is toward HP taking the lead.  @ > Also HP's numbers have declined faster than Gartner originally
 > thought. > F > Seems that the origional HP numbers were estimates (HP publish their@ > quarterly results a month later than Sun IBM etc). Gartner ranA > with an initial figure that was close to the IDC one, HP didn't A > correct them, now that Gartner have taken HP's actual quarterly ( > numbers the story is rather different.  D No, the underlying problem was deeper than time skew.  IDC's numbersE have consistently tracked more closely to the published data HP, IBM, C and Sun released as part of their earnings statements.  Gartner has F admitted their model was broken and needed work, and made changes (andE restatements) recently.  Time is required to verify if corrections to D their model will now result in accurate numbers (at least for Q1'03,F it appears the model still needs a bit of work, when you compare theirC results with the public data from each vendor).  Historically, they > had tried to produce Unix market share numbers without workingC directly with HP, but now, Gartner has now started working directly @ with HP (and presumably IBM and Sun) to try to get more-accurate> results.  IDC had been working directly with HP all along (andB undoubtedly all the other vendors as well) to help ensure accurate results.    Gartner released this statement:  F "We'd like to bring your attention to some adjustments we have made to? the U.S. historical data. These adjustments were made to better E reflect the relative business performance of a small number of server F vendors and more closely align with their announced financial results.B The period we have reviewed goes back to Q1 2001. This does affectC relative market shares in some areas (most notably RISC servers) as B well as growth rates.  If you have any specific questions on these? changes, please feel free to contact us for further background.   F At Gartner, we continue to strive to provide you with the best quality< market data that is available. We review our data collectionA methodology regularly in order to ensure this. Very occasionally, A circumstances lead us to review historical data points which have ? proven erroneous. As our methodology is refined and matured the C necessity for such corrections will greatly diminish. Please accept 4 our sincere apologies for any inconvenience caused."   ------------------------------  # Date: Thu, 12 Jun 2003 11:23:26 GMT - From: "labadie" <tonari_no_tottoro@127.0.0.1> 6 Subject: Re: Knowledge of VMS tests for job candidates2 Message-ID: <OoZFa.2426$YE2.1508@news.cpqcorp.net>  ? "Cheryl Hoefelmeyer" <texascheryl@hotmail.com> wrote in message 7 news:7112dd15.0306111645.7db403ad@posting.google.com... F > Does anyone out there have any exams to administer to job candidatesH > to see if they do, if fact, know anything at all about VMS, and if so,H > are you willing to send me a copy? We are interviewing candidates, andH > we really need to weed out the people who are not truthful about their< > knowledge before any actual interview. Please send text to! > cheryl.hoefelmeyer@premera.com.  > 	 > Thanks!  > CherylK At Decus Europe (May 2002) , in Lyon, It was possible to pass the following  Vms certifications  G -          Openvms V7 Advanced Administration, Performance and Support.   ) -          Openvms System Administration.     Regards       Grard   ------------------------------  % Date: Thu, 12 Jun 2003 10:48:02 -0400 < From: "Peter Weaver" <WeaverConsultingServices@sympatico.ca>6 Subject: Re: Knowledge of VMS tests for job candidates5 Message-ID: <bca3r4$gs3gl$1@ID-141708.news.dfncis.de>    Stanley F. Quayle wrote:4 > On 11 Jun 2003 at 17:45, Cheryl Hoefelmeyer wrote:& >> We are interviewing candidates, and= >> we really need to weed out the people who are not truthful  about / >> their knowledge before any actual interview.  > = > Easy enough -- require that they are a HP Certified OpenVMS 	 v7 System # > Administrator or System Engineer.  >...  8 Not really a good indication. I wrote and passed the VMS8 Network test when it was in beta mode. It was focused on; DECNet Phase V and UCX. I should not have passed! If it was < DECNet Phase IV and TCPWare then I would expect to pass, but= it had questions on where specific UCX information is stored; ; I had no idea so I guessed. The fact that I passed the test 6 was enough to convince me that the test (and the whole; certification by multi-choice test idea) was useless, other < than the fact I got the"OpenVMS Systems Engineer" pen out of it.    -- Peter Weaver Weaver Consulting Services Inc. ) Serving Southern Ontario/Western New York    ------------------------------  % Date: Thu, 12 Jun 2003 12:18:35 -0400 * From: "Syltrem" <syltremzulu@videotron.ca>6 Subject: Re: Knowledge of VMS tests for job candidates4 Message-ID: <dN1Ga.9851$945.28733@tor-nn1.netcom.ca>   I used this before. E NOTE: It's in french, but I converted it with SYSTRAN, which does not K exactly make it an english document... But that should be enough for you to F understand (with the few modifications I had to do after the automatic conversion).J This is very simple and does not actually cover much, so that someone doesJ not typically need more than 10 minutes to complete the test. If they failF this, they certainly don't know enough to do any more technical stuff.  J And I hope no one will come and say my answers (imbedded) are wrong... :-)     Briefly answer the questions.B  G E.g.: Q: How to erase an entry in a queue: R: I take order DELETE/ENTRY       = 1. To know which partition a machine in cluster for the boot,    how do you make use?   $ SHOW LOGICAL sys$sysroot       2. Name 10 lexical functions  ' you should know the answer to this one!a      = 3. People complain that the access has a file RMS is slow fora  @ some time. Which tools (orders) are at your disposal to find the   problem?   $ MONITOR DISK   $ ANALYZE /RMS   $ CONVERT /FDL   $ EDIT /FDL/ANALYZE =r   etc.      = 4. You want to make so that an account user is closed after 3g  : attempts to enter an incorrect password. What do you make?  > Modify in SYSGEN the values of parameters LGI_BRK__DISUSER and   LGI_BRK_LIMg      F 5. Which are, in the order, the logical tables of names crossed by VMS    to find equivalence (the value)?   lnm$jobe  	 lnm$groupE  
 lnm$system   lnm$syscluster  ; (of course you may have DECW and clusterwide logicals now.)n      B 6. How definite you logical names in APP_SETUP.COM, if you have an  B application installed on $1$dua20:[app... ] and that you want that  ; the logical names point with your files and programs in thee  G sub-directories? Take into account that the application must be able toL  > be moved on another disc or repertory without having to modify   APP_SETUP.COM.  ( Structure: $1$dua20: [ APP]APP_SETUP.COM       $1$dua20: [APP.PROG]*.EXE        $1$dua20: [ APP.DATA]*.DAT   Give all the commands whicht  B you would put in APP_SETUP.COM, to define of the logical names for   each repertory above.t  # $ PROC = f$environment("procedure")n  L $ ROOT =f$parse(proc,, "DEVICE") + f$parse(proc,, "DIRECTORY") - "]" + ". ]"  . $ DEFINE APP_ROOT 'ROOT /TRANSLATION=(CONCEAL)  " $ DEFINE APP_DATA APP_ROOT:[DATA ]  " $ DEFINE APP_PROG APP_ROOT:[PROG ]              & 7. Which privileges are necessary for:  % a) To make a SUBMIT /USER=XYZ? CMKRNLo  D b) SET PROCESS /PRIORITY=5 (if the process doesn't belong to you and  + is not in your UIC group)? ALTPRI and WORLDK      @ 8. Which DCL command do you use for concatenate 2 indexed files?   CONVERT /MERGE       9. What is a generic queue?S  D A queue which sends work to a predetermined list of execution queues      < 10 Which is the order which makes it possible to carry out a  D program in detached mode, with the support for DCL (i.e. environment  , DCL is available for this detached process).  0 $ RUN /DETACH sys$system:LOGINOUT /INPUT=XXX.COM         -- -   OpenVMS 7.2-1, Oracle 8.1.6.08   SyltremmI http://pages.infinit.net/syltrem (OpenVMS related web site - en franais)y8 To reply to myself directly, remove zulu from my address  I "Cheryl Hoefelmeyer" <texascheryl@hotmail.com> a crit dans le message deE7 news:7112dd15.0306111645.7db403ad@posting.google.com... F > Does anyone out there have any exams to administer to job candidatesH > to see if they do, if fact, know anything at all about VMS, and if so,H > are you willing to send me a copy? We are interviewing candidates, andH > we really need to weed out the people who are not truthful about their< > knowledge before any actual interview. Please send text to! > cheryl.hoefelmeyer@premera.com.i >i	 > Thanks!  > Cheryl   ------------------------------    Date: 12 Jun 2003 11:48:29 -0500; From: koehler@eisner.nospam.encompasserve.org (Bob Koehler)C6 Subject: Re: Knowledge of VMS tests for job candidates3 Message-ID: <2AA1m6$I8p7K@eisner.encompasserve.org>   a In article <dN1Ga.9851$945.28733@tor-nn1.netcom.ca>, "Syltrem" <syltremzulu@videotron.ca> writes:uH > 5. Which are, in the order, the logical tables of names crossed by VMS > " > to find equivalence (the value)? > 	 > lnm$job  >  > lnm$groupk >  > lnm$system >  > lnm$syscluster > = > (of course you may have DECW and clusterwide logicals now.)b      You sure you passed?:   ------------------------------  % Date: Thu, 12 Jun 2003 12:57:25 -0400i* From: "Syltrem" <syltremzulu@videotron.ca>6 Subject: Re: Knowledge of VMS tests for job candidates4 Message-ID: <hl2Ga.9853$945.28725@tor-nn1.netcom.ca>  G "Bob Koehler" <koehler@eisner.nospam.encompasserve.org> a crit dans le 8 message de news:2AA1m6$I8p7K@eisner.encompasserve.org...@ > In article <dN1Ga.9851$945.28733@tor-nn1.netcom.ca>, "Syltrem"" <syltremzulu@videotron.ca> writes:J > > 5. Which are, in the order, the logical tables of names crossed by VMS > > $ > > to find equivalence (the value)? > >$ > > lnm$joba > > 
 > > lnm$group< > >4 > > lnm$system > >> > > lnm$syscluster > >.? > > (of course you may have DECW and clusterwide logicals now.)m >  >    You sure you passed?u >   ? Hummmm.. Cut-n-paste problem... So it's Windows fault again ;-)    -- i   OpenVMS 7.3-1, Oracle 8.1.6.0u   Syltrem I http://pages.infinit.net/syltrem (OpenVMS related web site - en franais)p8 To reply to myself directly, remove zulu from my address   ------------------------------  % Date: Thu, 12 Jun 2003 12:57:35 -0400e" From: JF Mezei <jfmezei@istop.com>6 Subject: Re: Knowledge of VMS tests for job candidates) Message-ID: <3EE8B0FE.DF07B389@istop.com>n  K If one starts to narrow the field to only those with a piece of paper, thenhN the issue of finding qualified VMS people will become even harder (if there isI such an issue, from my point of view, there is a surplus of VMS qualifiedu% people, and a shortage of VMS sites).e   ------------------------------  + Date: Thu, 12 Jun 2003 10:19:00 +0200 (MET)d9 From: Phillip Helbig <HELBPHI@sysdev.deutsche-boerse.com>i2 Subject: Re: Motif 1.3, Euro sign not standard ?!?; Message-ID: <01KX052QGPCYAM67GX@sysdev.deutsche-boerse.com>t  J > The <compose> <x> <o> is because before the Euro sign was invented, the= reI > was a 'universal currency symbol' at that character position.  It lookseH > like a circle with four rays at the diagonals coming from it.  Thus itF > looks somewhat like a superimposed x and o.  Most of the keyboard=20F > sequences make sense and can be usually guessed correctly to produce > such characters. >=20J > I have never ever seen that currency symbol used as such, or for anythi= ngI > other than a dingbat or a little sun symbol.  I guess that's why it wasR > recycled into the euro.    Here it is: =A4 .$  H I've seen it used in diagrams for electric circuits, IIRC to indicate=20D that current is flowing perpendicular to and away from the observer.   ------------------------------    Date: 12 Jun 2003 08:32:25 -0500- From: Kilgallen@SpamCop.net (Larry Kilgallen)t2 Subject: Re: Motif 1.3, Euro sign not standard ?!?3 Message-ID: <8a+B1+tH3Flw@eisner.encompasserve.org>r  w In article <01KX052QGPCYAM67GX@sysdev.deutsche-boerse.com>, Phillip Helbig <HELBPHI@sysdev.deutsche-boerse.com> writes: K >> The <compose> <x> <o> is because before the Euro sign was invented, the=u > reJ >> was a 'universal currency symbol' at that character position.  It looksI >> like a circle with four rays at the diagonals coming from it.  Thus it G >> looks somewhat like a superimposed x and o.  Most of the keyboard=20vG >> sequences make sense and can be usually guessed correctly to producee >> such characters.c >>=20dK >> I have never ever seen that currency symbol used as such, or for anythi=e > ngJ >> other than a dingbat or a little sun symbol.  I guess that's why it was >> recycled into the euro. >  > Here it is: =A4 .a  D Are you perhaps presuming that everyone reads comp.os.vms with a web	 browser ?a   ------------------------------  + Date: Thu, 12 Jun 2003 15:37:00 +0100 (MET)c9 From: Phillip Helbig <HELBPHI@sysdev.deutsche-boerse.com>l2 Subject: Re: Motif 1.3, Euro sign not standard ?!?; Message-ID: <01KX0G5GI1DSAOKN0V@sysdev.deutsche-boerse.com>n  J > > > I have never ever seen that currency symbol used as such, or for an= ythingJ > > > other than a dingbat or a little sun symbol.  I guess that's why it=  was > > > recycled into the euro.a > >=20 > > Here it is: =3DA4 .x >=20F > Are you perhaps presuming that everyone reads comp.os.vms with a web > browser ?t  I Not at all.  With VMS MAIL via Info-VAX or (will check tonight, but am=20lK pretty sure from other 8-bit experiences in the past) it shows up in the=20aG Newsgroup as I type it in, namely =A4, which is the circle with four=20h spikes.   G This has more to do with the issue of how 8-bit ASCII is transported=20 = throughout the internet than with the original topic, though.g   ------------------------------  + Date: Thu, 12 Jun 2003 15:03:48 +0000 (UTC)w7 From: moroney@world.std.spaamtrap.com (Michael Moroney) 2 Subject: Re: Motif 1.3, Euro sign not standard ?!?( Message-ID: <bca4ok$1lq$1@pcls4.std.com>  ; Phillip Helbig <HELBPHI@sysdev.deutsche-boerse.com> writes:   K >> > > I have never ever seen that currency symbol used as such, or for an=n >ythingnK >> > > other than a dingbat or a little sun symbol.  I guess that's why it=a > wasi >> > > recycled into the euro. >> >=20t >> > Here it is: =3DA4 . >>=20rG >> Are you perhaps presuming that everyone reads comp.os.vms with a webt >> browser ?  J >Not at all.  With VMS MAIL via Info-VAX or (will check tonight, but am=20L >pretty sure from other 8-bit experiences in the past) it shows up in the=20H >Newsgroup as I type it in, namely =A4, which is the circle with four=20 >spikes.  H >This has more to do with the issue of how 8-bit ASCII is transported=20> >throughout the internet than with the original topic, though.  A I deliberately did not try to type it in because I know sometimessH the transport drops the high bit, tries to convert it somehow (it showedK up here as =A4, 3 characters) or some even (blasphemy!) read this group on vF a M$ system which has its own idea what character is at that hex code.' (standards? we make our own standards!)    --   -Miker   ------------------------------    Date: 12 Jun 2003 11:25:23 -0500- From: Kilgallen@SpamCop.net (Larry Kilgallen) 2 Subject: Re: Motif 1.3, Euro sign not standard ?!?3 Message-ID: <iC1nkQGrm2R0@eisner.encompasserve.org>i  w In article <01KX0G5GI1DSAOKN0V@sysdev.deutsche-boerse.com>, Phillip Helbig <HELBPHI@sysdev.deutsche-boerse.com> writes:cK >> > > I have never ever seen that currency symbol used as such, or for an=  > ythingK >> > > other than a dingbat or a little sun symbol.  I guess that's why it=l >  was >> > > recycled into the euro. >> >=20c >> > Here it is: =3DA4 . >>=20cG >> Are you perhaps presuming that everyone reads comp.os.vms with a webR >> browser ? > K > Not at all.  With VMS MAIL via Info-VAX or (will check tonight, but am=20 M > pretty sure from other 8-bit experiences in the past) it shows up in the=20tI > Newsgroup as I type it in, namely =A4, which is the circle with four=20A	 > spikes.o  @ With ANUNEWS reading comp.os.vms it shows up as three charactersD "=", "A" and "4", and replying here with TECO I see that it is threeE characters (whereas reading an eight bit character with TECO shows upbB as the hex value in square brackets and is a single character when stepping over it.l  I > This has more to do with the issue of how 8-bit ASCII is transported=20 ? > throughout the internet than with the original topic, though.m  @ Yes, but original topic drift is a day-one feature of Usenet :-)   ------------------------------  % Date: Thu, 12 Jun 2003 18:06:42 +0200c$ From: Michael Unger <unger@decus.de>2 Subject: Re: Motif 1.3, Euro sign not standard ?!?5 Message-ID: <bcaa3k$guf74$1@ID-152801.news.dfncis.de>i  , On 12-Jun-2003 15:32, Larry Kilgallen wrote:  T > In article <01KX052QGPCYAM67GX@sysdev.deutsche-boerse.com>, Phillip Helbig writes:	 >>> [...]  >>>uL >>> I have never ever seen that currency symbol used as such, or for anythi= >> ngcK >>> other than a dingbat or a little sun symbol.  I guess that's why it was  >>> recycled into the euro.  >> l >> Here it is: =A4 . > F > Are you perhaps presuming that everyone reads comp.os.vms with a web > browser ?   H Phillip has an additional "Content-Type: text/plain; charset=iso-8859-1"D header in his mail; the "currency" character is visible. You, Larry,E have no "Content-Type" header; the "currency" ist displayed as "=A4",  apparently the ASCII value.7  - Mozilla Mail-News (V 1.3.1 EN) is doing fine.:   Michael    -- g  @ Please do *not* send "Security Patch Notifications" or "SecurityA Updates"; this system isn't running a Micro$oft operating system.a= And don't annoy me <mailto:postmaster@[127.0.0.1]> please ;-)    ------------------------------    Date: 12 Jun 2003 11:45:25 -0500; From: koehler@eisner.nospam.encompasserve.org (Bob Koehler)r2 Subject: Re: Motif 1.3, Euro sign not standard ?!?3 Message-ID: <MlytZRsLJA55@eisner.encompasserve.org>i  \ In article <bcaa3k$guf74$1@ID-152801.news.dfncis.de>, Michael Unger <unger@decus.de> writes: > J > Phillip has an additional "Content-Type: text/plain; charset=iso-8859-1"F > header in his mail; the "currency" character is visible. You, Larry,G > have no "Content-Type" header; the "currency" ist displayed as "=A4",- > apparently the ASCII value.2  E    I'm reading it via putty, and ANUNEWS, which displays it as "=A4".e?    I believe is I was using a DECterm instead of putty it wouldrC    come out better (other characters have in the past), even thoughi    I'd still be using ANUNEWS.   ------------------------------  % Date: Thu, 12 Jun 2003 10:39:18 -0500a* From: Patrick Spinler <pspinler@yahoo.com> Subject: Re: perl question+ Message-ID: <bca6rm$5ka$1@tribune.mayo.edu>n   Chris Olive wrote: >   >? >>> 3) SET VERIFY and F$VERIFY() come in handy at times in DCL.y >> >> >>I >> Yeah, this would be nice in Perl.  I don't know of a way to do this inhH >> Perl.  Maybe with the debugger.  At least Perl *has* a debugguer.  :) >> > H > Data::Dumper helps some in lieu of SET VERIFY, but not the same thing F > for sure.  But Perl has "use strict" and DCL doesn't, so there is a  > tradeoff there, somewhat.m  F As hinted above, the perl debugger can do this handily, using the 't' H (trace), and 'c' (continue) commands inside the debugger.  for instance  (forgive the line wrapping):      $ perl -d test.pl     Default die handler restored.  3    Loading DB routines from perl5db.pl version 1.07     Editor support available.  C    Enter h or `h h' for help, or `perldoc perldebug' for more help.   8    main::(test.pl:2):      $ exit++ + ++$status !=0 and  $exit=$status=undef;  whilee1 ($#ARGV > -1 && $ARGV[$#ARGV] eq '') {pop @ARGV;}0      DB<1> t
    Trace = ono      DB<1> c8    main::(test.pl:2):      $ exit++ + ++$status !=0 and  $exit=$status=undef;  while 1 ($#ARGV > -1 && $ARGV[$#ARGV] eq '') {pop @ARGV;}i9    main::(test.pl:4):      open (IN, "<", 'sys$command');y+    main::(test.pl:5):      my $resp = <IN>;d    foo'    main::(test.pl:6):      chomp $resp;I8    main::(test.pl:7):      print "Response = ($resp)\n";    Response = (foo)y    F Also check out the "-D" option in the perl run documentation (perldoc 	 perlrun).e  I In particular, perl "-Dt" or "-Dtls", while _considerably_ more verbose,  E comes close to what set verify does, although it does require a perl q compiled with debugging.     -- Pat   ------------------------------  % Date: Thu, 12 Jun 2003 09:40:05 -0500i' From: Chris Olive <nospam@raytheon.com>  Subject: Re: perl question> Message-ID: <1i0Ga.3540$c6.3329@bos-service2.ext.raytheon.com>   Ryan Moore wrote:e) > On Wed, 11 Jun 2003, Chris Olive wrote:c >  >>Craig A. Berry wrote:n >>p >>>Chris Olive <nospam@raytheon.com> wrote in message news:<cK7Fa.3506$c6.3314@bos-service2.ext.raytheon.com>... >>>-3 >>>Just curious.  What things do you miss the most?  >>G >>1) The exact way that the ' character works in DCL is absent in Perl.  >> >>Trite Example: >>	 >>$ i = 0 
 >>$ while: >>$    if (i .ge. 10) goto end >>$    var'i = i >>$    i = i + 1 >>$    goto whilei >>$ end: >  >  > You can do this with eval(). >  > for $i (1..10) > {, >         eval "\$var$i = $i"; > }i >   > print "$var1, $var2, $var3\n"; > 
 > will print:e	 > 1, 2, 3i >   E Right.  There is also at least one other way to do it that I've seen hE before.  Can't quite remember it off the top of my head...  It's not a much more involved than eval().v   > 1 >>2) I also like the := and :== operators in DCL.u >  > L > You mean like '=' vs. ':='?  Kinda like the difference between a string inG > single quotes vs. a string in double quotes in Perl.  That's the best 
 > analogy. >   1 Eeeh... KINDA of...  I'm thinking more like this:   ) $ username := 'f$getjpi( "", "USERNAME" )   G You can't do that in Perl exactly that way.  I think it can be done on lC one line in Perl, but not as easily.  This is what I mean by the ' t> character and the := sign.  Just seems easier in DCL this way.   > = >>3) SET VERIFY and F$VERIFY() come in handy at times in DCL.m >  > H > Yeah, this would be nice in Perl.  I don't know of a way to do this inG > Perl.  Maybe with the debugger.  At least Perl *has* a debugguer.  :)e >   F Data::Dumper helps some in lieu of SET VERIFY, but not the same thing D for sure.  But Perl has "use strict" and DCL doesn't, so there is a  tradeoff there, somewhat.o   > I >>Perl has been my "DCL fix" language everywhere else for reasons alreadym0 >>mentioned.  It just "feels like" a "Unix DCL." >  > F > Of course, it's mostly based on shell and C.  Both of those are Unix > creatures. >   F Yeah, but C Shell, Bourne Shell and Korn Shell don't give me the same  "DCL feeling," only Perl.u   Chrisl -----S Chris Olive  Systems Consultant' Raytheon Technical Services Corporation  Indianapolis, IN  * email: olivec(AT)indy(DOT)raytheon(DOT)com   ------------------------------  % Date: Thu, 12 Jun 2003 10:48:21 -0500f* From: Patrick Spinler <pspinler@yahoo.com> Subject: Re: perl question+ Message-ID: <bca7ck$6b3$1@tribune.mayo.edu>e   Patrick Spinler wrote: > Chris Olive wrote: >=20 >> >  > >=20@ >>>> 3) SET VERIFY and F$VERIFY() come in handy at times in DCL. >>>o >>>A >>>a >>>oJ >>> Yeah, this would be nice in Perl.  I don't know of a way to do this i= niJ >>> Perl.  Maybe with the debugger.  At least Perl *has* a debugguer.  :)=   >>>  >>J >> Data::Dumper helps some in lieu of SET VERIFY, but not the same thing =  I >> for sure.  But Perl has "use strict" and DCL doesn't, so there is a=20n >> tradeoff there, somewhat. >=20 >=20   (snip earlier methods)  J Sometimes I'm an idiot.  I keep forgetting about this, but the following =  4 is cut directly from the perlrun documentation page:  J      If you=92re just trying to get a print out of each line of Perl code=  E      as it executes, the way that "sh -x" provides for shell scripts,A9      you can=92t use Perl=92s -D switch.  Instead do thise         # If you have "env" utilityJ      env=3DPERLDB_OPTS=3D"NonStop=3D1 AutoTrace=3D1 frame=3D2" perl -dS p= rogram        # Bourne shell syntaxJ      $ PERLDB_OPTS=3D"NonStop=3D1 AutoTrace=3D1 frame=3D2" perl -dS progr= am        # csh syntaxtJ      % (setenv PERLDB_OPTS "NonStop=3D1 AutoTrace=3D1 frame=3D2"; perl -d= S=20 program)  .      See perldebug for details and variations.   Or, on VMS:e  =    $ define perldb_opts "NonStop=3D1 AutoTrace=3D1 frame=3D2":    $ perl "-dS" test.plu     Default die handler restored.    Package test.pl.aE    2:      $ exit++ + ++$status !=3D0 and $exit=3D$status=3Dundef;=20m while($#ARGV > -1 &&! $ARGV[$#ARGV] eq '') {pop @ARGV;}oE    2:      $ exit++ + ++$status !=3D0 and $exit=3D$status=3Dundef;=20e while($#ARGV > -1 &&! $ARGV[$#ARGV] eq '') {pop @ARGV;}r)    4:      open (IN, "<", 'sys$command');e    5:      my $resp =3D <IN>;     foo    6:      chomp $resp;c*    7:      print "Response =3D ($resp)\n";    Response =3D (foo)a    $  ( Again, please forgive the line wrapping.   -- Pat   >=20 >=20 > -- Pat >=20   ------------------------------  % Date: Thu, 12 Jun 2003 07:33:44 -0500m, From: Gary L. Ross <rossgl@parknicollet.com>, Subject: Re: PIPE output defined to a symbol8 Message-ID: <ijsgevoarhnibgv1a0aqlka4n8rskaqbh1@4ax.com>  F On Wed, 11 Jun 2003 10:21:47 -0500, brandon@dalsemi.com (John Brandon) wrote:  H >I used the examples provided earlier in this thread to do a performanceE >comparison between the PIPE (NEW) and /OUTPUT to disk (OLD) methods.b >  >  >The NEW method: >e >$ TYPE GET_MEMORY_NEW.COM;d >$!c? >$! this is the NEW method to get a SEARCH string into a symbol  >$!a >$ pipe ( -a >         show memory | --. >         search sys$pipe "Private Memory" | -
 >         ( -.# >           read sys$pipe $tmp$ ; -H+ >           define/job/nolog $tmp$ &$tmp$ -K
 >         ) -e	 >       ) G >$ private_memory = f$element(0,")",f$element(1,"(",f$trnlnm("$tmp$")))oD >$ write sys$output "This server has ''private_memory' total memory" >S >$ @GET_MEMORY_NEW.COM;n' >This server has 6332.00Mb total memoryw >w >  >The OLD method: >g >$ TYPE GET_MEMORY_OLD.COM;v >$!e? >$! this is the OLD method to get a SEARCH string into a symbol> >$!t >$ show memory /out=search.txt; 8 >$ search search.txt; "Private Memory" /out=results.txt; >$ open rdfl results.txt; /read  >$ read rdfl datat
 >$ close rdfl $ >$ delete results.txt;*,search.txt;*: >$ private_memory = f$element(0,")",f$element(1,"(",data))D >$ write sys$output "This server has ''private_memory' total memory" >K >$ @GET_MEMORY_OLD.COM; ' >This server has 6332.00Mb total memorye >y >o
 >Analysis: > N >In this example, the NEW method will have the following impact on performance >over the OLD method:  > & >Increase	[Buffered I/O count]	by 230%# >Decrease	[Direct I/O count]	by 25%B >Decrease	[Page faults]		by 55%o# >Decrease	[Charged CPU time]	by 47%S >OL >Overall a reduction in system resources for said commands.  Your milage may >vary. >SK >This is not to imply that you will reduce overall costs by any great deal,aP >however if your process does multiple searches as demonstrated above, it may be> >worth the time and effort to learn and implement this method. >- >-# >Thanks for the submitted examples!0 >94 >Call me what you will, just do not call me Shirley! >c
 >John Brandonh >VMS Systems Administrator% >john.brandon.spam.me.not@dalsemi.coma  F Thanks everyone for the help and examples.  It has been very helpful!!   Gary L. Ross Park Nicollet Health Servicese   ------------------------------    Date: 12 Jun 2003 08:27:53 -0500 From: briggs@encompasserve.org, Subject: Re: PIPE output defined to a symbol3 Message-ID: <Z33+PuUgp51F@eisner.encompasserve.org>i  \ In article <03061110214745@dscis6-0.dalsemi.com>, brandon@dalsemi.com (John Brandon) writes:I > I used the examples provided earlier in this thread to do a performancelF > comparison between the PIPE (NEW) and /OUTPUT to disk (OLD) methods. >  >  > The NEW method:  >  > $ TYPE GET_MEMORY_NEW.COM; > $!@ > $! this is the NEW method to get a SEARCH string into a symbol > $! > $ pipe ( - >          show memory | -/ >          search sys$pipe "Private Memory" | -w >          ( -$ >            read sys$pipe $tmp$ ; -, >            define/job/nolog $tmp$ &$tmp$ - >          ) -
 >        )H > $ private_memory = f$element(0,")",f$element(1,"(",f$trnlnm("$tmp$")))E > $ write sys$output "This server has ''private_memory' total memory"   B As you know, PIPE is based heavily on what is essentially a SPAWN.( The syntax above involves three "SPAWN"s  B You can get major performance boosts from SPAWN by suppressing the@ propagation of DCL symbols and process logicals into the spawned subprocess.n  < You can get the same performance boost from PIPE in the same manner.u  & $ PIPE /NOSYMBOLS /NOLOGICAL_NAMES ...  I In the case at hand, DCL symbols and process logical names are irrelevantfG to the correct functioning of the PIPE subcommands.  So, if performanceiD is important, using /NOSYMBOLS and /NOLOGICAL_NAMES is a no brainer.  O > In this example, the NEW method will have the following impact on performancee > over the OLD method: > ' > Increase	[Buffered I/O count]	by 230%i$ > Decrease	[Direct I/O count]	by 25%  > Decrease	[Page faults]		by 55%$ > Decrease	[Charged CPU time]	by 47%  F The propagation of symbols and logicals is done by mailbox I/O betweenD the parent process and the subprocess.  Mailbox I/O is buffered I/O.G You should expect to see that buffered I/O count reduced significantly. E You are also likely to pick up a nice reduction in CPU usage as well.e  A I haven't done careful measurements in quite some time.  But I doo= remember that from a wall clock point of view, the benefit ofh? /NOSYMBOLS and /NOLOGICAL_NAMES is dramatic (but then I'm in an C environment where we have a ton and a half of DCL symbols defined).5   	John Briggs   ------------------------------  % Date: Thu, 12 Jun 2003 09:13:58 -0500c( From: brandon@dalsemi.com (John Brandon), Subject: Re: PIPE output defined to a symbol1 Message-ID: <03061209135827@dscis6-0.dalsemi.com>o  D > You can get major performance boosts from SPAWN by suppressing theB > propagation of DCL symbols and process logicals into the spawned
 > subprocess.  > > > You can get the same performance boost from PIPE in the same	 > manner.n > ( > $ PIPE /NOSYMBOLS /NOLOGICAL_NAMES ... > K > In the case at hand, DCL symbols and process logical names are irrelevantmI > to the correct functioning of the PIPE subcommands.  So, if performanceNF > is important, using /NOSYMBOLS and /NOLOGICAL_NAMES is a no brainer. > Q > > In this example, the NEW method will have the following impact on performance  > > over the OLD method: > > ) > > Increase	[Buffered I/O count]	by 230%t& > > Decrease	[Direct I/O count]	by 25%" > > Decrease	[Page faults]		by 55%& > > Decrease	[Charged CPU time]	by 47%  M I added the recommendations (/NOSYMBOLS,/NOLOGICAL_NAMES) and re-ran the teste$ code and came up with the following:  : Increase    [Buffered I/O count]    by 200%    (30% delta): Decrease    [Direct I/O count]      by 25%     (NO CHANGE): Decrease    [Page faults]           by 80%     (55% delta): Decrease    [Charged CPU time]      by 95%     (40% delta)   Much much better!   H > The propagation of symbols and logicals is done by mailbox I/O betweenF > the parent process and the subprocess.  Mailbox I/O is buffered I/O.I > You should expect to see that buffered I/O count reduced significantly.iG > You are also likely to pick up a nice reduction in CPU usage as well.  > C > I haven't done careful measurements in quite some time.  But I do ? > remember that from a wall clock point of view, the benefit ofiA > /NOSYMBOLS and /NOLOGICAL_NAMES is dramatic (but then I'm in anrE > environment where we have a ton and a half of DCL symbols defined).- >  > 	John Briggs  O And these number are not "careful" either - just shotgun (rounded) - however it G still demonstrates the savings.  Thanks to John Briggs for the insight!3       J*o*h*n B*r*a*n*d*o*nn VMS Systems Administratorr* firstname.lastname.spam.me.not@dalsemi.com   ------------------------------  + Date: Thu, 12 Jun 2003 08:43:58 +0000 (UTC) . From: "Sanface Software" <sanface@sanface.com> Subject: PLUG: txt2pdf 6.6H Message-ID: <f3358a3446b579b6d3bb8fb1b1b4958b.93245@mygate.mailgate.org>  / We would like to announce txt2pdf 6.6 version. n# http://www.sanface.com/txt2pdf.htmldE txt2pdf is shareware; it is a very flexible and powerful Perl5 scriptaH that converts text files to PDF format files, so you can use it in every8 operating systems supported by Perl5, including OpenVMS.; It's simple to design background like invoices, orders etc.u) Here nice examples made using txt2pdf PRO - http://www.sanface.com/pdf/Purchase_Order.pdfP& http://www.sanface.com/pdf/invoice.pdf$ http://www.sanface.com/pdf/hfmus.pdf) http://www.sanface.com/pdf/heraldbill.pdfI@ If you prefer we also distribute executables for Windows, Linux,F Solaris, AIX, HP-UX, and Mac OS X (if you need also Tru64). Inside the, Windows version is Visual txt2pdf, a VB GUI.   What's new in this version  7 It's possible to set the PDF version (default is 1.2). vD txt2pdf.vim update (vim plug-in to convert using vim text into pdf).   Test txt2pdf 6.6!06 You can find it at http://www.sanface.com/txt2pdf.html     -- a8 Posted via Mailgate.ORG Server - http://www.Mailgate.ORG   ------------------------------  % Date: Thu, 12 Jun 2003 01:52:02 -0400t* From: "Bill Todd" <billtodd@metrocast.net>& Subject: Re: Portents of Itanium death2 Message-ID: <vxSdncZE9sqRiHWjXTWcqw@metrocast.net>  8 "Rob Young" <young_r@encompasserve.org> wrote in message- news:i7jod7U5lWnC@eisner.encompasserve.org... H > In article <3EE7F0D2.89F34F6C@istop.com>, JF Mezei <jfmezei@istop.com> writes:W > > Rob Young wrote:G > >>         No.  That was all tongue-in-cheek.  Your less than stellar G > >>         analysis has it upside down.  Opteron will be hard-pressedf" > >>         to keep up with Xeon. > >dJ > > That is the crux of the problem, because your argument applies to IA64 asJ > > well. IA64 will be hard pressed to keep up with the 8086 architecture, exceptJ > > in certain configurations where the 8086 doesn't have the interconnect > > hardware needed. >n >e* > But Xeon won't be outperforming Madison.  K Of course it will, as will Opteron if its clock rate increases on schedule.tI Even if Madison attains the SPECint score of 1250 you recently predicted,cJ Xeon should equal that at 3.2 GHz and Opteron should equal it at between 2K and 2.2 GHz - both of which are scheduled to occur in a few months at most.n     Madison won'to8 > be blowing it away by any stretch (other than floating< > point and highend database work) - but it doesn't have to. >e9 > Here is another neat wrinkle.  Deerfield is a great HPC- > part.  >e- > You heard it from Paul DeMone on RWT first:e >t >rL http://www.realworldtech.com/forums/index.cfm?action=detail&PostNum=1437&Thr ead=19&entryID=18552&roomID=11 >8J > >> Maybe, maybe not. A 2.6 GHz Opteron would still consume less than 90W whilegL > >> a 1.5-1.8 Ghz Madison would be in the 130W range, 50% higher. Maybe the highK > >> wattage Madisons will generate 50% higher Linpack Gflops to compensatet butw% > >> it sure looks like a close call.  > > L > >How about a 1.3 GHz Madison in the mid 60 Watt range that cranks out >50% > >more Linpack MFLOP/s? >   > Very, very impressive if true.  L The operative phrase very definitely being the 'if true' part.  Last I knew,G Intel was stating that the 1 GHz, 1.5 MB cache version of Madison wouldiH consume 62 W, so either Intel or Paul wouldn't seem to know what they'reD talking about (perhaps he's just confused as you so often are by his# enthusiasm for most things Itanic):a  L http://www.computerworld.com/hardwaretopics/hardware/server/story/0,10801,78 628,00.html1   - bill   ------------------------------  % Date: Thu, 12 Jun 2003 02:22:19 -0400:" From: JF Mezei <jfmezei@istop.com>& Subject: Re: Portents of Itanium death) Message-ID: <3EE81C1A.9A938670@istop.com>a   Rob Young wrote:I >         Point is Deerfield has Opteron stomped when it comes to a lowera- >         powered higher performing HPC part.p    J Woopty doo. Deerfield is even more vapourware than Madison that isn't even officially out yet.   K Look at EV7 and EV79 that had been supposed to blow everyone out, but as itrL turns out, for whatever reason, they came out very late and with performanceM that hasn't made the news "HP hiding the gem it killed because it performs soV1 much better than the bloated IA64 HP has chosen".N  N Guess what while you might be impressed with what Intel tells you about futureK IA64s, you have to compare that with whatever else will be available at thetL time the chip makes it to customers. Comparing it to today's performance may+ make a "wow", but it doesn't mean anything.s  M If Deerfield doubles the speed, but at the same time, the 8086 gets 2.5 speedtL boots and Power gets 3 times a speed boost, then Deerfield will be seen as a% slow turtle by the time it comes out.-  N And so far, the track record of IA64 has been that by the time a chip actually5 makes it to customers, its performance is lacklustre.t  M I don't care about what performance some beta chip gets in a lab. If it takessN X amount of time to move it from the lab to production and then to customers, Q then it is the performance comparison at the end of X amount of time that counts.t  M Right now, the experience/history of IA64 is that every release (Merced was arG disaster, McKinley was unipressive) has been a letdown and Intel has toeH convince people that the future will be brighter. But those who need the: performance today aren't impressed with what is available.   ------------------------------  % Date: Thu, 12 Jun 2003 13:01:51 +0530m4 From: Kesav Tadimeti <Kesav_Tadimeti@KeaneIndia.com>& Subject: Re: Portents of Itanium deathK Message-ID: <8EA11405E59BD611BA7100104B93C2600181A051@exdel01.del.mgsl.com>a  J But even about 1-2 years back, when COMPAQ made the decision to stop AlphaH and adopt Itanic, it wasn't too late for Intel to abandon IA-64 and takeH Alpha under its wings (esp since they bought the Fabs, the compiler teamH ..). THey could have called the alpha as Itanium-2. It might have causedE them some loss of face, but they would have saved money. And IA-64 (aLH badge-engineered Alpha) could ship today. At the very least, Alpha couldK have continued to be produced until IA-64 became a mighty force IBM and Suna5 would have had to recon with, as someone pointed out.t  # JF Mezei <jfmezei@istop.com> wrote:g   Wiser ? Of course.  K But remember that IA64 started life a long long long time ago, when Digital J was still alive. HP wanted to replace Pa-Risc and teamed up with Intel.  IL don't think that Digital would have wanted to sell its Alpha which was stillG fairly new at the time to a competitor, especially since Digital wasn'tt" starved for cash yet at that time.  K And once HP decided with intel to come up with some totally new chip, there > was no going back. So now they are stuck with that IA64 thing.   ------------------------------  % Date: Thu, 12 Jun 2003 13:10:31 +0530y4 From: Kesav Tadimeti <Kesav_Tadimeti@KeaneIndia.com>& Subject: Re: Portents of Itanium deathK Message-ID: <8EA11405E59BD611BA7100104B93C2600181A06B@exdel01.del.mgsl.com>o  * From: "John Smith" <a@nonymous.com> wrote:7 >Sun/McNeely aren't stupid the way Compaq/Capellas was.vH I read somewhere (may be on c.o.v, can't be sure), that there was a timeC when SUn had to make a decision as to whether or not to make Wintel(J computers. Scott was the only person to say a flat *NO* to that, at a time6 when almost all big Unix vendors were making wintels.   % I wish Digital had someone like that.C   ------------------------------    Date: 12 Jun 2003 07:41:30 -0500; From: koehler@eisner.nospam.encompasserve.org (Bob Koehler) & Subject: Re: Portents of Itanium death3 Message-ID: <akaJzYAQJDyF@eisner.encompasserve.org>.   In article <8EA11405E59BD611BA7100104B93C2600181A051@exdel01.del.mgsl.com>, Kesav Tadimeti <Kesav_Tadimeti@KeaneIndia.com> writes:L > But even about 1-2 years back, when COMPAQ made the decision to stop AlphaJ > and adopt Itanic, it wasn't too late for Intel to abandon IA-64 and takeJ > Alpha under its wings (esp since they bought the Fabs, the compiler teamJ > ..). THey could have called the alpha as Itanium-2. It might have causedG > them some loss of face, but they would have saved money. And IA-64 (auJ > badge-engineered Alpha) could ship today. At the very least, Alpha couldM > have continued to be produced until IA-64 became a mighty force IBM and Sun 7 > would have had to recon with, as someone pointed out.o  H    And they could sell it to all those PC vendors so the customers could*    run how many of their x86 applications?   ------------------------------    Date: 12 Jun 2003 07:39:29 -0500; From: koehler@eisner.nospam.encompasserve.org (Bob Koehler)o& Subject: Re: Portents of Itanium death3 Message-ID: <LO7P8QiX1ReR@eisner.encompasserve.org>s  N In article <3EE7A1A5.8D872329@istop.com>, JF Mezei <jfmezei@istop.com> writes: > P > Sun is smart to keep its own products. Heck, with Linux threathening Unix, SunP > would be smart to buy VMS and provide a true enterprise system that would haveT > a huge advantage over HP and IBM's unix offerings. (think VMS with sun marketing).  F    Sun invested a major effort to get across the notion that if you doD    anything other than UNIX, you can't really be serious about UNIX.  E    Investing in VMS would make Sun look about as reputable as Compaq.o   ------------------------------   Date: 12 Jun 2003 15:22:14 GMT( From: bill@cs.uofs.edu (Bill Gunshannon)& Subject: Re: Portents of Itanium death5 Message-ID: <bca5r6$h4cn7$1@ID-135708.news.dfncis.de>>  3 In article <akaJzYAQJDyF@eisner.encompasserve.org>,s> 	koehler@eisner.nospam.encompasserve.org (Bob Koehler) writes: > In article <8EA11405E59BD611BA7100104B93C2600181A051@exdel01.del.mgsl.com>, Kesav Tadimeti <Kesav_Tadimeti@KeaneIndia.com> writes:M >> But even about 1-2 years back, when COMPAQ made the decision to stop AlphanK >> and adopt Itanic, it wasn't too late for Intel to abandon IA-64 and takesK >> Alpha under its wings (esp since they bought the Fabs, the compiler team K >> ..). THey could have called the alpha as Itanium-2. It might have caused H >> them some loss of face, but they would have saved money. And IA-64 (aK >> badge-engineered Alpha) could ship today. At the very least, Alpha couldlN >> have continued to be produced until IA-64 became a mighty force IBM and Sun8 >> would have had to recon with, as someone pointed out. > J >    And they could sell it to all those PC vendors so the customers could, >    run how many of their x86 applications?  3D The same number they can run on their IA64.  What's your point?  :-)   bill   --  J Bill Gunshannon          |  de-moc-ra-cy (di mok' ra see) n.  Three wolvesD bill@cs.scranton.edu     |  and a sheep voting on what's for dinner. University of Scranton   |A Scranton, Pennsylvania   |         #include <std.disclaimer.h>   y   ------------------------------  # Date: Thu, 12 Jun 2003 16:05:39 GMTa# From: "John Smith" <a@nonymous.com>-& Subject: Re: Portents of Itanium deathG Message-ID: <nx1Ga.65564$G_.40242@news02.bloor.is.net.cable.rogers.com>r  @ "Bob Koehler" <koehler@eisner.nospam.encompasserve.org> wrote in5 message news:LO7P8QiX1ReR@eisner.encompasserve.org... 4 > In article <3EE7A1A5.8D872329@istop.com>, JF Mezei <jfmezei@istop.com> writes:  > > ; > > Sun is smart to keep its own products. Heck, with Linuxo threathening Unix, SunB > > would be smart to buy VMS and provide a true enterprise system that would haveiF > > a huge advantage over HP and IBM's unix offerings. (think VMS with sun marketing).e >oE >    Sun invested a major effort to get across the notion that if yout doF >    anything other than UNIX, you can't really be serious about UNIX.  B Actually it was that if you do any other unix than Solaris, you're@ joking, and that if you don't do unix at all, you can't be taken
 seriously.      ? >    Investing in VMS would make Sun look about as reputable aso Compaq.   E They could probably 'spin' it to work out ok. It was the marketing of D their brand of 'snake oil' that caused DEC so much grief in the late 80's/early 90's.   ------------------------------    Date: 12 Jun 2003 11:53:07 -0500; From: koehler@eisner.nospam.encompasserve.org (Bob Koehler)b& Subject: Re: Portents of Itanium death3 Message-ID: <R1$Yiva2qMxS@eisner.encompasserve.org>>  ` In article <bca5r6$h4cn7$1@ID-135708.news.dfncis.de>, bill@cs.uofs.edu (Bill Gunshannon) writes:5 > In article <akaJzYAQJDyF@eisner.encompasserve.org>,e@ > 	koehler@eisner.nospam.encompasserve.org (Bob Koehler) writes:  K >>    And they could sell it to all those PC vendors so the customers coulds- >>    run how many of their x86 applications?y >  rF > The same number they can run on their IA64.  What's your point?  :-)  F    So all that compatability stuff Intel puts in the IA64 architecture'    manuals is just there to kill trees?h   ------------------------------  % Date: Thu, 12 Jun 2003 13:20:40 -0400e* From: "Bill Todd" <billtodd@metrocast.net>& Subject: Re: Portents of Itanium death2 Message-ID: <y92dnSrXuL3hK3WjXTWcpQ@metrocast.net>  H "Bob Koehler" <koehler@eisner.nospam.encompasserve.org> wrote in message- news:R1$Yiva2qMxS@eisner.encompasserve.org...dH > In article <bca5r6$h4cn7$1@ID-135708.news.dfncis.de>, bill@cs.uofs.edu (Bill Gunshannon) writes:a7 > > In article <akaJzYAQJDyF@eisner.encompasserve.org>, A > > koehler@eisner.nospam.encompasserve.org (Bob Koehler) writes:= >wG > >>    And they could sell it to all those PC vendors so the customers0 could / > >>    run how many of their x86 applications?. > >iH > > The same number they can run on their IA64.  What's your point?  :-) >eH >    So all that compatability stuff Intel puts in the IA64 architecture) >    manuals is just there to kill trees?t  K Pretty much so.  According to an HP white paper released with McKinley lastmL year, IA32 code on McKinley runs at approximately Pentium Pro 300 MHz speed.I This makes Itanic's 'compatibility' about as useful as Alpha's FX!32 was:-I Alpha ran IA32 code comparatively faster than Itanic's hardware emulation.H does, but my impression is could not run *all* IA32 applications without change, so it's about a wash.a  L Another way of assessing the quality of Itanic's IA32 hardware support is inH the light of the recent Intel announcement that they're going to releaseD their own (FX!32-like?) software emulation as well to offer improved performance...   - bill   ------------------------------  # Date: Thu, 12 Jun 2003 14:34:53 GMT:* From: Paul Anderson <paul.anderson@hp.com>H Subject: Re: Printer recommendation, hobbyist/decwindows/mozilla and VMS5 Message-ID: <120620031015418353%paul.anderson@hp.com>b  F In article <3EE79F0C.DE53DC8F@istop.com>, JF Mezei <jfmezei@istop.com> wrote:   > Paul Anderson wrote:G > > DCPS today requires bi-directional communication with the printer'sa > > PostScript interpreter.f > D > Why the "require" ?  Couldn't you define a class of printers whereE > bidirectional comms would not be needed and where DCPS would simplycF > rely on the line's flow control mechanisms ? One would still benefit@ > from DCPS services such as on-the fly translation from text to > postscript etc.d   Note that I said *today*.   H We're adding LPD support in the next version, which is uni-directional. B A locally-attached printer would not use LPD, but as I said, we're= considering parallel and/or USB support for a future release.m  " > Isn't VMS going to support USB ?  G Yes.  In addition, we would also have to do some work in DCPS to get ite to support USB.i  5 In article <3ee73f79$1@cpns1.saic.com>, Mark Berryman.# <Mark.Berryman@Mvb.Saic.Com> wrote:d   > Paul Anderson wrote:F > > But how many people would have an inkjet printer attached to their4 > > OpenVMS system via a parallel or USB connection? > : > Um, I currently have two, one on each type of interface. : B > If DCPS supported the USB or parallel port I suspect more people > would hook them up that way.  : I don't doubt there are people, like yourself, with such aA configuration.  The problem is determining how many people do and > whether engineering effort should be spent on this, vs. other,	 features.2  F > How about an execution queue where DCPS interfaces with Ghostscript F > rather than a physical printer?  Ghostscript will output to several J > different native printer formats and the resulting file could be passed 7 > on to a print queue for actual output to the printer.r  D That's an interesting idea.  But DCPS uses PostScript to control theA job itself, so even a PCL print job would have PostScript wrappedt7 around it.  I wonder how Ghostscript would handle that.o   Paul   --    Paul Anderson   OpenVMS Engineering    Hewlett-Packard Companyr   ------------------------------  % Date: Thu, 12 Jun 2003 12:17:04 +0200 ! From: "Tim" <tim_a_b@hotmail.com>09 Subject: Problems running a .com file using a Java socketf/ Message-ID: <vegkovn8natp79@corp.supernews.com>a  
 Hello all,  K I have written a socket program to run on our VMS box that I want to use ton run .com files remotely.  K The socket program opens a server socket on a known port and just waits for  connections.  J Once it receives a connection it reads commands from the client. If one ofL these commands is 'switch A' then it needs to close the already open version+ of a program and then start up version 'A'.   ? To do this manually from the VMS console I have 2 .com scripts:h     run.comw     switch.com  ( So to switch to A manually, I would type
     @run stop8
     @switch Ai     @run  H All of those scripts work fine at the command prompt, but when I execute- them using, (just run by itself for example):   7     switchParams = new String [] {"iris:[src]run.com"};EE     BufferedReader br = new BufferedReader(new InputStreamReader((p =y* rt.exec(switchParams)).getInputStream()));     p.waitFor();  K I read the input from br and I see that the script has run because there is@I a lot of output saying that it is doing certain stuff and the p.waitFor()2K seems to wait for the same amount of time as the run command usually takes.r  K However, once 'run' has completed, a new application should have started on5G my VMS box and there should be an application window along with it, but. there is nothing.   H So what am I missing? I know that as well as parameters, I can also sendE environment variables to the exec command. So is there an environment"K variable in VMS that I need to set? I am from a Unix background and totallyrG new to VMS so I don't know what these are, although I think they may ber logicals in VMS?  J Just to clarify - I can run com files successfully that do not start otherJ GUI applications, but when I run a com script that launches a GUI, the GUIL part is not visible to me. oh yes, and I am running java 1.1.8 and have been* told not to upgrade to a later version :o(   Any help is much appreciated,e   Regards,   Timo   ------------------------------  % Date: Thu, 12 Jun 2003 07:40:07 +0200n From: Dirk Munk <munk@home.nl>, Subject: Re: Rookie question for the experts2 Message-ID: <bc941o$v7n$1@news3.tilbu1.nb.home.nl>  
 JustMe wrote: K > Just make sure you start up DECnet on the required line (EWA0) before youlJ > start LAT.  My recollection from older (VAX) versions was that LAT wouldL > "take over" the adapter if started first, and then DECnet wouldn't be able > to start the line.  N Yep, when the interface is used for DECnet phase IV (compatible) traffic, the N physical address is changed to AA-00-04-00-aa-bb. This can only be done if no  other protocol is active.    > D > The SYSTARTUP_VMS.com template has the sequencing like that, iirc. > J > (btw, I think the IP equivalent of MOP is BOOTP not TFTP, but I could be' > mistaken -- wouldn't be the 1st time)o  Q If I'm not mistaken the 'call for help' is BOOTP, and the actual loading is TFTP.    > K > If you install TSM for your server management, and define the servers andqL > their characteristics, its pretty straightforward from there.  We only addI > the info to the NCP (Phase IV) database to permit an alternate means of$L > 'talking' to the server for mgt. purposes (NCP > conn node xxxxxx), pretty. > sure that TSM maintains its info separately. >  > - > "Dirk Munk" <munk@home.nl> wrote in message:. > news:bc8247$vs9$1@news3.tilbu1.nb.home.nl... > Rusty Ferrell wrote: >  >>Greetings all!H >>I'm a UNIX kind-of-guy that's been put into a VMS-kind-of-guy position >>so please be gentle with me! >>D >>Our existing system is an AlphaServer 2100 and it's being replacedE >>with a DS10L AlphaServer. The 2100 is FDDI attached to our internalpE >>TCP/IP based network. There are no other network connections to theaC >>2100. In order for the old DECserver's and devices to "talk" (viae >>DECnet I assume) >  > G > No, 'classic' DECservers use MOP (Maintenance Operations Protocol) to 
 > downloadL > their software, and LAT (Local Area Transport) for their normal operation. > 1 > Newer versions can use TFTP and Telnet as well.  > J > MOP and LAT are non-routable protocols and have nothing (!!!) to do with	 > DECnet.  > N > The only problem is that the MOP entries for the software for the DECservers > areLM > placed in the DECnet databases (Phase IV) or configuration files (Phase V /i> > OSI), so you need DECnet to be able to load your DECservers. >  > 6 >>to the 2100, a small ethernet hub was installed thatG >>acts as a media converter/bridge between the internal network and the  >>thinwire DEC network.3 >>F >>The DS10L has two internal ethernet adapters. Is it possible to bind/ >>TCPIP to one adapter and DECnet to the other?C >  > K > Yes, easy. The configuration utilities will ask you which interface(s) to  > use. >  >  >>The network guys herenE >>really, REALLY want to eliminate the little ethernet hub (it annoyslF >>them) and I thought that I could accomplish that task if I could getA >>the DS10L connected to both the internal IP network and the DEC G >>thinwire directly (yes, I know I'll have to convert the TP connection,@ >>to BNC but I'll have the little hub available if I can get the  >>ethernet adapters configured). >>! >>Thanks in advance for any help!h >> >>Rusty Ferrell  >>rferrell@smcwv.com >  >  >    ------------------------------  + Date: Thu, 12 Jun 2003 11:27:08 +0000 (UTC)a+ From: david20@alpha2.mdx.ac.uk (David Webb)b, Subject: Re: Rookie question for the experts+ Message-ID: <bc9o2c$noc$1@aquila.mdx.ac.uk>   S In article <bc8247$vs9$1@news3.tilbu1.nb.home.nl>, Dirk Munk <munk@home.nl> writes:, >Rusty Ferrell wrote:  >> Greetings all!rI >> I'm a UNIX kind-of-guy that's been put into a VMS-kind-of-guy positiont >> so please be gentle with me!  >> lE >> Our existing system is an AlphaServer 2100 and it's being replacedeF >> with a DS10L AlphaServer. The 2100 is FDDI attached to our internalF >> TCP/IP based network. There are no other network connections to theD >> 2100. In order for the old DECserver's and devices to "talk" (via >> DECnet I assume)- >SP >No, 'classic' DECservers use MOP (Maintenance Operations Protocol) to download K >their software, and LAT (Local Area Transport) for their normal operation.- >-0 >Newer versions can use TFTP and Telnet as well. >    And also BOOTP.e    
 David Webb VMS and Unix team leader CCSS Middlesex University        Q >MOP and LAT are non-routable protocols and have nothing (!!!) to do with DECnet.  >0R >The only problem is that the MOP entries for the software for the DECservers are M >placed in the DECnet databases (Phase IV) or configuration files (Phase V /  = >OSI), so you need DECnet to be able to load your DECservers.* >d7 >> to the 2100, a small ethernet hub was installed that H >> acts as a media converter/bridge between the internal network and the >> thinwire DEC network. >> sG >> The DS10L has two internal ethernet adapters. Is it possible to binde1 >> TCPIP to one adapter and DECnet to the other? t >gO >Yes, easy. The configuration utilities will ask you which interface(s) to use.n >g >> The network guys hereF >> really, REALLY want to eliminate the little ethernet hub (it annoysG >> them) and I thought that I could accomplish that task if I could gettB >> the DS10L connected to both the internal IP network and the DECH >> thinwire directly (yes, I know I'll have to convert the TP connectionA >> to BNC but I'll have the little hub available if I can get theP! >> ethernet adapters configured).m >> "" >> Thanks in advance for any help! >> t >> Rusty Ferrell >> rferrell@smcwv.com  >o   ------------------------------    Date: 12 Jun 2003 07:56:08 -0500; From: koehler@eisner.nospam.encompasserve.org (Bob Koehler)l, Subject: Re: Rookie question for the experts3 Message-ID: <klfNFgmQ6PJT@eisner.encompasserve.org>s  h In article <9b72a38f.0306111611.7d717407@posting.google.com>, rferrell@smcwv.com (Rusty Ferrell) writes: > D > You know, after I wrote that I knew that it didn't sound right. WeC > have an old SNA gateway that an application uses to get chemistry&H > information from our mainframe. Would I be correct in assuming that IT > uses DECnet? >   G Gee, could be DECnet/SNA gateway.  DEC sold that for a while.  SomebodypC might have done an IP/SNA gateway.  I would assume DECnet until youo find out for sure.   > @ > The old DECservers (200's, 300's and 500's) are being used forD > terminals, printers and serial devices such as mass spectrometers,E > xray systems, atomic scales, etc. I probably should have been a bit & > more descriptive of the environment.  G IIRC the 200 and 300 can't do IP, the 500 might support it but probably B won't boot that way.  You have to have MOM/MOP and LAT to use this stuff.  9 > So, looks like my focus in the morning is to figure outdE > how to bind LAT to the "EWA0" port and get TSM running (I think the C > software failed to install since the network wasn't up during theH > software installation).   D I think LAT always wants to bind to the first ethernet it finds.  IfD that's not the one you want you can define lat$link before you start LAT.   ------------------------------    Date: 12 Jun 2003 07:57:27 -0500; From: koehler@eisner.nospam.encompasserve.org (Bob Koehler)m, Subject: Re: Rookie question for the experts3 Message-ID: <yJH6kuvYOdUC@eisner.encompasserve.org>O  b In article <YjQFa.8049$0c4.770361@news20.bellglobal.com>, "JustMe" <mc.moore@sympatico.ca> writes:K > Just make sure you start up DECnet on the required line (EWA0) before yousJ > start LAT.  My recollection from older (VAX) versions was that LAT wouldL > "take over" the adapter if started first, and then DECnet wouldn't be able > to start the line. > D > The SYSTARTUP_VMS.com template has the sequencing like that, iirc. > J > (btw, I think the IP equivalent of MOP is BOOTP not TFTP, but I could be' > mistaken -- wouldn't be the 1st time)   D    Generally IP boot functions require BOOTP or DHCP to establish an$    IP address and TFTP to get files.   ------------------------------  % Date: Thu, 12 Jun 2003 09:15:24 -0400	/ From: "Hank Vander Waal" <hvanderw@mansply.com>/, Subject: RE: Rookie question for the experts? Message-ID: <MDEGJFAOHGLNLHONIEBMEEEGCLAA.hvanderw@mansply.com>i    FWIW   decserver 300's do use IP   Hank   -----Original Message-----B From: Bob Koehler [mailto:koehler@eisner.nospam.encompasserve.org]% Sent: Thursday, June 12, 2003 8:56 AMu To: Info-VAX@Mvb.Saic.Coma, Subject: Re: Rookie question for the experts    = In article <9b72a38f.0306111611.7d717407@posting.google.com>,m* rferrell@smcwv.com (Rusty Ferrell) writes: >uD > You know, after I wrote that I knew that it didn't sound right. WeC > have an old SNA gateway that an application uses to get chemistryhH > information from our mainframe. Would I be correct in assuming that IT > uses DECnet? >e  G Gee, could be DECnet/SNA gateway.  DEC sold that for a while.  SomebodyrC might have done an IP/SNA gateway.  I would assume DECnet until youy find out for sure.   >U@ > The old DECservers (200's, 300's and 500's) are being used forD > terminals, printers and serial devices such as mass spectrometers,E > xray systems, atomic scales, etc. I probably should have been a bit & > more descriptive of the environment.  G IIRC the 200 and 300 can't do IP, the 500 might support it but probablyMB won't boot that way.  You have to have MOM/MOP and LAT to use this stuff.  9 > So, looks like my focus in the morning is to figure out E > how to bind LAT to the "EWA0" port and get TSM running (I think the C > software failed to install since the network wasn't up during them > software installation).   D I think LAT always wants to bind to the first ethernet it finds.  IfD that's not the one you want you can define lat$link before you start LAT.   ------------------------------  + Date: Thu, 12 Jun 2003 15:12:52 +0000 (UTC)t, From: Michael Moroney <moroney@TheWorld.com>, Subject: Re: Rookie question for the experts( Message-ID: <bca59k$1lq$2@pcls4.std.com>    Dirk Munk <munk@home.nl> writes:   >JustMe wrote:L >> Just make sure you start up DECnet on the required line (EWA0) before youK >> start LAT.  My recollection from older (VAX) versions was that LAT wouldeM >> "take over" the adapter if started first, and then DECnet wouldn't be ablem >> to start the line.c  O >Yep, when the interface is used for DECnet phase IV (compatible) traffic, the iO >physical address is changed to AA-00-04-00-aa-bb. This can only be done if no t >other protocol is active.  F No longer true as of V5.late or V6.0.  Protocols, when starting up canJ specify an options bit that states it will allow another protocol (DECnet)E to change the physical address out from underneath it.  LAT sets this F bit so you can start LAT before DECnet IV.  TCPIP does not so you must start it after DECnet IV.  r -- O -Miker   ------------------------------  % Date: Thu, 12 Jun 2003 12:54:20 -0400C" From: JF Mezei <jfmezei@istop.com>, Subject: Re: Rookie question for the experts) Message-ID: <3EE8B03B.2766F528@istop.com>n   Just a tought.  K If your network people are "forcing" you to get rid of DECNET, then perhapsiK you could arrange for all that work to be billed to the network people. AndgN they should also force the IBM mainframe to provide IP based solutiosn such as5 NFS or FTP to allow your vaxes to excahnge the files.P  M If you have 2 ethernets devides, then it shouldn't really be a problem to useoN one for the "proprietary" protocols such as LAT , MOP and DECNET, and then use the other one for TCPIP.   ------------------------------    Date: 12 Jun 2003 10:34:24 +0200C From: vaxinf@chclu.chemie.uni-konstanz.de (Eberhard Heuser-Hofmann) $ Subject: Re: TK70 tape drive problem- Message-ID: <3ee83b10$1@news.uni-konstanz.de>e  F In article <3EE79F81.A6E2359C@istop.com>, JF Mezei <jfmezei@istop.com> writes:n |>"Jerome H. Fine" wrote: ; |>> If you don't have a head cleaner, then I have found ther; |>> best solution is a Q-tip (wood stick with cotton ends - * |>> a bit bigger than a wood match stick)! |>I |>Well, I didn't want to suggest this since I had been told the heads are[ |>quiteOF |>fragile. I use lens cleaning paper myself with a single drop of lens
 |>cleaningG |>fluid or alchool. I found that Qtips may leave little strands (unlesse |>youe |>have |>foam qtips). |>  < As chemist is must state this: DON'T USE ALKOHOL (Ethanol), 8 because Ethanol contains water, that destroys the head.    Isopropanol is OK.   eberhard   ------------------------------    Date: 12 Jun 2003 07:47:13 -0500; From: koehler@eisner.nospam.encompasserve.org (Bob Koehler)h$ Subject: Re: TK70 tape drive problem3 Message-ID: <tfxUdG8bWJAB@eisner.encompasserve.org>a  o In article <3EE7811D.29A63163@b9rvnospamcompsys.to>, "Jerome H. Fine" <jhfineb9rv@b9rvnospamcompsys.to> writes:i5 > Since you already seem to have access to the drive,t6 > completely removing it from the system should not be; > too difficult.  There is a latch in the front which holds 7 > the drive in place, just push the latch down.  I haveM8 > never found there to be a problem doing this on a live5 > system, but I use the TK70 on a PDP-11 under RT-11,y9 > so decide what is best for you.  PLUS, I did once shorte4 > the power supply when I plugged the power back in,9 > so connecting the power is NOT trivial.  I suggest thatg7 > IF you are working with a live system, disconnect they9 > power first and connect the power last relative the thei< > signal cable.  If the cable does not have a sticker saying  > THIS SIDE  UP, mark the cable.  F    I never found it necessary to disconnect the power when sliding theC TK70 out of my VAX 3300 and operating on it's mechanical inards for D stuck tapes and the like.  Just didn't let any user think they could@ actually use it while I was holding it (system was up and user's/ generally didn't know the front cover was off).t   ------------------------------  + Date: Thu, 12 Jun 2003 09:45:48 +0000 (UTC)rA From: Adrian Stapley <adrian@semleystation.spammed-to-hell.co.uk>c. Subject: Re: TK70 tape drive problem - Updatye0 Message-ID: <bc9i4b$gr4$1@sparta.btinternet.com>   Adrian Stapley wrote:n  J Thanks to all the replies - I have tried cleaning the heads, and also the 3 small led? sensors either side of the rear capstan.i  K Error count - does not get reset after a reboot, I have already tried this  K and am currently looking for a utility I remember seeing on an ftp site to .# reset the error count for a device.s  I The symptoms I am seeing would suggest to me that it is a head failure -  K after closing the handle on tape (either a fresh one for an init or a used i. one for a mount), the following actions occur:  K 1) The leader engages the tape and it is hauled through the capstans until -3 the hole in the tape is almost on the take up spool   I 2) Nothing else happens until an init or mount command, when the tape is II run back slightly until the hole has reached the heads, then the tape is u8 forwarded until the hole is almost on the take up spool.  $ 3) The mount error message is given.  , From 1 and 2 I assume that the LED's are OK.  J Trying to manually wind the tape by turning the take up spool gives all 3 H led's flashing (so movement sensors are also OK ?). Pressing the unload 9 button will rewind the tape and clear the flashing LED's.   I Of course, I am going to be totally wrong, and find it is the controller u6 board or error count, as has already been suggested :)   Adrian    (Why did I stop using my QL's ?)   ------------------------------  % Date: Thu, 12 Jun 2003 12:47:16 -0400v" From: JF Mezei <jfmezei@istop.com>. Subject: Re: TK70 tape drive problem - Updatye) Message-ID: <3EE8AE94.30B2715B@istop.com>r   Adrian Stapley wrote:eJ > The symptoms I am seeing would suggest to me that it is a head failure -L > after closing the handle on tape (either a fresh one for an init or a used0 > one for a mount), the following actions occur:  J Of you could have a bad tape. I've had TK50s that just didn't want to workK anymore, but putting a different tape in worked. (perhaps a question of therG heads being at questionable "cleanliness" and working on good tapes bute2 failing below a certain quality of tape (old etc).   ------------------------------    Date: 12 Jun 2003 12:08:54 -0500B From: clubley@remove_me.eisner.decus.org-Earth.UFP (Simon Clubley). Subject: Re: TK70 tape drive problem - Updatye3 Message-ID: <$1siYWkgqcnA@eisner.encompasserve.org>>  t In article <bc9i4b$gr4$1@sparta.btinternet.com>, Adrian Stapley <adrian@semleystation.spammed-to-hell.co.uk> writes: > M > Error count - does not get reset after a reboot, I have already tried this  M > and am currently looking for a utility I remember seeing on an ftp site to 5% > reset the error count for a device.y >   H The error count most certainly does get reset on a reboot which makes meO wonder if you seeing errors at system startup during controller initialisation.n  < Are the errors logged against the device or the controller ?  J After the system reboot, but _before_ accessing the tape drive in any way,K is the error count greater than zero (your above statement suggests that ituE so), and if so, what are the last couple of errors in the error log ?m  I Post those errors here. Those people familiar with this hardware might be2  able to tell you what is broken.   Simon.   -- jB Simon Clubley, clubley@remove_me.eisner.decus.org-Earth.UFP       L VMS advocate: One who makes a Mac advocate look like a beginner at advocacy.   ------------------------------  % Date: Thu, 12 Jun 2003 10:49:25 -0500p( From: brandon@dalsemi.com (John Brandon) Subject: Tracking VMS SMTP mail 1 Message-ID: <03061210492538@dscis6-0.dalsemi.com>t  J I am looking to track outbound VMS Mail (using SMTP).  I want to track theJ header information and have the ability to (turn on/off by user) track the message of the mail.  + I was looking through Google for solutions.e  : 1)  Define the logical UCX$SMTP_LOG_LEVEL was one of them.  M What sort of loggin does this do?  I do not want to find out I am running outc of disk space or such.  Q 2)  Use the MX software provided by MadGoat.  I believe this would imply that theo) VMS server is now a Mail eXchange server.   2 What impact does this have on my normal VMS Mail?           H Is there anyone out there tracking their SMTP mail from VMS?  How do you
 acomplish it?      J*o*h*n B*r*a*n*d*o*n  VMS Systems Administratorm* firstname.lastname.spam.me.not@dalsemi.com   ------------------------------  % Date: Thu, 12 Jun 2003 13:18:10 -0400r" From: JF Mezei <jfmezei@istop.com># Subject: Re: Tracking VMS SMTP mail ) Message-ID: <3EE8B5D0.C85CFFA0@istop.com>    John Brandon wrote:e > L > I am looking to track outbound VMS Mail (using SMTP).  I want to track theL > header information and have the ability to (turn on/off by user) track the > message of the mail.  L What version of software are you using ? If using version 5.0 and above, you can look at:  : http://vaxination.dyndns.orgf/vms/tcpip_smtp_systartup.com  L which gives you a template to define the various logicals, in particular the logging logicals.-4 Note that after changing the logicals, you have to :   TCPIP> STOP MAIL TCPIP> START MAIL_  + to get teh logicals to enter into function.    Here is a sample output:  "TCPIP$SMTP_LOG_LEVEL" = "2"C    "TCPIP$SMTP_NOSEY" = "1"c ----- / Printing debug_level 2, Domains and recipients:h Domain: istop.comA'    Recipient address: JFMEZEI@ISTOP.COMe    Domain part:       istop.coma    Local part:        JFMEZEIs-    Address Status:    Done, delivered. (Sent).  ) Printing debug_level 2, relevant headers:m" Return-Path: jfmezei@chocolate.com% Date: Sat, 17 May 2003 14:08:45 -0400 / Message-Id: <03051714084542@bike.vaxination.ca>l From: jfmezei@chocolate.coms To: JFMEZEI@pastry.com+ Subject: test message while smtp is stoppedg! X-VMS-To: SMTP%JFMEZEI@PASTRY.COMh -------o  " Now, with the SYMB TRACE, you get:P 19-APR-2003 02:06:23.78: next open file _$2$DKA200:[JFMEZEI.MAILBOX]030419020619$ 30_JFMEZEI-22000098.TCPIP_BIKE;1(72)- recv buf=220 smtp.isp.com ESMTP Postfix\0d\0a & send buf=HELO bike.chocolate.com\0d\0a recv buf=250 smtp.isp.com\0d\0au0 send buf=MAIL FROM:<jfmezei@chocolate.com>\0d\0a recv buf=250 Ok\0d\0at- send buf=RCPT TO:<ghislaing@pastry.com>\0d\0at recv buf=250 Ok\0d\0aa send buf=DATA\0d\0an2 recv buf=354 End data with <CR><LF>.<CR><LF>\0d\0a4 send buf=Date: Sat, 19 Apr 2003 02:06:19 -0400\0d\0a> send buf=Message-Id: <03041902061930@bike.vaxination.ca>\0d\0aB send buf=From: jfmezei@chocolate.com (Jean-Fran\e7ois Mezei)\0d\0a' send buf=To: ghislaing@pastry.com\0d\0aP6 send buf=Subject: Re: Choose David or Golliath ?\0d\0a4 send buf=X-VMS-To: smtp%"ghislaing@pastry.com"\0d\0aK send buf=X-VMS-True-From: jfmezei@chocolate.com (Jean-Fran\e7ois Mezei)\0d\m 0a send buf=\0d\0a P send buf=>an FTP server and I can setup any other type of server if I want to. I n\0d\0aeP send buf=>addition, I had a requirement to use a non-PPPoE connection and a fixe
 d IP\0d\0aP send buf=>address for my VPN server. I use their residential 3.5M service.\0d\0a ...  send buf=\0d\0ah send buf=.\0d\0a+ recv buf=250 Ok: queued as 1627136B1B\0d\0ae send buf=QUIT\0d\0a0 recv buf=221 Bye\0d\0a      O And after the actual trace, you get the normal logging as in the first example.e  N You cannot enable this on a per-user basis. It is all or nothing. However, theJ SMTP server does have a snapshot buffer which can stop a log in memory andA only write it to disk if there was a problem sending the message.O   ------------------------------   Date: 12 Jun 2003 08:49 CDT'' From: carl@gerg.tamu.edu (Carl Perkins)  Subject: Re: uptime-alikeg- Message-ID: <12JUN200308492685@gerg.tamu.edu>   / young_r@encompasserve.org (Rob Young) writes...a] }In article <03061110464556@dscis6-0.dalsemi.com>, brandon@dalsemi.com (John Brandon) writes:-O }>> I need a program functionally like the UN*X "uptime" command.  Seems to me d }>> that I could write it. }>> O }>> Getting the BOOTTIME and therefore the up-time of the machine is easy.  Thes }>> partR }>> that has me stumped is how I'm going to figure out the rest, meaning the countP }>> of pending COM[O] processes.  The most recent internals data I've got is for* }>> VAX/VMS 3.6, so I'm kinda at sea here. }>>  }>> Suggestions? }> sQ }> For some of us lacking in UN*X could you show us sample output of what you are8 }> wanting?  Thanks! }> e } % }nodename_is_here:/home/youngr>uptimeoI }  11:51AM   up 4 days,  20:39,  4 users,  load average: 0.36, 0.61, 0.74s }  }				Rob  G If you are running Multinet, then it provides some load information viar
 a device,  (LNM$SYSTEM_TABLE)  !   "$$VMS_LOAD_AVERAGE" = "_LAV0:"e  9 A read on this device gives you data formatted like this:. ;+ ---) ; Local Data for the Load Average Watchere ;- ---B M1:     .FLOAT  0.0                             ;1  Min. Load Ave.B M5:     .FLOAT  0.0                             ;5  Min. Load Ave.B M15:    .FLOAT  0.0                             ;15 Min. Load Ave.  F P1:     .FLOAT  0.0                             ;1  Min. Priority Ave.F P5:     .FLOAT  0.0                             ;5  Min. Priority Ave.F P15:    .FLOAT  0.0                             ;15 Min. Priority Ave.  K Q1:     .FLOAT  0.0                             ;1  Min. Ave disk queue lenEK Q5:     .FLOAT  0.0                             ;5  Min. Ave disk queue len<K Q15:    .FLOAT  0.0                             ;15 Min. Ave disk queue lenn  * Here's a Fortran program to show the info:           OPTIONS /EXTEND_SOURCE       PROGRAM load         IMPLICIT NONEA  ,       REAL              Load1, Load5, Load15/       REAL              Block1, Block5, Block15 >       REAL              DQueLength1, DQueLength5, DQueLength15       CHARACTER*23      Date$       Parameter LoadDevice = 'LAV0:'C c  ! Site-specific: This is the load average pseudo-device.  If notlF c  available, omit this section.  Or leave it and it will still be OK.       Open(Unit=13,h      1  File=LoadDevice,      2  Type='NEW',       3  RecordSize=36)       Call LIB$DATE_TIME (Date)oH       Read(13,fmt='(9A4)') Load1, Load5, Load15,Block1, Block5, Block15,.      1  DqueLength1, DQueLength5, DQueLength15       Close(Unit=13)-       write (6,100),Date, Load1,Load5,Load15,c       1  Block1, Block5, Block15,.      1  Dquelength1, Dquelength5, Dquelength159 100   FORMAT ('0RECENT LOAD AVERAGE STATISTICS at ',A23,/c7      1  '0 Period (minutes)',t37,'1',t47,'5',t56,'15',/.1      1  ' ',t35,'=====',t45,'=====',t55,'=====',/h-      1  ' Jobs waiting for CPU:',t30,3f10.2,/"<      1  ' Lowest Executing Priority:          ',t30,3f10.2,/*      1  ' Disk queue length: ',t30,3f10.2)       Call EXITe	       End   F (Originally posted by Ed Wilts, it's probably an excerpt from a larger	 program.)i  8 This does nothing for you if you are not using Multinet.   --- Carl   ------------------------------   Date: 12 Jun 03 08:12:21 PST From: mckinneyj@cpva.saic.comH Subject: Re: uptime-alikei( Message-ID: <X$7O$k07lQD9@cpva.saic.com>  - In article <12JUN200308492685@gerg.tamu.edu>,d*  carl@gerg.tamu.edu (Carl Perkins) writes:1 > young_r@encompasserve.org (Rob Young) writes...e_ > }In article <03061110464556@dscis6-0.dalsemi.com>, brandon@dalsemi.com (John Brandon) writes:tQ > }>> I need a program functionally like the UN*X "uptime" command.  Seems to me o > }>> that I could write it. > }>> Q > }>> Getting the BOOTTIME and therefore the up-time of the machine is easy.  The 
 > }>> partT > }>> that has me stumped is how I'm going to figure out the rest, meaning the countR > }>> of pending COM[O] processes.  The most recent internals data I've got is for, > }>> VAX/VMS 3.6, so I'm kinda at sea here. > }>>  > }>> Suggestions? > }> -S > }> For some of us lacking in UN*X could you show us sample output of what you aree > }> wanting?  Thanks! > }> : > } ' > }nodename_is_here:/home/youngr>uptime$K > }  11:51AM   up 4 days,  20:39,  4 users,  load average: 0.36, 0.61, 0.74n > } 
 > }				Rob > I > If you are running Multinet, then it provides some load information viao > a device,  > (LNM$SYSTEM_TABLE) > # >   "$$VMS_LOAD_AVERAGE" = "_LAV0:"i > ; > A read on this device gives you data formatted like this:h  0 [... data structures and code sample snipped...]  : > This does nothing for you if you are not using Multinet. > 
 > --- Carl  E But the LAVDRIVER was written by David Kashtan (of TGV/SRI fame) if I6F recall correctly and the source was made public years ago. It predatedF Alpha, but may have been ported by someone other than Process. It willE demonstrate how to walk the state queues that you're interested in. Ah- google search might turn up an Alpha version..   ------------------------------  % Date: Thu, 12 Jun 2003 08:00:46 +0200s+ From: "IdrEASY" <idreasyMAKNI_BAGRU@vip.hr>o0 Subject: Re: Wininet.dll (FindFirstFile) and VMS0 Message-ID: <bc94nr$4af$1@garrison.globalnet.hr>  G "Simon Clubley" <clubley@remove_me.eisner.decus.org-Earth.UFP> wrote inh5 message news:GrFVJuAAHznR@eisner.encompasserve.org...s< > In article <bc44c2$j2i$1@garrison.globalnet.hr>, "IdrEASY"# <idreasyMAKNI_BAGRU@vip.hr> writes:u< > > How to get directory list of VMS account in VisualBasic? > >hD > > When filename is very long, FindFirstFile function breaking file informationiI > > in two lines (filename and file attributes), so next call of function  returnK > > file attributes of previous file instead of next filename informations.y > >m >eI > You need to install a FTP server on the VMS system that emits directory I > listings in Unix style format, or switch your existing FTP server to do  > the same if it is possible.l >aK > This is required because the API you are using assumes that the directoryr > listing is in Unix format. >tI > If you provide details of which TCP/IP stack you are using, people here * > will be able to tell you how to proceed. >  Thank you Simon!H I'm using Digital TCP/IP Services for OpenVMS Alpha Version V4.2 - ECO 2 (UCX)u       ---e& Outgoing mail is certified Virus Free.: Checked by AVG anti-virus system (http://www.grisoft.com).A Version: 6.0.489 / Virus Database: 288 - Release Date: 10.06.2003u   ------------------------------    Date: 12 Jun 2003 06:55:56 -0500B From: clubley@remove_me.eisner.decus.org-Earth.UFP (Simon Clubley)0 Subject: Re: Wininet.dll (FindFirstFile) and VMS3 Message-ID: <CKQNIzrFZ89F@eisner.encompasserve.org>C  ^ In article <bc94nr$4af$1@garrison.globalnet.hr>, "IdrEASY" <idreasyMAKNI_BAGRU@vip.hr> writes:I > "Simon Clubley" <clubley@remove_me.eisner.decus.org-Earth.UFP> wrote in 7 > message news:GrFVJuAAHznR@eisner.encompasserve.org...u= >> In article <bc44c2$j2i$1@garrison.globalnet.hr>, "IdrEASY"w% > <idreasyMAKNI_BAGRU@vip.hr> writes:s= >> > How to get directory list of VMS account in VisualBasic?d >> >E >> > When filename is very long, FindFirstFile function breaking filen
 > informationdJ >> > in two lines (filename and file attributes), so next call of function > returnL >> > file attributes of previous file instead of next filename informations. >> > >>J >> You need to install a FTP server on the VMS system that emits directoryJ >> listings in Unix style format, or switch your existing FTP server to do >> the same if it is possible. >>L >> This is required because the API you are using assumes that the directory >> listing is in Unix format.  >>J >> If you provide details of which TCP/IP stack you are using, people here+ >> will be able to tell you how to proceed.f >> > Thank you Simon!J > I'm using Digital TCP/IP Services for OpenVMS Alpha Version V4.2 - ECO 2 > (UCX)b >   I In that case, the second option I gave you, reconfigure your existing FTPtG server to emit Unix format directory listings, is not available to you.e  J You will have to install a new FTP server and configure it to work in Unix5 mode. The one that I used was Hunter Goatley's HGFTP.l  G Hunter's main VMS page is at: http://www.process.com/openvms/index.htmlf  ( The list of packages for download is at:  . 	http://vms.process.com/fileserv-software.html  G (HGFTP needs NETLIB and I cannot remember if NETLIB came with the HGFTPr@ package, or if you have to download NETLIB from the above page).  J Note that I have had problems browsing sub-directories from some MicrosoftK applications, but otherwise I have found the HGFTP server to be fine for my- needs.  M You can install the HGFTP server without having to also run the HGFTP client.a  H Also be aware of potential differences in the case of filenames returnedL via an MGET command. the HGFTP server returns ODS-2 filenames in lower case.8 (I'm not sure about ODS-5 as I don't use ODS-5 volumes).   Simon.   -- mB Simon Clubley, clubley@remove_me.eisner.decus.org-Earth.UFP       L VMS advocate: One who makes a Mac advocate look like a beginner at advocacy.   ------------------------------  % Date: Thu, 12 Jun 2003 14:28:51 +0200e+ From: "IdrEASY" <idreasyMAKNI_BAGRU@vip.hr> 0 Subject: Re: Wininet.dll (FindFirstFile) and VMS0 Message-ID: <bc9rfh$cfe$1@garrison.globalnet.hr>  G "Simon Clubley" <clubley@remove_me.eisner.decus.org-Earth.UFP> wrote in 5 message news:CKQNIzrFZ89F@eisner.encompasserve.org... < > In article <bc94nr$4af$1@garrison.globalnet.hr>, "IdrEASY"# <idreasyMAKNI_BAGRU@vip.hr> writes:iK > > "Simon Clubley" <clubley@remove_me.eisner.decus.org-Earth.UFP> wrote in}9 > > message news:GrFVJuAAHznR@eisner.encompasserve.org...e? > >> In article <bc44c2$j2i$1@garrison.globalnet.hr>, "IdrEASY".' > > <idreasyMAKNI_BAGRU@vip.hr> writes:d? > >> > How to get directory list of VMS account in VisualBasic?  > >> >G > >> > When filename is very long, FindFirstFile function breaking file  > > informationmL > >> > in two lines (filename and file attributes), so next call of function
 > > return@ > >> > file attributes of previous file instead of next filename
 informations.a > >> > > >>L > >> You need to install a FTP server on the VMS system that emits directoryL > >> listings in Unix style format, or switch your existing FTP server to do  > >> the same if it is possible. > >>D > >> This is required because the API you are using assumes that the	 directoryw > >> listing is in Unix format.  > >>L > >> If you provide details of which TCP/IP stack you are using, people here- > >> will be able to tell you how to proceed.  > >> > > Thank you Simon!L > > I'm using Digital TCP/IP Services for OpenVMS Alpha Version V4.2 - ECO 2	 > > (UCX)a > >v >eK > In that case, the second option I gave you, reconfigure your existing FTPdI > server to emit Unix format directory listings, is not available to you.u >0L > You will have to install a new FTP server and configure it to work in Unix7 > mode. The one that I used was Hunter Goatley's HGFTP.1 >.I > Hunter's main VMS page is at: http://www.process.com/openvms/index.htmlu >d* > The list of packages for download is at: >l/ > http://vms.process.com/fileserv-software.htmlu > I > (HGFTP needs NETLIB and I cannot remember if NETLIB came with the HGFTPeB > package, or if you have to download NETLIB from the above page). >aL > Note that I have had problems browsing sub-directories from some MicrosoftJ > applications, but otherwise I have found the HGFTP server to be fine for my > needs. > G > You can install the HGFTP server without having to also run the HGFTPe client.  >MJ > Also be aware of potential differences in the case of filenames returnedH > via an MGET command. the HGFTP server returns ODS-2 filenames in lower case. : > (I'm not sure about ODS-5 as I don't use ODS-5 volumes). >a > Simon. >  > --= > Simon Clubley, clubley@remove_me.eisner.decus.org-Earth.UFPmD > VMS advocate: One who makes a Mac advocate look like a beginner at	 advocacy.r  B Thank you again! I am sure that I will be in situation to use your% informations somewhere in the future.   I I think that would be easiest for me to develop my own parser o responded H string from VMS, because I have no administrator's rights to do what you	 describe.   I In this case I must detect when "filename" starts with number and contain-L slash, then add this to previous filename and after that proceed like normal responding string.  I For example, I was tried connections on my VMS host with many FTP clientso and CuteFTP works perfectly. Maybe author done same thing.t   Thank you once again.  Bye!     ---r& Outgoing mail is certified Virus Free.: Checked by AVG anti-virus system (http://www.grisoft.com).A Version: 6.0.489 / Virus Database: 288 - Release Date: 10.06.2003    ------------------------------  % Date: Thu, 12 Jun 2003 12:37:39 -0400.2 From: John Eisenschmidt <jweisen@eisenschmidt.org>> Subject: [OT] SCO may revoke IBM's license to SysV tomorrow...5 Message-ID: <20030612163739.GJ30359@eisenschmidt.org>m  C I know the last thing we need is another off topic discussion here, 9 and I really thought twice about posting it, but with thefE Oracle/Peoplesoft/Edwards soap opera going on I'm just giddy. Anyway:C  M http://asia.reuters.com/newsArticle.jhtml?type=technologyNews&storyID=2915659a  > "IBM, which had licensed Unix code in order to develop its ownE Unix-based system called AIX, is also one of the biggest champions ofpF Linux, which it supports in order to sell its hardware and services toC corporations. SCO claims that IBM transferred some AIX code over to  Linux.  F 'If we don't have a resolution by midnight on Friday the 13th, the AIXC world will be a different place,' SCO President and Chief Executive  Darl McBride told Reuters.  C 'We've basically mapped out what we will do. People will be runninglC AIX without a valid license,' said McBride, who offered no specifict' details on what action SCO would take."t   -- a/ John W. Eisenschmidt (jweisen@eisenschmidt.org)t.   http://www.eisenschmidt.org/jweisen/pgp.html  M "binary four n. :: [Usenet] The finger, in the sense of `digitus impudicus'. -M  This comes from an analogy between binary and the hand, i.e. 1=00001=thumb, rL  2=00010=index finger, 3=00011=index and thumb, 4=00100."   -The Jargon File   ------------------------------   End of INFO-VAX 2003.324 ************************