1 INFO-VAX	Wed, 17 May 2006	Volume 2006 : Issue 273       Contents: Re: Cross-architecture booting Re: Cross-architecture booting Re: Hardware Question  Re: Hardware Question  Re: Hobbyist on Itanium  Re: Hobbyist on Itanium  Re: Hobbyist on Itanium % Re: MCHKMCHK Bugcheck in VMS VAX V6.2 % Re: MCHKMCHK Bugcheck in VMS VAX V6.2 % Re: MCHKMCHK Bugcheck in VMS VAX V6.2 . Re: Multiple DECW$Sessions on one X terminal ?1 Re: OT: race for 8086s continues: 4 cores for AMD 1 Re: OT: race for 8086s continues: 4 cores for AMD 1 Re: OT: race for 8086s continues: 4 cores for AMD  Ping in web server command file 7 Re: problem using mail from php-script in apache server F Re: problem using mail from php-script in apache server [cf: Multinet] Re: Results of my straw poll.  Re: Results of my straw poll.  Re: Results of my straw poll.  Re: Results of my straw poll.  Re: Results of my straw poll.  Re: Results of my straw poll.  Re: Results of my straw poll.  Re: Results of my straw poll.  Re: Results of my straw poll.  Re: Results of my straw poll.  Re: Results of my straw poll.  Re: Results of my straw poll.  Re: Results of my straw poll.  Re: Results of my straw poll.  Re: Results of my straw poll.  Re: Results of my straw poll.  Re: Results of my straw poll.  Re: Results of my straw poll.  Re: Results of my straw poll.  SORT, LEF, SYS$PUT and COBOL, speeding up LAVC with switch instead of hub?0 Re: speeding up LAVC with switch instead of hub?0 Re: speeding up LAVC with switch instead of hub?0 Re: speeding up LAVC with switch instead of hub?0 Re: speeding up LAVC with switch instead of hub?0 Re: speeding up LAVC with switch instead of hub? Thundering Hurd pacing HP gains # Re: Thundering Hurd pacing HP gains / Re: what should Bad-Clients: in SMTP.CONFIG do? / Re: what should Bad-Clients: in SMTP.CONFIG do?   F ----------------------------------------------------------------------    Date: 17 May 2006 01:35:21 -0700# From: "H Vlems" <hvlems@freenet.de> ' Subject: Re: Cross-architecture booting C Message-ID: <1147854921.800300.291330@g10g2000cwb.googlegroups.com>   B Don't see why that is tricky: use SIMH to build the desired systemG disk. Next, perform a BACKUP/IMAGE/IGNORE=INTERLOCK to a backup saveset D and copy that saveset to the  target boot host. On that host restoreD the saveset to a disk.  Set up LANCP or NCP to boot the diskless VAX from. E I did exactly the opposite thing: boot an Alpha 2100 from a VAX 8550.  Piece of cake. Hans   ------------------------------  # Date: Wed, 17 May 2006 13:35:22 GMT , From: Hoff Hoffman <hoff-remove-this@hp.com>' Subject: Re: Cross-architecture booting 0 Message-ID: <uEFag.603$kD1.182@news.cpqcorp.net>   mpocciot@gmail.com wrote:   @ > [discussion of installing  OpenVMS VAX on a disk on a VAX box,? > and swapping the disk back over to the target VAX expurgated]  > That should fix it, right?  7    For some definitions of "should" and "it", yes.  :-)   + > I'll also look into the B set suggestion.   C    The classic OpenVMS VAX installation uses standalone BACKUP, and C that requires direct access to the target disk.  (This is the basis + for the core requirement for local access.)   E    Installation in a cluster environment requires cluster membership, : and until OpenVMS is installed you don't really have that.  D    I'm not at all certain performing a BACKUP/IMAGE of VMSxxx.B ontoD a target disk using an OpenVMS Alpha box, and a subsequent satelliteA boot of the VAX (for the rest of the upgrade) will work, in other > words.  But since you are not in a production environment, any> failures here won't have a particularly big effect, obviously.  A    You could upgrade to a current version of OpenVMS Alpha (circa 9 V8.2, or the SDK 8.3 kit) and use the embedded InfoServer * capabilities to ISL boot the VAX, as well.   ------------------------------    Date: 17 May 2006 08:56:57 -0500' From: jim@info2.uah.edu (Jim McCullars)  Subject: Re: Hardware Question( Message-ID: <e4fa39$fvg$1@info2.uah.edu>  : johnhreinhardt@yahoo.com (johnhreinhardt@yahoo.com) wrote:   : Jim,I :    Another decent place to try questions like this is on the HP site at D : the IT Resource Center.  They have VMS and Storage related forums.I : Click this link http://www1.itrc.hp.com then click on the "Fourms" link  : about halfway down the page.      OK, thanks for the heads-up.   H : retired) called a Modular Data Router.  This lets me present the tapesF : to my VMS systems and they work pretty much as if the libraries wereA : directly connected to the machines.  The latest version of this E : hardware is called a Network Data Router (I think) and I beleive it E : works similarly.  It is connected to the SAN switch rather than the   0    Yeah, that's exactly what I am looking to do.   Jim    ------------------------------    Date: 17 May 2006 09:03:40 -0500' From: jim@info2.uah.edu (Jim McCullars)  Subject: Re: Hardware Question( Message-ID: <e4fafs$g3a$1@info2.uah.edu>  : johnhreinhardt@yahoo.com (johnhreinhardt@yahoo.com) wrote:  I :    Another decent place to try questions like this is on the HP site at D : the IT Resource Center.  They have VMS and Storage related forums.  K    Well, guess what?  I went there and saw a thread called "Plug MSL5026 on % a SAN with VMS boxes".  Thanks again!   
 Jim McCullars    ------------------------------  # Date: Wed, 17 May 2006 13:24:23 GMT , From: Hoff Hoffman <hoff-remove-this@hp.com>  Subject: Re: Hobbyist on Itanium0 Message-ID: <buFag.602$YH1.581@news.cpqcorp.net>   healyzh@aracnet.com wrote:/ > Hoff Hoffman <hoff-remove-this@hp.com> wrote:   N > I've played with PHP on my PWS433au and it seems just fine, same with Perl. M > I don't remember even seeing Ruby.  What I'm interested in on the XP1000 is N > Python performance.  I was interested in using it for a project, the problem> > is "Hello, World" takes 10 seconds to run on my PWS433au :^(  ,    A Ruby port is on the Freeware V7 distro.  L > How can you tell which ones it will work on?  I'm in the market for one ofJ > the 667Mhz XP1000's, as after 6+ years, I think it is time to retire the( > PWS433au and move to something faster.  F    There was a discussion here of a Plextor series DVD recorder drive G that did work and that did boot.  The PX-716A series.  (I've not tried  E booting with one, however.)  I know that various of the HP DVDwriter  H drives (which are not supported by OpenVMS), including the 740i series,  don't.  I    The "fun" here with any of these sorts of questions, however, is that  H even a particular series of drives can change its firmware and even its H core hardware, and that can toggle the "it works" state either way.  By D the time you get a configuration together, you might get completely % different results than somebody else.   G    Given the storage bays available on an XP1000, you can also load an  G older IDE/ATAPI CD-ROM drive or a SCSI CD-ROM drive should you need to  ? bootstrap, if whatever drive you select doesn't happen to work.   D    Most DVD drives seem to last roughly six months in the market in I recent times, give or take, before the product line is either overhauled  H (euphemistically known as being refreshed) or until the product line is ; replaced.  Approaching the shelf-lifetime of a fruit fly...   ( > Also, will HVD SCSI work in an XP1000?  H    I've stuffed a few ancient (and specifically also flavours supported ; by OpenVMS) SCSI controllers into the box.  They've worked.   H    Key here is the controller you choose to use.  SCSI should negotiate B mixtures of newer and older.  But if I can banish the older disks I entirely, or at least onto an older SCSI controller, I'll tend that way.  F   This avoids having a newer controller and a newer widget hobbled by   the presence of an older widget.  C    Will an arbitrary (and ancient)SCSI disk work on the integrated  H controller?  I don't know.  The embedded controller is UltraSCSI on the 2 EV6 variants, and Wide Ultra on the EV67 variants.  E    I've been upgrading to (ancient) StorageWorks (Midnight-coloured)  < replacement shelves for the FWSE (and ancient) StorageWorks C (Lichen-coloured) hardware previously used -- the bulk of the disk  I storage is external, so I don't have to consider the thermals inside the  F older AlphaStation and AlphaServer series boxes.  (And I've basically < eliminated all of the StorageWorks Lichen-coloured shelves.)   ------------------------------  % Date: Wed, 17 May 2006 07:03:20 -0700 # From: "Tom Linden" <tom@kednos.com>   Subject: Re: Hobbyist on Itanium) Message-ID: <op.s9o7zut8zgicya@hyrrokkin>   L On Wed, 17 May 2006 06:24:23 -0700, Hoff Hoffman <hoff-remove-this@hp.com>   wrote:  H >   There was a discussion here of a Plextor series DVD recorder drive  J > that did work and that did boot.  The PX-716A series.  (I've not tried   > booting with one, however.) C >   I know that various of the HP DVDwriter drives (which are not   : > supported by OpenVMS), including the 740i series, don't. > $ I can confirm both those assertions.   ------------------------------  % Date: Wed, 17 May 2006 11:21:46 -0400 3 From: "Richard B. Gilbert" <rgilbert88@comcast.net>   Subject: Re: Hobbyist on Itanium0 Message-ID: <0eqdnZq3UeEWovbZRVn-tw@comcast.com>   Bill Gunshannon wrote:  . > In article <4cs0nuF17ajsiU1@individual.net>,- > 	bill@cs.uofs.edu (Bill Gunshannon) writes:  > 1 >>In article <jt3ag.487$cC.117@news.cpqcorp.net>, 1 >>	Hoff Hoffman <hoff-remove-this@hp.com> writes:  >> >>>Bill Gunshannon wrote:  >>>  >>> K >>>>I would imagine those of us who are currently getting by on a VAX could I >>>>work with something a little less.  Just out of curiosity, can anyone 5 >>>>tell me how an XP1000 compares to a VAX 7000-640?  >>> K >>>   Very nicely, in my experience.  I can set the AlphaStation XP1000 on  J >>>my desk, and it hasn't collapsed to the floor yet.  I can also connect K >>>it to an office-grade UPS, too.  Well, without the UPS starting to glow.  >>> K >>>   One of the folks in the community has the AlphaStation XP1000 posted  K >>>at circa 260 VUPs, if that helps.  (That's likely an unofficial number;  I >>>AFAIK DIGITAL/Compaq/HP hasn't ever posted VUPs numbers more an a few  H >>>(if any?) VUPs ratings for systems other than VAX systems.)  The VAX A >>>7000 model 640 was officially rated somewhere around 125 VUPs.  >>> B >>>http://h18002.www1.hp.com/alphaserver/performance/vups_297.htmlB >>>http://h18002.www1.hp.com/alphaserver/performance/perf_tps.html >>> J >>>   In the latter chart, the VAX 7000 model 640 is 590 units, while the K >>>slowest EV6-class box I can find (an EV6 at 466 MHz) is well over 2,000  G >>>units -- I don't see the AlphaStation (it's a workstation and not a  D >>>server, and the workstations tended not to see the same sorts of H >>>performance testing sequences performed and published) listed in the ( >>>server performance comparison charts. >>> 6 >>>   Most any low-end Integrity server is faster yet. >>J >>Thanks Hoff.  That answers my qyestion.  I am hoping to acquire a coupleJ >>of XP1000's and if I do will replace the VAX with at least one and maybeI >>a pair of them, clustered.  I just wanted to know what the impact on my H >>users was likely to be.  Looks like a performance improvement for whatF >>they do now.  I guess the next wuestion will be how well it performs
 >>doing Java.  >  > K > Something else just hit me after I made that posting.  What am I going to I > do with three of those huge VAXen?  :-)  They make my 11/44's look like  > MicroVAX.  >  > bill >     G When faced with a similar problem, I offered the boxes here, free to a  C good home if someone would haul them away.  No takers.  Then I got  C really lucky and found somebody local who would haul them away and  E actually gave us $500 for two VAX 6000-320s and accessories (HSC95's  G star coupler, three racks full of RA series disks, etc.)  (By shutting  L those boxes down we also saved something like $7000/month on maintenance!!!)  # But that was five or six years ago!   H If you are very lucky someone might haul them away for free.  Otherwise 1 you'll probably have to pay someone to take them!    ------------------------------    Date: 17 May 2006 09:45:02 -0500; From: koehler@eisner.nospam.encompasserve.org (Bob Koehler) . Subject: Re: MCHKMCHK Bugcheck in VMS VAX V6.23 Message-ID: <yz589oWjeNU1@eisner.encompasserve.org>   q In article <1147841733.989622.93580@j33g2000cwa.googlegroups.com>, "Stanley F. Quayle" <quayle@pobox.com> writes: I > I'm at one of those "undisclosed locations" migrating a bunch of 8-inch H > floppies (RT-11 and RSX-11) to a VAX.  Twice today, the operator got aE > MCHKMCHK Bugcheck, which is "hardware check during hardware check".  > G > I *think* he's operated the handle on the floppy drive after the disk H > read began, since the first instance showed an OPCOM message asking toF > install a disk in DUB0: (the floppy drive).  Shortly thereafter, the? > request was satisified, and then the system did the BUGCHECK.  > 6 > Would this be it, or could there be another problem?  F   No I/O device error like this should cause a bugcheck, but there areF   potential issues with UNIBUS signal delays.  Many UNIBUS controllersE   will shut themselves down rather than do a double error in order to B   avoid a double error halt on the 11/78x models, but not all (the-   odds of getting the second error are slim).   0   I'd look for weak cables in the I/O subsystem.   ------------------------------  # Date: Wed, 17 May 2006 16:19:01 GMT , From: Hoff Hoffman <hoff-remove-this@hp.com>. Subject: Re: MCHKMCHK Bugcheck in VMS VAX V6.20 Message-ID: <V1Iag.618$CO1.355@news.cpqcorp.net>   Stanley F. Quayle wrote:1 > I'm at one of those "undisclosed locations" ...   @    Well, I'm working at an undisclosed location today, too.  :-)    > ...migrating a bunch of 8-inchH > floppies (RT-11 and RSX-11) to a VAX.  Twice today, the operator got aE > MCHKMCHK Bugcheck, which is "hardware check during hardware check".  > G > I *think* he's operated the handle on the floppy drive after the disk H > read began, since the first instance showed an OPCOM message asking toF > install a disk in DUB0: (the floppy drive).  Shortly thereafter, the? > request was satisified, and then the system did the BUGCHECK.  > 6 > Would this be it, or could there be another problem?  I    Machine checks are more typically hardware problems, or (given one of  H the product lines you work with) problems deep within the VAX emulator. I   This can include errors in main memory, the caches, the I/O bus or I/O  0 controllers, or within the VAX processor itself.  I    There have been low-level device driver issues on occasion, and these  / can also manifest themselves as machine checks.   E    I've never seen a problem caused by operating the drive door on a  G floppy, even something as old as an eight-inch RX01 drive.  Other than  D the obvious potential for read- and/or write-errors with the medium H presently loaded in the drive, that is.  (If you could crash an OpenVMS I box by operating the floppy drive door, well, that would very likely get  F the full attention of at least a few of the engineering folks here in 	 OpenVMS.)   C    Which VAX?  (Or is this environment running on a VAX emulator?)  : Which floppy?  RX01?  How is the drive connected?  Unibus?  E    The only way to troubleshoot this is with details of the VAX, and  I details of the machine check stack frame, and (potentially) with details  I from the error log.  And there are various driver and mandatory ECO kits  I available for OpenVMS VAX V6.2 (which should address the known low-level  F driver issues), so I'd check that the system is "current" with those. G (And if this is a VAX emulator, with the assistance and involvement of  @ the folks support whichever emulator is in use here, obviously.)  F    As was also suggested in another reply, do check your media and do  check the cabling and such.   G    And when all else fails :-), do contact the customer support center  > and/or the locally-preferred hardware support vendor directly.   ------------------------------  % Date: Wed, 17 May 2006 08:54:37 -0400 * From: "FredK" <fred.nospam@nospam.dec.com>. Subject: Re: MCHKMCHK Bugcheck in VMS VAX V6.2, Message-ID: <446b1cf9$1@usenet01.boi.hp.com>  - Bad hardware, possibly the floppy controller.   7 "Stanley F. Quayle" <quayle@pobox.com> wrote in message < news:1147841733.989622.93580@j33g2000cwa.googlegroups.com...I > I'm at one of those "undisclosed locations" migrating a bunch of 8-inch H > floppies (RT-11 and RSX-11) to a VAX.  Twice today, the operator got aE > MCHKMCHK Bugcheck, which is "hardware check during hardware check".  > G > I *think* he's operated the handle on the floppy drive after the disk H > read began, since the first instance showed an OPCOM message asking toF > install a disk in DUB0: (the floppy drive).  Shortly thereafter, the? > request was satisified, and then the system did the BUGCHECK.  > 6 > Would this be it, or could there be another problem? >    ------------------------------    Date: 17 May 2006 09:41:58 -0500; From: koehler@eisner.nospam.encompasserve.org (Bob Koehler) 7 Subject: Re: Multiple DECW$Sessions on one X terminal ? 3 Message-ID: <eBXCo7Jiaeqq@eisner.encompasserve.org>   ] In article <446AB0BE.26DEB838@teksavvy.com>, JF Mezei <jfmezei.spamnot@vaxination.ca> writes: F > When my workstation (BIKE) boots, I get a default login screen whichG > then gives me a session manager on node BIKE. From there, I can start   > applications that run on BIKE. > J > I'd like to get a second window, a session manager that runs on VELO but > displays on BIKE.  > D > It seems to be technically possible to run DECW$SESSION on 2 nodesG > targetted to the one X-TERMINAL. However, if you end the VELO session H > manager, VELO insists on starting a login screen on that display (even) > though it is still logged in to BIKE).   >   H    Do you really need DECW$SESSION, or can you get away with VUE$MASTER?   ------------------------------  % Date: Wed, 17 May 2006 02:39:29 -0400 ' From: Dave Froble <davef@tsoft-inc.com> : Subject: Re: OT: race for 8086s continues: 4 cores for AMD/ Message-ID: <vL6dnf__1PCoWffZRVn-iA@libcom.com>    JF Mezei wrote:  > Bill Todd wrote:H >> If you ignore FP performance (as most commercial software does), thatJ >> 'clear lead' is likely to arrive in only a couple of months, when Intel( >> ships its 'Woodcrest' server family:  > F > One must look at the "system level" performance, aka, the ability to  > scale to multiple processors.  > I > Is it fair to state that IA64 still enjoys an edge because it can scale  > to 64 processors ?    G Wrong, if you're talking about what's on the chip.  When talking about  H on chip routers and such, SMP glue, the itanic still scales to, are you H ready, one processor.  As far as I'm aware, all scaling is done off the H CPU chip, and there's nothing that prevents the same being done for any & other CPU, including Intel's x86 CPUs.  G > With the announced AMD chips, could a superdome class system be built I > with those chips (aka 64 cores) or is there still a lack of chipsets to $ > support such scaling of the 8086 ?     --  4 David Froble                       Tel: 724-529-0450> Dave Froble Enterprises, Inc.      E-Mail: davef@tsoft-inc.com DFE Ultralights, Inc.  170 Grimplin Road  Vanderbilt, PA  15486    ------------------------------  % Date: Wed, 17 May 2006 03:22:12 -0400 ( From: Bill Todd <billtodd@metrocast.net>: Subject: Re: OT: race for 8086s continues: 4 cores for AMD= Message-ID: <1Y6dnSTBg9BtU_fZRVn-qw@metrocastcablevision.com>    JF Mezei wrote:  > Bill Todd wrote:H >> If you ignore FP performance (as most commercial software does), thatJ >> 'clear lead' is likely to arrive in only a couple of months, when Intel( >> ships its 'Woodcrest' server family:  > F > One must look at the "system level" performance, aka, the ability to  > scale to multiple processors.  > I > Is it fair to state that IA64 still enjoys an edge because it can scale  > to 64 processors ?    I No - because IBM's X3 ('Hurricane') chipset already allows Xeon to scale  G to 32 sockets (64 cores) as well (and while I haven't seen any 64-core  I benchmarks for it, its 32-socket/32-core performance matches Itanic's in  < the benchmarks I've seen that they have in common; Unisys's > 16-socket/32-core Xeon offering also matches Itanic's 32-core  performance in TPC-C).   > G > With the announced AMD chips, could a superdome class system be built I > with those chips (aka 64 cores) or is there still a lack of chipsets to $ > support such scaling of the 8086 ?  I Horus developed a chipset supporting 32-socket (64-core) Opteron systems  E (using the currently-available processors) that reportedly seemed to  E scale at least comparably to IBM's 32-socket Xeon configuration, but  F AFAIK hasn't yet released it (even though it was planned to have been F released a while ago and has been demoed).  Recently a rumor surfaced E that AMD's glueless 32-socket configuration planned for 2007 or 2008  E might have taken the wind out of their sales, but even if that's the  F case I'd have expected them to have tried to make some profits in the  interim.   - bill   ------------------------------  % Date: Wed, 17 May 2006 05:20:27 -0400 - From: JF Mezei <jfmezei.spamnot@teksavvy.com> : Subject: Re: OT: race for 8086s continues: 4 cores for AMD, Message-ID: <446AEA9E.F088669E@teksavvy.com>   Dave Froble wrote:I > ready, one processor.  As far as I'm aware, all scaling is done off the I > CPU chip, and there's nothing that prevents the same being done for any ( > other CPU, including Intel's x86 CPUs.  G IA64 currently has supporting chipsets to allow a superdome to be built B with 64 IA64 cores. Since no 64 CPU 8086s have been built, I wouldC assume that there are no chipsets supporting 64 8086s in production  today ?     F Would system level chipsets deal exclusively with CPUs or with cores ?  F Isn't a core considered a full fledged standalone CPU when you look atH the system level architectures ? (so a 64 CPU superdome would consist of/ 32 dual cores CPU chips for instance, right ?).    ------------------------------  % Date: Wed, 17 May 2006 10:50:47 -0400 + From: Chuck Aaron <caaron@ceris.purdue.edu> ( Subject: Ping in web server command file5 Message-ID: <e4fd30$bau$1@mailhub227.itcs.purdue.edu>   , This is a multi-part message in MIME format.& --------------0001000701080407050203089 Content-Type: text/plain; charset=us-ascii; format=flowed  Content-Transfer-Encoding: 7bit   D Does anyone have the dcl tcpip command "ping" set up in a webserver  command fileG that returns the results and tests the results for validity...anything   like that at all? 7 If so, please email me the dcl setup to take a look at.    Thanks,  Chuck   & --------------000100070108040705020308) Content-Type: text/html; charset=us-ascii  Content-Transfer-Encoding: 7bit   ? <!DOCTYPE html PUBLIC "-//W3C//DTD HTML 4.01 Transitional//EN">  <html> <head>I   <meta content="text/html;charset=ISO-8859-1" http-equiv="Content-Type">    <title></title>  </head> ' <body bgcolor="#339999" text="#ffffff"> C Does anyone have the dcl tcpip command "ping" set up in a webserver  command file<br>F that returns the results and tests the results for validity...anything like that at all?<br> ; If so, please email me the dcl setup to take a look at.<br>  <br> Thanks,<br> 	 Chuck<br>  </body>  </html>   ( --------------000100070108040705020308--   ------------------------------  + Date: Wed, 17 May 2006 11:49:23 +0000 (UTC)  From: david20@alpha2.mdx.ac.uk@ Subject: Re: problem using mail from php-script in apache server) Message-ID: <e4f2k3$je4$1@news.mdx.ac.uk>   \ In article <446A2CF0.CF56BF0E@teksavvy.com>, JF Mezei <jfmezei.spamnot@teksavvy.com> writes: >Hoff Hoffman wrote:I >> whatever you want just before you send the mail.  That's how I "forge" G >> my return email address to the corporate email address, using TCP/IP  >> Services and standard MAIL.)  >  >----------- >BREAKING NEWS:  > > >SENIOR VMS ENGINEER ADMITS TO FORGING EMAIL ADDRESS AND GIVES, >INSTRUCTIONS TO THE WORLD ON HOW TO DO IT. H >Men in black are on their way. Hoff should be hearing black helicopters' >hovering over his car really soon now.  >  >:-) >  >--------------- >  > H >Seriously, there should be a way to safely authorize a username to send >on behalf of another. >  >something like: > q >$DEFINE/TABLE=(obscure name)/EXEC  TCPIP$SMTP_FROM_HOFFMAN  "chef@pastry.com","System@star.hp.com","hurd@hp.com"  > H >Then, when user "HOFFMAN" is logged in, he could define TCPIP$SMTP_FROMC >to any of the above 3 values, but not any other values. MAIL would F >ignore the user definition if its translation wasn't contained in the >"secure" definition.  > . >Right now, TCPIP$SMTP_FROM is all or nothing.  F But what is the point of securing SMTP mail at that level on VMS when O anybody can just connect to the smtp port and forge the FROM using the standard " documented SMTP protocol commands.  J As long as the syntax of the commands is correct the SMTP protocol does no- checking that the MAIL FROM address is valid.           
 David Webb Security team leader CCSS Middlesex University   ------------------------------    Date: 17 May 2006 06:35:51 -0700 From: davidc@montagar.com O Subject: Re: problem using mail from php-script in apache server [cf: Multinet] C Message-ID: <1147872951.811828.187000@u72g2000cwu.googlegroups.com>   B Here's a php module I use, lifted from the phpbb2 code.  It shouldF function much like the built-in mail() function, but uses SMTP, so youD can use an arbitrary from, headers, etc.  Note that this uses a mailF server at localhost, if you need to redirect that somehwere else, just5 replace "localhost" with the name of the SMTP server.   
 smtpmail.php:  <?php L /***************************************************************************(  *                              smtp.php,  *                       -------------------+  *   begin                : Wed May 09 2001 4  *   copyright            : (C) 2001 The phpBB Group-  *   email                : support@phpbb.com   *>  *   $Id: smtp.php,v 1.2 2006/01/26 19:53:10 taxsavercvs Exp $  *  L ***************************************************************************/  L /***************************************************************************  *B  *   This program is free software; you can redistribute it and/or modifyF  *   it under the terms of the GNU General Public License as published byF  *   the Free Software Foundation; either version 2 of the License, or(  *   (at your option) any later version.  *  L ***************************************************************************/   define('SMTP_INCLUDED', 1);    //. // This function has been modified as provided. // by SirSir to allow multiline responses when // using SMTP Extensions //; function server_parse($socket, $response, $line = __LINE__)  {          $server_response = '';5         while (substr($server_response, 3, 1) != ' ') 	         { >                 if (!($server_response = fgets($socket, 256)))                 { E                         message_die(GENERAL_ERROR, "Couldn't get mail - server response codes", "", $line, __FILE__);                  } 	         }   ;         if (!(substr($server_response, 0, 3) == $response)) 	         { E                 message_die(GENERAL_ERROR, "Ran into problems sending 8 Mail. Response: $server_response", "", $line, __FILE__);	         }  } 3 // Replacement or substitute for PHP's mail command F function smtpmail($mail_from, $mail_to, $subject, $message, $headers = '')  { B         // Fix any bare linefeeds in the message to make it RFC821
 Compliant.C         $message = preg_replace("#(?<!\r)\n#si", "\r\n", $message);            if ($headers != '') 	         { '                 if (is_array($headers))                  { 1                         if (sizeof($headers) > 1)                          { @                                 $headers = join("\n", $headers);                         }                          else                         { 7                                 $headers = $headers[0];                          }                  } *                 $headers = chop($headers);  G                 // Make sure there are no bare linefeeds in the headers @                 $headers = preg_replace('#(?<!\r)\n#si', "\r\n",
 $headers);  E                 // Ok this is rather confusing all things considered, C                 // but we have to grab bcc and cc headers and treat  them differentlyE                 // Something we really didn't take into consideration 
 originally:                 $header_array = explode("\r\n", $headers);&                 @reset($header_array);                   $headers = '';<                 while(list(, $header) = each($header_array))                 { F                         $headers .= ($header != '') ? $header . "\r\n" : '';                  }   *                 $headers = chop($headers);	         }   !         if (trim($subject) == '') 	         { <                 message_die(GENERAL_ERROR, "No email Subject$ specified", "", __LINE__, __FILE__);	         }   !         if (trim($message) == '') 	         { E                 message_die(GENERAL_ERROR, "Email message was blank",  "", __LINE__, __FILE__);	         }   C         // Ok we have error checked as much as we can to this point  let's get on         // it already.G         if( !$socket = @fsockopen('localhost', 25, $errno, $errstr, 20)  ) 	         { E                 message_die(GENERAL_ERROR, "Could not connect to smtp 2 host : $errno : $errstr", "", __LINE__, __FILE__);	         }            // Wait for reply /         server_parse($socket, "220", __LINE__);   7         fputs($socket, "HELO " . 'localhost' . "\r\n"); /         server_parse($socket, "250", __LINE__);   F         // From this point onward most server response codes should be 250 +         // Specify who the mail is from.... >         fputs($socket, "MAIL FROM: <" . $mail_from . ">\r\n");/         server_parse($socket, "250", __LINE__);   <         // Specify each user to send to and build to header.         $to_header = '';  C         // Add an additional bit of error checking to the To field. F         $mail_to = (trim($mail_to) == '') ? 'Undisclosed-recipients:;' : trim($mail_to); 3         if (preg_match('#[^ ]+\@[^ ]+#', $mail_to)) 	         { :                 fputs($socket, "RCPT TO: <$mail_to>\r\n");7                 server_parse($socket, "250", __LINE__); 	         }   G         // Ok now we tell the server we are ready to start sending data #         fputs($socket, "DATA\r\n");   F         // This is the last response code we look for until the end of the message./         server_parse($socket, "354", __LINE__);   #         // Send the Subject Line... 0         fputs($socket, "Subject: $subject\r\n");           // Now the To Header. +         fputs($socket, "To: $mail_to\r\n");   %         // Now any custom headers.... +         fputs($socket, "$headers\r\n\r\n");   1         // Ok now we are ready for the message... '         fputs($socket, "$message\r\n");   B         // Ok the all the ingredients are mixed in let's cook this puppy...          fputs($socket, ".\r\n");/         server_parse($socket, "250", __LINE__);   B         // Now tell the server we are done and close the socket...#         fputs($socket, "QUIT\r\n");          fclose($socket);           return TRUE; }    ?>   ------------------------------    Date: 17 May 2006 09:24:50 -0500; From: koehler@eisner.nospam.encompasserve.org (Bob Koehler) & Subject: Re: Results of my straw poll.3 Message-ID: <hK9SNBfzlpKo@eisner.encompasserve.org>   W In article <4cuf8sF17mleiU3@individual.net>, bill@cs.uofs.edu (Bill Gunshannon) writes:  > H > I won't argue this with you as I have no idea how logicals are done in > the meaty part of the OS.   C    If a newbie asked you to explain the difference between logicals G    and DCL symbols, I think you'ld cover all the relavent stuff without     knowing any inernals.   ------------------------------    Date: 17 May 2006 09:23:07 -0500; From: koehler@eisner.nospam.encompasserve.org (Bob Koehler) & Subject: Re: Results of my straw poll.3 Message-ID: <1MEVD8Q59FeU@eisner.encompasserve.org>   W In article <4cuf8sF17mleiU3@individual.net>, bill@cs.uofs.edu (Bill Gunshannon) writes:  > M > There is no way that links are going to let multiple users access different N > files using the same name, that's pretty much the reciprical of the way theyJ > work.  Links can give one file multiple names, but not one name multiple > files.  C    Yes you can.  The software opens "able".  You link whatever file E    you want to open to the name "able".  Your link is in your cwd and C    the next fellows link is in his cwd, so they don't have to point     to the same file.   ------------------------------    Date: 17 May 2006 09:33:49 -0500; From: koehler@eisner.nospam.encompasserve.org (Bob Koehler) & Subject: Re: Results of my straw poll.3 Message-ID: <KJbLWOC+sbkW@eisner.encompasserve.org>   W In article <4cv03mF17mvibU1@individual.net>, bill@cs.uofs.edu (Bill Gunshannon) writes: . > In article <446A2DA5.9A7D3EA7@teksavvy.com>,2 > 	JF Mezei <jfmezei.spamnot@teksavvy.com> writes: >> Bill Gunshannon wrote: H >>> The exact same thing can be done on Unix by defining $exampl_data inL >>> the same manner and then using that value when you open the file in your >>> program.   >>   >>  D >> You mean to tell me that in unix, fopen("$exampl_data","r") wouldK >> automatically result in "$exampl_data" being tested for a shell variable E >> and if so, automatically translated and  "JOES.DAT" being opened ?  > E > No, and that isn't what I said.  I specifically said the syntx was  / > different but the functionality was the same.  >   G    The functionality is not the same.  There are no job, group, system, H    or cluster shell symbols.  There are no inner mode (trusted) symbols.D    There are no user-defined symbol tables.  In all these ways shell3    symbols are like DCL symbols, not logical names.    ------------------------------    Date: 17 May 2006 09:31:39 -0500; From: koehler@eisner.nospam.encompasserve.org (Bob Koehler) & Subject: Re: Results of my straw poll.3 Message-ID: <DbHuyUiSUF2J@eisner.encompasserve.org>   W In article <4cunisF17om01U1@individual.net>, bill@cs.uofs.edu (Bill Gunshannon) writes:  > J > Unless I have totally misunderstood the original comment, it was basiclyF > singing the praises of having the abiltiy for EXAMPL$DATA mapping toK >  a file named JOES.DAT for joe and MARYS.DAT for mary so that when either G > of them run the same program the file that get's opened is different. F > The exact same thing can be done on Unix by defining $exampl_data inJ > the same manner and then using that value when you open the file in yourH > program.  The code may differ a bit, but the result is the same and in+ > both cases, you only write the code once.   D    You can use the same code, but its easier on VMS.  You don't haveC    to translate EXAMPL$DATA on VMS if its a logical name.  The file B    system will do it for you.  All you have to do is open the file&    with the name string "EXAMPL$DATA".  >    And then get critised for using the reserved character '$'.  J > Create a link to what?  If the program wants usr/config.dat then someone= > has to have created the file. No additional links involved.   G    If the program wants usr/config.dat and Joe has a link from ./usr to I    ~Joe/wanabe then the program will open ~Joe/wanabe/config.dat for Joe. D    If Mary has a link from ./usr to ~Steven/theboss then the program1    will open ~Steven/theboss/config.dat for Mary.    ------------------------------   Date: 17 May 2006 14:48:55 GMT( From: bill@cs.uofs.edu (Bill Gunshannon)& Subject: Re: Results of my straw poll., Message-ID: <4d0renF181hqsU1@individual.net>  3 In article <1MEVD8Q59FeU@eisner.encompasserve.org>, > 	koehler@eisner.nospam.encompasserve.org (Bob Koehler) writes:Y > In article <4cuf8sF17mleiU3@individual.net>, bill@cs.uofs.edu (Bill Gunshannon) writes:  >>  N >> There is no way that links are going to let multiple users access differentO >> files using the same name, that's pretty much the reciprical of the way they K >> work.  Links can give one file multiple names, but not one name multiple 	 >> files.  > E >    Yes you can.  The software opens "able".  You link whatever file G >    you want to open to the name "able".  Your link is in your cwd and E >    the next fellows link is in his cwd, so they don't have to point  >    to the same file.  C True, but looking at the original problem, if the needed data files D are always going to be in differnt directories then they can just be@ named "able" and there is no need to play around with links.  MyA understanding was that the desire was for the program to open the G correct file based on the logged in user who ran the program regardless E of what directory they were in.  The "link" solution fails if joe and G mary need to run the program from the same "cwd".  My example does not.   D I guess the problem comes back to software engineering.  The correctC solution can't be accurately derived until the problem is copletely C stated.  To be honest, I can't really think of a practical use for  D links.  I can think of no reason why a file needs more than one nameD any more than I do.  But then, it's not the first thing in Unix thatF serverd no practical purpose.  Look at groups. (and then you also haveC the SUID bit which has caused more problems than it likely solved.)    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>       ------------------------------   Date: 17 May 2006 14:51:59 GMT( From: bill@cs.uofs.edu (Bill Gunshannon)& Subject: Re: Results of my straw poll., Message-ID: <4d0rkfF181hqsU2@individual.net>  3 In article <hK9SNBfzlpKo@eisner.encompasserve.org>, > 	koehler@eisner.nospam.encompasserve.org (Bob Koehler) writes:Y > In article <4cuf8sF17mleiU3@individual.net>, bill@cs.uofs.edu (Bill Gunshannon) writes:  >>  I >> I won't argue this with you as I have no idea how logicals are done in  >> the meaty part of the OS. > E >    If a newbie asked you to explain the difference between logicals I >    and DCL symbols, I think you'ld cover all the relavent stuff without  >    knowing any inernals.  F I think you'ld be wrong.  Other than possibly globality, I don't thinkF I understand the difference myself.  And I have worked with VMS at the user level since at least 1981.    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>       ------------------------------   Date: 17 May 2006 14:58:55 GMT( From: bill@cs.uofs.edu (Bill Gunshannon)& Subject: Re: Results of my straw poll., Message-ID: <4d0s1fF181hqsU3@individual.net>  3 In article <DbHuyUiSUF2J@eisner.encompasserve.org>, > 	koehler@eisner.nospam.encompasserve.org (Bob Koehler) writes:Y > In article <4cunisF17om01U1@individual.net>, bill@cs.uofs.edu (Bill Gunshannon) writes:  >>  K >> Unless I have totally misunderstood the original comment, it was basicly G >> singing the praises of having the abiltiy for EXAMPL$DATA mapping to L >>  a file named JOES.DAT for joe and MARYS.DAT for mary so that when eitherH >> of them run the same program the file that get's opened is different.G >> The exact same thing can be done on Unix by defining $exampl_data in K >> the same manner and then using that value when you open the file in your I >> program.  The code may differ a bit, but the result is the same and in , >> both cases, you only write the code once. > F >    You can use the same code, but its easier on VMS.  You don't haveE >    to translate EXAMPL$DATA on VMS if its a logical name.  The file D >    system will do it for you.  All you have to do is open the file( >    with the name string "EXAMPL$DATA".  F But what makes that such a big deal?  Once the program is written, theG functionality (in this particualr case!) is the same.  I can change the E actual file used by the program without having to change the program. I That is the example given and my method does it just as well as LOGIGALS. G It may be (and probably is) that LOGICALS can do a lot of other things, H too.  But I was addressing this particular application, as badly defined
 as it was.   > @ >    And then get critised for using the reserved character '$'. > K >> Create a link to what?  If the program wants usr/config.dat then someone > >> has to have created the file. No additional links involved. > I >    If the program wants usr/config.dat and Joe has a link from ./usr to K >    ~Joe/wanabe then the program will open ~Joe/wanabe/config.dat for Joe. F >    If Mary has a link from ./usr to ~Steven/theboss then the program3 >    will open ~Steven/theboss/config.dat for Mary.   F But that depends on where they are located in the filesystem more thanF on who the user is.  Not the way I saw the problem defined.  I saw theE problem as wanting user specific files rather than files that changed  depending on what your CWD was.    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>       ------------------------------  % Date: Wed, 17 May 2006 10:37:08 -0400 # From: sol gongola <sol@adldata.com> & Subject: Re: Results of my straw poll.0 Message-ID: <1147876655.994460@nntp.acecape.com>   Bob Koehler wrote:Y > In article <4cuf8sF17mleiU3@individual.net>, bill@cs.uofs.edu (Bill Gunshannon) writes: N >> There is no way that links are going to let multiple users access differentO >> files using the same name, that's pretty much the reciprical of the way they K >> work.  Links can give one file multiple names, but not one name multiple 	 >> files.  > E >    Yes you can.  The software opens "able".  You link whatever file G >    you want to open to the name "able".  Your link is in your cwd and E >    the next fellows link is in his cwd, so they don't have to point  >    to the same file. > M That only works if you are running from your cwd. That isn't always the case. M Rephrase to say if able is under your home directory and references to "able" . are always used relative to that as in ~/able.  F It still ain't logical but it can work if all links are done that way.   ------------------------------    Date: 17 May 2006 09:10:20 -0700- From: "Doug Phillips" <dphill46@netscape.net> & Subject: Re: Results of my straw poll.B Message-ID: <1147882220.495415.65410@u72g2000cwu.googlegroups.com>   Bill Gunshannon wrote:5 > In article <hK9SNBfzlpKo@eisner.encompasserve.org>, @ > 	koehler@eisner.nospam.encompasserve.org (Bob Koehler) writes:[ > > In article <4cuf8sF17mleiU3@individual.net>, bill@cs.uofs.edu (Bill Gunshannon) writes:  > >>K > >> I won't argue this with you as I have no idea how logicals are done in  > >> the meaty part of the OS. > > G > >    If a newbie asked you to explain the difference between logicals K > >    and DCL symbols, I think you'ld cover all the relavent stuff without  > >    knowing any inernals. > H > I think you'ld be wrong.  Other than possibly globality, I don't think% > I understand the difference myself.  >   F Is that because the difference is too complex for you to understand or0 because you've never taken the time to find out?  C > And I have worked with VMS at the user level since at least 1981.   G And in all that time you've never really needed to know the deep, dark, @ VMS secrets, have you. But, with *nix you've learned the secretsC because you *had* to. You're more familiar with *nix because it has  demanded more of your time.   8 Is that too obvious to see or too logical to comprehend?   ------------------------------   Date: 17 May 2006 16:29:57 GMT( From: bill@cs.uofs.edu (Bill Gunshannon)& Subject: Re: Results of my straw poll., Message-ID: <4d11c5F183pjvU1@individual.net>  B In article <1147882220.495415.65410@u72g2000cwu.googlegroups.com>,0 	"Doug Phillips" <dphill46@netscape.net> writes: > Bill Gunshannon wrote:6 >> In article <hK9SNBfzlpKo@eisner.encompasserve.org>,A >> 	koehler@eisner.nospam.encompasserve.org (Bob Koehler) writes: \ >> > In article <4cuf8sF17mleiU3@individual.net>, bill@cs.uofs.edu (Bill Gunshannon) writes: >> >> L >> >> I won't argue this with you as I have no idea how logicals are done in >> >> the meaty part of the OS.  >> >H >> >    If a newbie asked you to explain the difference between logicalsL >> >    and DCL symbols, I think you'ld cover all the relavent stuff without >> >    knowing any inernals.  >>I >> I think you'ld be wrong.  Other than possibly globality, I don't think & >> I understand the difference myself. >> > H > Is that because the difference is too complex for you to understand or2 > because you've never taken the time to find out?  G Probably because I have never run into a task I had to accomplish where E I needed the functionality that is unique to them.  That and the fact B that I haven't a clue where to look for the answer and I have moreF important thigs to do than read obscure documentation for the it's own sake.    > D >> And I have worked with VMS at the user level since at least 1981. > I > And in all that time you've never really needed to know the deep, dark, B > VMS secrets, have you. But, with *nix you've learned the secrets > because you *had* to.   E Nice try, but no.  I learned the Unix Internals because they are well B documented in a lot of places and I had jobs that involved dealingE with device drivers and Unix Internal Data Structures.  Had to in the G sense that I had employers who expected me to do this, but more because D I could find easy to read books that taught me all I needed to know.G I have yet to see the same books for VMS.  But then, may they are being + marketed by the same people who market VMS.   E >                       You're more familiar with *nix because it has  > demanded more of your time.   E No, actually, it demands less of my time.  I wish it demanded more as D I enjoy working with it a lot more thant he alternatives, but I haveF to but my time in where my boss thinks it is best spent.  As I pointedF out here in the past, my daughter learned how to use Unix when she was" about 6.  How difficult can it be?   > : > Is that too obvious to see or too logical to comprehend?  $ Not obvious or logical, just absurd.   bill   --  J Bill Gunshannon          |  de-moc-ra-cy (di mok' ra see) n.  Three wolvesD bill@cs.scranton.edu     |  and a sheep voting on what's for dinner. University of Scranton   |A Scranton, Pennsylvania   |         #include <std.disclaimer.h>       ------------------------------   Date: 17 May 2006 16:52:27 GMT( From: bill@cs.uofs.edu (Bill Gunshannon)& Subject: Re: Results of my straw poll., Message-ID: <4d12mbF18a2aoU1@individual.net>  0 In article <BfIag.621$wM1.421@news.cpqcorp.net>,/ 	Hoff Hoffman <hoff-remove-this@hp.com> writes:  > Bill Gunshannon wrote: > I >> But what makes that such a big deal?  Once the program is written, the 8 >> functionality (in this particualr case!) is the same. > F >    The difference is one of transparency -- a program that gives no ? > stead to logical names can usually be gotten to work through  I > logical-name based redirections, and a program that correctly uses the  I > primary and default specifications on OpenVMS I/O is far more flexible.  > K >    If I can modify the code, then I can make it do most anything.  Up to  H > and including having the tool use a relational database rather than C I > I/O.  If I can modify the code, I can alter the underpinnings of the C  K > fopen call and the kernel to provide whatever capabilities I want, up to  I > and including providing logical name support within Unix/Linux.  (That  G > way, I can completely avoid modifying the source code, too.  I'm not  E > kidding here, either.  I've ported code where the programmers have  F > chosen to modify standard C calls for new and different behaviours. I > That's a construct that takes some thought and some time to extricate,   > too.)  > I >    Like RMS and the integrated support for various record formats, and  K > like the integrated support for file locking, the point of logical names  A > is more that you don't have to deal with it.   "It just works".  > E >    If I am willing to "deal with it" myself and to invest the time  K > necessary (from adding a getenv call into the application up to altering  J > the kernel, or simply patching the executables to alter the processing, G > to an industry favorite time-sink of re-re-re-porting source code to  J > re-re-re-re-getenv the logic), I can make "it" do most anything I might E > want.  Locks.  Clustering.  I/O redirection.  Indexed file access.   > Whatever.    I And after way too much heat and very little light, Hoff comes thru again. ? I think I can now see the primary advantage to this approach.     
 Thank you.   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>       ------------------------------    Date: 17 May 2006 09:53:19 -0700- From: "Doug Phillips" <dphill46@netscape.net> & Subject: Re: Results of my straw poll.B Message-ID: <1147884799.669860.39840@i40g2000cwc.googlegroups.com>   Dave Froble wrote: > Bill Gunshannon wrote:0 > > In article <446A674C.563A37A0@teksavvy.com>,4 > > 	JF Mezei <jfmezei.spamnot@teksavvy.com> writes: > >> Bill Gunshannon wrote: 6 > >>>         if ((filename = getenv(name)) == NULL) {> > >>>                 printf("\nEXAMPLE_DATA not defined.\n"); > >>>                 exit(-1);  > >>>                 } 7 > >>>         if ((fp = fopen(filename,"r")) == NULL) { < > >>>                 printf("\nCan't open %s\n",*filename); > >>>                 exit(-1); K > >> OK, this is fair enough. Not quite as transparent as logical names but K > >> fair enough. Question is what percentage of applications will actually I > >> be built that way to use environment variables (and add codes to use L > >> some default file name if the environment variable isn't translatable). > > O > > I would imagine any one that requires it.  That's what Software Engineering  > > is all about.  > > B > >> What this lacks though is the ability to have something like: > >>I > >> 	fopen("CHOCOLATE$DIR:PASTRY.TXT")  where you combine a logical name  > >> with a fixed file name. > > L > > Sigh.....   One more line of code and I can duplicate this functionalityJ > > as well.  Of course, we could start a contest trying to see who can beM > > the first one to find an obscure way of doing things that the other can't 
 > > mimic. > >  > >  > > bill > >  > G > As a developer who has worked on VMS, and now windoz, but not Unix, I B > really miss logicals when I'm not using VMS.  So far, I see manyI > examples for filenames, file locations, and such.  Logicals give you so F > much more.  You can use them for synchronization, status, flags, andJ > much more.  Generally, when I miss them the most, it's not for files and5 > locations.  I've always had other methods for that.  >    Dave,   G I've used windoz Environment Variables to perform those functions. They B work more like symbols but with a few characteristics of logicals:  F o You can set a value at the operating system levels that is available to all processes. F o You can set a value at the user level that is only available to that user. " o You can set nested local values.) o (see help->index->environment variable)   F Not quite as powerful as logicals, but better than nothing. Anything ID can't figure out how to do with ev's, I revert to the old pass via a; file or control record method. I'm probably not alone here.     -Doug   ------------------------------    Date: 17 May 2006 12:12:24 -0500; From: koehler@eisner.nospam.encompasserve.org (Bob Koehler) & Subject: Re: Results of my straw poll.3 Message-ID: <JxVLRRvZ0Psa@eisner.encompasserve.org>   W In article <4d0rkfF181hqsU2@individual.net>, bill@cs.uofs.edu (Bill Gunshannon) writes:  > H > I think you'ld be wrong.  Other than possibly globality, I don't thinkH > I understand the difference myself.  And I have worked with VMS at the! > user level since at least 1981.   F    OK, then, I guess you don't know as much about VMS as I thought you    did.    ------------------------------  # Date: Wed, 17 May 2006 17:05:13 GMT , From: Hoff Hoffman <hoff-remove-this@hp.com>& Subject: Re: Results of my straw poll.0 Message-ID: <dJIag.631$HM1.435@news.cpqcorp.net>   Bill Gunshannon wrote:  7 > ...I learned the Unix Internals because they are well D > documented in a lot of places and I had jobs that involved dealingG > with device drivers and Unix Internal Data Structures.  Had to in the I > sense that I had employers who expected me to do this, but more because F > I could find easy to read books that taught me all I needed to know.I > I have yet to see the same books for VMS.  But then, may they are being - > marketed by the same people who market VMS.   E    The Internals and Data Structures Manuals (the IDSM), the Writing  H Device Drivers in C book, and the OpenVMS source listings are the usual I tomes for this task, and there have been any number of internals courses  D at the symposia over the years (and many of which are now posted at C various web sites), and there will be a barrage of new courses and  E materials presented at the bootcamp next week.  (Which reminds me, I  F need to write my presentations.  :-) )  There are also various formal E training programs and courses available for OpenVMS, from user-level  A commands, to system management operations, to advanced internals   programming classes.  H    Linux champions the ability of the end-user to change and to improve E and to tailor the kernel to meet local requirements and the relative  G forward speed of the platform development, and is rather less centered  C on the source and executable compatibility of applications, and on   cross-platform compatibility.   H    OpenVMS classically targets cross-platform and upward-compatibility, @ and the consistency and transparency of the integrated features.  H    These are two quite different approaches, and two different customer / markets, though both are obviously quite valid.   C    The customers here do differ in the individual requirements and  I expectations, of course.  OpenVMS customers can tend to become concerned  G if OpenVMS Engineering should make incompatible changes to an API, and  C thus requires rebuilding code.  Linux customers can tend to become  I concerned if the latest APIs aren't available, and don't mind rebuilding  F source code to get there.   These are two very different (and equally 7 valid) customer environments and customer requirements.   D    All that written, when any product moves up into the range where B supported products, platforms and related services are available, G however, you'll find the kernel access and the site-specific tailoring  D requirements and the associated limitations are rather more similar G across the available products.  There are differences in offerings and  H expectations with self-support, with community support, and with formal  vendor support, obviously.   ------------------------------   Date: 17 May 2006 17:26:50 GMT( From: bill@cs.uofs.edu (Bill Gunshannon)& Subject: Re: Results of my straw poll., Message-ID: <4d14mpF16mbtjU1@individual.net>  3 In article <JxVLRRvZ0Psa@eisner.encompasserve.org>, > 	koehler@eisner.nospam.encompasserve.org (Bob Koehler) writes:Y > In article <4d0rkfF181hqsU2@individual.net>, bill@cs.uofs.edu (Bill Gunshannon) writes:  >>  I >> I think you'ld be wrong.  Other than possibly globality, I don't think I >> I understand the difference myself.  And I have worked with VMS at the " >> user level since at least 1981. > H >    OK, then, I guess you don't know as much about VMS as I thought you	 >    did.   D I have never claimed to be anywhere near as proficient with VMS as IG am with Unix.  I would have assumed that was obvious from the technical F questions I come here with every so often.  But all that aside, I haveE written production programs for the VMS environment in the commercial C world and could again.  While there are probably better ways to get J some of the things I do done, I can still get the job done when it counts.J Getting back to where this all began, this is probably another good reasonI why getting VMS back into the EDU world would be a good thing. If someone G like me who has a commercial as well as academic background (no, I have I not been working in a school for my whole career) and has been successful F at it knows so little of the real advantages and of the better ways toC do things on VMS what does that bode for the future if VMS can't be I brought back into the classroom?  I have had people here ask, "Why didn't F you learn how to use LOGICALS?"  The answer is really quite simple.  II have never had to do a task where a more common (in the textbook meaning) E method wasn't available.  If I had learned VMS in depth, I might have L opted for a less portable but more elegant in it's own environment solution.I I can assure you that after only one course using VMS our students may be I able to compile a simple program, but they actually know a lot less about G VMS than even I do.  I'm glad for what little exposure they get, but if H the survival of VMS is paramount, we need to find ways to get them a lot more.    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>       ------------------------------    Date: 17 May 2006 12:14:10 -0500; From: koehler@eisner.nospam.encompasserve.org (Bob Koehler) & Subject: Re: Results of my straw poll.3 Message-ID: <Fq+YXPQUDUGG@eisner.encompasserve.org>   W In article <4d11c5F183pjvU1@individual.net>, bill@cs.uofs.edu (Bill Gunshannon) writes:  > I > Probably because I have never run into a task I had to accomplish where G > I needed the functionality that is unique to them.  That and the fact D > that I haven't a clue where to look for the answer and I have moreH > important thigs to do than read obscure documentation for the it's own > sake.   2    Gee, I got it all out of HELP.  That's obscure?  D    Now I'm starting to understand why you need so much time when you    deal with VMS.    ------------------------------    Date: 17 May 2006 12:20:40 -0500; From: koehler@eisner.nospam.encompasserve.org (Bob Koehler) & Subject: Re: Results of my straw poll.3 Message-ID: <P4vrTdyrTzeV@eisner.encompasserve.org>   W In article <4d0s1fF181hqsU3@individual.net>, bill@cs.uofs.edu (Bill Gunshannon) writes:  > H > But what makes that such a big deal?  Once the program is written, the7 > functionality (in this particualr case!) is the same.   *    Transparency to the programmer is good.  E    If you're going to deal with non-transparent techniques, you could E    just have one file which lists all the other files you need.  Then H    all you have to do is find that first file (in other words you didn'tF    really solve the problem).  Whatever solution you come up with, youC    it may not be usefull in every language someone wants to work in C    (e.g. is there an f77 interface to getenv() in the POSIX spec?).   F    Meanwhile all I'm doing is coding a string in an open statement, noC    matter what language I'm programming in.  Problem solved for me.    ------------------------------    Date: 17 May 2006 10:32:22 -0700- From: "Doug Phillips" <dphill46@netscape.net> & Subject: Re: Results of my straw poll.C Message-ID: <1147887142.041661.156310@i39g2000cwa.googlegroups.com>    Bill Gunshannon wrote:D > In article <1147882220.495415.65410@u72g2000cwu.googlegroups.com>,2 > 	"Doug Phillips" <dphill46@netscape.net> writes: > > Bill Gunshannon wrote:8 > >> In article <hK9SNBfzlpKo@eisner.encompasserve.org>,C > >> 	koehler@eisner.nospam.encompasserve.org (Bob Koehler) writes: ^ > >> > In article <4cuf8sF17mleiU3@individual.net>, bill@cs.uofs.edu (Bill Gunshannon) writes: > >> >> N > >> >> I won't argue this with you as I have no idea how logicals are done in! > >> >> the meaty part of the OS.  > >> >J > >> >    If a newbie asked you to explain the difference between logicalsN > >> >    and DCL symbols, I think you'ld cover all the relavent stuff without > >> >    knowing any inernals.  > >>K > >> I think you'ld be wrong.  Other than possibly globality, I don't think ( > >> I understand the difference myself. > >> > > J > > Is that because the difference is too complex for you to understand or4 > > because you've never taken the time to find out? > I > Probably because I have never run into a task I had to accomplish where G > I needed the functionality that is unique to them.  That and the fact D > that I haven't a clue where to look for the answer and I have moreH > important thigs to do than read obscure documentation for the it's own > sake.  >    Here's a place to start:  - <http://www.openvms.org/pages.php?page=Books>   @ Your local book store should have a catalog you can reference ifC shopping on-line isn't your thing. There have been VMS books around B since way-back-when and if you believe the excellent documentationA available from (the owner) is too complicated or too simple, then @ you're not looking at the right manual. You want hardcopy? OrderG hardcopy. Order a book. Your "bad documentation / no books" excuse only & proves that you haven't really looked.   > > F > >> And I have worked with VMS at the user level since at least 1981. > > K > > And in all that time you've never really needed to know the deep, dark, D > > VMS secrets, have you. But, with *nix you've learned the secrets > > because you *had* to.  > G > Nice try, but no.  I learned the Unix Internals because they are well D > documented in a lot of places and I had jobs that involved dealingF > with device drivers and Unix Internal Data Structures. Had to in the8 > sense that I had employers who expected me to do this,   Okay. That makes sense.    > but more becauseF > I could find easy to read books that taught me all I needed to know.+ > I have yet to see the same books for VMS.   
 That doesn't.   E I'm looking at a set of documentation from the early 80's. There's an E "Introduction to VAX/VMS" and about 9 other little books whose titles E start with "Guide to..." and touch everything from DCL to Clustering. C For more in depth understanding there's a wall of books that covers 5 nearly every aspect of VMS from its surface to heart.   G Any company involved in development on VMS needed to order at least the 4 level of documentation that covered that level. Duh!  E  All the information you need is available, and at whatever level you E need. That was true then and it's true now. The "no documentation, no & good books" excuse doesn't hold water.   >  But then, may they are being - > marketed by the same people who market VMS.  >    Some are, some aren't.  G > >                       You're more familiar with *nix because it has  > > demanded more of your time.  > + > No, actually, it demands less of my time.   F You're saying that VMS demands more of *your* time than *nix? Methinks3 you're confusing VMS with that there Winders thang.    >  I wish it demanded more as F > I enjoy working with it a lot more thant he alternatives, but I haveH > to but my time in where my boss thinks it is best spent.  As I pointedH > out here in the past, my daughter learned how to use Unix when she was$ > about 6.  How difficult can it be?  B Proves nothing. Your daughter is probably a genius or at least hasD above average intellegence along with a father willing to spend time	 with her.    >  > > < > > Is that too obvious to see or too logical to comprehend? > & > Not obvious or logical, just absurd. >   , Denial of the obvious is the only absurdity.   ------------------------------   Date: 17 May 2006 17:37:33 GMT( From: bill@cs.uofs.edu (Bill Gunshannon)& Subject: Re: Results of my straw poll., Message-ID: <4d15atF16mbtjU2@individual.net>  3 In article <P4vrTdyrTzeV@eisner.encompasserve.org>, > 	koehler@eisner.nospam.encompasserve.org (Bob Koehler) writes:Y > In article <4d0s1fF181hqsU3@individual.net>, bill@cs.uofs.edu (Bill Gunshannon) writes:  >>  I >> But what makes that such a big deal?  Once the program is written, the 8 >> functionality (in this particualr case!) is the same. > , >    Transparency to the programmer is good. > G >    If you're going to deal with non-transparent techniques, you could G >    just have one file which lists all the other files you need.  Then J >    all you have to do is find that first file (in other words you didn'tH >    really solve the problem).  Whatever solution you come up with, youE >    it may not be usefull in every language someone wants to work in E >    (e.g. is there an f77 interface to getenv() in the POSIX spec?).  > H >    Meanwhile all I'm doing is coding a string in an open statement, noE >    matter what language I'm programming in.  Problem solved for me.   E As I said in response to Hoff's fine post, I can see the advantage in B view of transparency, but it has a cost in portability (as does myE solution, but to a lesser extent).  Long before even POSIX the getenv + solution was recognized as standard by ISO.    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>       ------------------------------  % Date: Wed, 17 May 2006 17:55:28 +0100 ; From: Gerald Marsh <marsh_family--spammenot.btconnect.com@> % Subject: SORT, LEF, SYS$PUT and COBOL 8 Message-ID: <elkm62tsv305ibiuk91r8doj3buist1gvn@4ax.com>   Hello VMSers  = I'm hoping someone out there will have a clue about this one:   ( OpenVMS - V7.1 I think - running on VAX.  E Old COBOL program hangs in LEF in the middle of a large sort. I think B the sort is the one in the PROCEDURE-DIVISION and not the callable SOR$ routines.  $ SDA shows busy channel to SORTWORK1.  B Program counter in $PUT within SORTSHR. Oh, after the WAITFR call.  D Monitored the process for three hours and it doesn't use CPU neither does DIO or BIO increase.   E Previous attempts to delete the process have resulted in an RWAST and  reboot.   E I cannot see any quotas being exhausted including the pooled quotas - ( there are no other processes in the job.  @ Sometimes it works ok but we've seen this a few times now and it0 always requires a reboot to "sort" the mess out.  D We were hoping to spread the load around more files but I cannot seeC any way of increasing the number of work files which COBOL uses for > its own sort. I'm quite sure there's no logical name which can override the default of two.  A Any help gratefully recieved as ever and keep that flag a'flying.    Bye for now,     Gerald.    ------------------------------  + Date: Wed, 17 May 2006 11:49:16 +0000 (UTC) P From: helbig@astro.multiCLOTHESvax.de (Phillip Helbig---remove CLOTHES to reply)5 Subject: speeding up LAVC with switch instead of hub? $ Message-ID: <e4f2js$m11$1@online.de>  H I'm looking to speed up communication in my LAN.  I have a cluster with H all communication on the LAN (SCS, TCPIP, LAT etc; shadow copies/merges H go across the network as well).  Now, everything is connected via a hub.  G If I understand things correctly, a switch gives me two advantages with F regard to speed: ethernet controllers can operate in full-duplex mode, and collisions are eliminated.  D Full duplex would double the bandwidth, whereas avoiding collisions C would only increase bandwidth if collisions are reducing it, right?   F Normally, collisions aren't a problem and there are many of them only   during shadow copies and merges.  G To benefit from full duplex, the ethernet controllers would have to be  I capable of full duplex.  Would that be the case for any of the following  1 machines (which have the "stock" ethernet cards)?       o  VAXstation 4000/90      o  VAX 4000/105A       o  DEC 3000/600      o  DEC 3000/300LX  0    o  ALPHAserver 1200 (aka Digital Server 5305)   ------------------------------    Date: 17 May 2006 14:39:27 +01006 From: peter@langstoeger.at (Peter 'EPLAN' LANGSTOEGER)9 Subject: Re: speeding up LAVC with switch instead of hub? * Message-ID: <446b359f@news.langstoeger.at>  w In article <e4f2js$m11$1@online.de>, helbig@astro.multiCLOTHESvax.de (Phillip Helbig---remove CLOTHES to reply) writes: H >To benefit from full duplex, the ethernet controllers would have to be J >capable of full duplex.  Would that be the case for any of the following 2 >machines (which have the "stock" ethernet cards)? >  >   o  VAXstation 4000/90  >  >   o  VAX 4000/105A >  >   o  DEC 3000/600  >  >   o  DEC 3000/300LX    No. 10MBit/Half Duplex only   1 >   o  ALPHAserver 1200 (aka Digital Server 5305)    Probably, Yes.K If it is 100MBit, then FD is likely, if 10MB (which I doubt) then unlikely.    --   Peter "EPLAN" LANGSTOEGER % Network and OpenVMS system specialist  E-mail  peter@langstoeger.atF A-1030 VIENNA  AUSTRIA              I'm not a pessimist, I'm a realist   ------------------------------  % Date: Wed, 17 May 2006 13:48:42 +0100 * From: "Richard Brodie" <R.Brodie@rl.ac.uk>9 Subject: Re: speeding up LAVC with switch instead of hub? 2 Message-ID: <e4f63a$7rl$1@blackmamba.itd.rl.ac.uk>  V "Phillip Helbig---remove CLOTHES to reply" <helbig@astro.multiCLOTHESvax.de> wrote in & message news:e4f2js$m11$1@online.de...  I > If I understand things correctly, a switch gives me two advantages with H > regard to speed: ethernet controllers can operate in full-duplex mode,  > and collisions are eliminated.  B OK, so far - although they are pretty much the same thing. Two bigF things you omitted: the backplane of a switch will usually have enoughF capacity to drive all ports simultaneously at full speed (assuming youC can match data sources and sinks). Secondly, you can run systems at  different speeds.   E > Full duplex would double the bandwidth, whereas avoiding collisions E > would only increase bandwidth if collisions are reducing it, right?   F You have to be slightly careful about 'double'. With a small number ofD systems (unlike a big network trunk), you are likely to find trafficA predominately in one direction at any given moment. You win a bit A in that your bulk data and acknowledgements don't contend for the  network.  G > Normally, collisions aren't a problem and there are many of them only " > during shadow copies and merges.  H Best just to forget about them, their overhead in a five node network is going to be minute.   H > To benefit from full duplex, the ethernet controllers would have to beJ > capable of full duplex.  Would that be the case for any of the following3 > machines (which have the "stock" ethernet cards)?   F Yes. An Alphaserver 1200 would likely have a 100Mb/s full duplex card. No to the VAXes., Maybe - Turbochannel Alphas. I can't recall.  A You don't give a lot of information about your disk layout or how  your load is balanced:  E  If most of your traffic were to/from the 1200 then letting it rip at @ 100Mb/s looks like a big win. Although having two faster servers,  to run the shadow sets off would be better.   ------------------------------  + Date: Wed, 17 May 2006 13:25:54 +0000 (UTC) P From: helbig@astro.multiCLOTHESvax.de (Phillip Helbig---remove CLOTHES to reply)9 Subject: Re: speeding up LAVC with switch instead of hub? $ Message-ID: <e4f892$f46$1@online.de>  F In article <446b359f@news.langstoeger.at>, peter@langstoeger.at (Peter 'EPLAN' LANGSTOEGER) writes:     > >   o  VAXstation 4000/90  > >  > >   o  VAX 4000/105A > >  > >   o  DEC 3000/600  > >  > >   o  DEC 3000/300LX  >  > No. 10MBit/Half Duplex only  > 3 > >   o  ALPHAserver 1200 (aka Digital Server 5305)  >  > Probably, Yes.M > If it is 100MBit, then FD is likely, if 10MB (which I doubt) then unlikely.   F Since collisions are usually not a problem (and when they are, such asC when a shadow copy is taking place, "normal use" is not recommended F anyway), then without full-duplex capability, I suppose a switch isn'tG worth the investment.  (I only use the 5305 when I need to use Mozilla, G and that tends to generate a lot of traffic.  Since I have a spare port F in the switch in the DSL router (to which the cluster hub is connectedF by an uplink), it would make more sense to plug the 5305 directly into! a switch port in the DSL router.)    ------------------------------  # Date: Wed, 17 May 2006 13:53:51 GMT , From: Hoff Hoffman <hoff-remove-this@hp.com>9 Subject: Re: speeding up LAVC with switch instead of hub? 0 Message-ID: <PVFag.604$dH1.189@news.cpqcorp.net>  / Phillip Helbig---remove CLOTHES to reply wrote: J > I'm looking to speed up communication in my LAN.  I have a cluster with J > all communication on the LAN (SCS, TCPIP, LAT etc; shadow copies/merges J > go across the network as well).  Now, everything is connected via a hub.  E    I would not expect the LAN infrastructure to be the central limit  F here.  A switch will certainly help isolate the traffic, but it won't $ help make a 10Mb VAX NIC any faster.  G    I'd look to see if it is the VAX boxes, or the 10Mb connections, or  F both.  With a shadow copy, you're jamming a whole pile of data over a C comparatively small pipe, so you will see collisions and back-offs.   E    (Run some math, and see what the theoretical best case for a full  I copy is, given your disk sizes and your NIC speeds.  If that calculation  G is not hugely different from what you are experiencing, then the limit  % likely isn't the LAN infrastructure.)   H    With current releases of OpenVMS, you can use some of the throttling G capabilities within shadowing (see the documentation for the shadowing  D ECO kits, or the OpenVMS documentation for shadowing on the current I releases) and dial back the network usage (trading off network bandwidth  H for time), but I'd expect the core configuration limit is comparatively ' big data and comparatively small pipes.    ------------------------------  % Date: Wed, 17 May 2006 15:10:39 +0100 * From: "Richard Brodie" <R.Brodie@rl.ac.uk>9 Subject: Re: speeding up LAVC with switch instead of hub? 2 Message-ID: <e4fasv$a1g$1@blackmamba.itd.rl.ac.uk>  V "Phillip Helbig---remove CLOTHES to reply" <helbig@astro.multiCLOTHESvax.de> wrote in & message news:e4f8ub$gvk$1@online.de...  K >> Best just to forget about them, their overhead in a five node network is  >> going to be minute. > ; > I do see the collision LED light up during a shadow copy.   A Yes, I expect you would. The LED indicators are invariably hugely  pulse stretched, though.     ------------------------------  % Date: Wed, 17 May 2006 09:55:16 -0400  From: norm.raphael@metso.com( Subject: Thundering Hurd pacing HP gainsQ Message-ID: <OFAC1B2AC0.1E511450-ON85257171.004BED15-85257171.004C70E4@metso.com>    Thundering Hurd pacing HP gains I Under CEO's guidance, quarterly earnings above expectations, and forecast I for current quarter better than expected ; shares rally 4.5% after hours. - By Amanda Cantrell, CNNMoney.com staff writer  May 16, 2006: 6:41 PM EDT     F NEW YORK (CNNMoney.com) - Hewlett-Packard Co. on Tuesday reported a 45I percent quarterly profit increase in an earnings report that handily beat E Wall Street analysts' expectations, thanks to strong sales of PCs and , printing supplies and cost-cutting measures.    B Shares of the computer and peripherals company rose 4.5 percent in after-hours trading.    I Analysts and investors said they remain impressed with the performance of I HP CEO Mark Hurd, who joined the company just over a year ago, succeeding  ousted CEO Carly Fiorina.     H "Since Hurd came aboard a year ago, it seems the company finally has itsI mojo back and is coming out with new products and an innovative marketing J campaign," said Samir Bhavnani, director of research for Current Analysis,= a technology research firm. "It's almost like a new company."     H HP (down $0.52 to $31.11, Research) reported earnings, excluding certainF charges, of 54 cents a share in the fiscal second quarter, up from proG forma earnings of 37 cents in the year-earlier quarter and beating Wall 2 Street analysts' expectations of 49 cents a share.    K Including charges, HP reported earnings of 51 cents a share, up from the 33 B cents a share, including charges, that HP reported a year earlier.    I The company reported $22.6 billion in sales for the quarter, in line with 9 analysts' estimates and up 5 percent from a year earlier.     G HP also issued a better than expected forecast for the current quarter, D which is a seasonally slower quarter for the company than the secondF quarter. HP said it expects earnings, excluding certain charges, of 45J cents to 48 cents a share, which is better than analyst expectations of 43J cents a share. It forecast sales for the current quarter of $21.7 billion,! in line with analysts' estimates.     G For the full fiscal 2006, HP expects sales of $91 billion, also in line F with estimates. HP said it expects fiscal 2006 earnings, minus certainK charges, of between $2.04 and $2.08 per share, which also beats the analyst & consensus forecast of $1.96 per share.    H The company announced that roughly 1,600 jobs were eliminated during theH quarter, bringing the total to about 8,100 jobs cut since Hurd announcedE plans to eliminate 15,300 jobs as part of his plan to restructure HP.     G "We posted solid revenue growth, controlled cost and expenses, expanded J margins and generated cash flow," said Hurd of the quarter on a conference call to analysts.     K But he acknowledged the company needs to grow revenue, which he plans to do B in part by increasing the number of products HP sells to corporateJ customers. Hurd said the company plans to add to its sales staff, with newC hires potentially numbering "in the hundreds" over the next several 	 quarters.     /                            Strong growth in PCs     I HP has gained market share and increased profitability in the division at I the same time rival Dell has been losing share and grappling with slowing I growth. HP's share of the PC market rose 22.3 percent during the quarter, K while Dell's dropped 10.2 percent, according to data from industry research 
 firm Gartner.     J Dell, which reports its earnings Thursday, issued a warning last week thatH its sales and earnings for its most recent quarter will fall short of an earlier forecast.     K "Given Dell's pre-announcement, we expected things to maybe be a little bit J soft, so we're pleasantly surprised" at HP's results, said Kim Caughey, anI analyst at Fort Pitt Capital Group. Caughey said she is also pleased with E the company's operating margin of 7.3 percent for the quarter, up 1.9 . percentage points from the prior-year quarter.    H The company's PC division reported $7 billion in revenue on a 16 percentJ increase in shipments, up 10 percent from the prior-year quarter, and a 69J percent increase in its operating profit. The strength in that division isG based largely on a 27 percent increase in laptop sales. Desktop revenue K grew just 1 percent year over year. The division posted an operating profit > of 3.6 percent of revenue, up from 2.3 percent a year earlier.    F Hurd said the PC division, which accounts for about 30 percent of HP'sF business, has been gradually improving for several quarters, a fact he0 believes is sometimes overlooked by Wall Street.    J "The performance has been a prolonged, sustained march of improvement thatI pre-dates me," he said during a conference call. "It's been a very steady * rise, and it just hasn't been recognized."    K HP's most profitable division, the printing business, reported $6.7 billion I in sales, up 5 percent year over year, fueled by a 10 percent increase in C sales of its printing supplies. Revenue from printers for corporate G customers grew 4 percent, but sales of printers to consumers declined 8 I percent. The division posted an operating profit of 15.5 percent, up from " 12.7 of revenue in the prior year.    K The company's enterprise storage and servers business reported $4.3 billion 7 in revenue, up 2 percent from the year-earlier quarter.    ------------------------------  % Date: Wed, 17 May 2006 07:07:29 -0700 # From: "Tom Linden" <tom@kednos.com> , Subject: Re: Thundering Hurd pacing HP gains) Message-ID: <op.s9o76rrnzgicya@hyrrokkin>   C On Wed, 17 May 2006 06:55:16 -0700, <norm.raphael@metso.com> wrote:   G > The company's enterprise storage and servers business reported $4.3   	 > billion 9 > in revenue, up 2 percent from the year-earlier quarter.   E  From which I guess you can cocnlude that VMS represents about 25% of J that unit.  Be interesting to see the gross profit of the different units.   ------------------------------  % Date: Wed, 17 May 2006 09:32:20 -0400 # From: sol gongola <sol@adldata.com> 8 Subject: Re: what should Bad-Clients: in SMTP.CONFIG do?0 Message-ID: <1147872767.364800@nntp.acecape.com>  / Phillip Helbig---remove CLOTHES to reply wrote:  > I've put   >  > Bad-Clients: 211.213.103.30  > G > in SMTP.CONFIG.  However, I don't see any change in behaviour.  I've  C > done TCPIP STOP|START MAIL, TCPIP DISABLE|ENABLE SERVICE SMTP and # > @TCPIP$SMTP_SHUTDOWN|STARTUP.COM.  > J > Do I have to stop and restart TCPIP?  (Other changes in SMTP.CONFIG are ' > picked up without having to do this.)  > G > What should the Bad-Clients: entry actually do?  I'm getting a lot of I > spam from the address above to a domain I don't even accept email for.  H > However, the sender appears to "exist" in that bounces are delivered. H > It would be nice to not send the bounces but still be notified (OPCOM,I > logfile entry) that mail from a Bad-Clients: address was "received" and  > appropriate action taken.   E The client is the relay/system that makes a connection to your system ? via port 25. It is the machine that passes the email on to your % system, possibly from somewhere else.   > You want to put a client in your bad client list when you knowA spam comes to you through that system. However, it may not be the G source of the spam, just the machine that actually delivered it to you.   ? This works only when the email is coming directly into your vms ? smtp server. If you have an smtp proxy/firewall in front of you = then this may not work since the only email client connecting  to you is that proxy.   A If you look at your email headers, the latest "Received:" line is , the one you would put into your client list.  G You probably want to use reject-mail-from to drop the offending emails.   G > I've been experimenting mostly with TCPIP 5.4 on ALPHA, since it has  C > more anti-spam features than 5.3 on VAX (in particular, I can set D > Symbiont-Checks-Deliverability to FALSE).  With 5.4, this is what  > TCPIP$SMTP_COMMON looks like:  > P >    "TCPIP$SMTP_COMMON" [exec] = "SYS$SPECIFIC:[TCPIP$SMTP]" (LNM$SYSTEM_TABLE)  >         = "CLUSTER_TCPIP_SMTP"R > 1  "CLUSTER_TCPIP_SMTP" [exec] = "CLUSTER_DISK:[SYSTEM.TCPIP_SMTP]" (LNM$SYSTEM_ > < >    "CLUSTER_DISK" [exec] = "DISK$USER:" (LNM$SYSTEM_TABLE)K > 1  "DISK$USER" [exec] = "DSA510:" [concealed,terminal] (LNM$SYSTEM_TABLE)  > G > This seems to work OK.  (Beware that shutting down SMTP will deassign I > this logical, so it needs to be redefined before starting up SMTP again / > since otherwise one will get just the default A > SYS$SPECIFIC:[TCPIP$SMTP].)  With 5.3, however, it appears that J > SMTP.CONFIG must be in SYS$SPECIFIC:[TCPIP$SMTP] and not somewhere else J > in the search list (or perhaps it has to be in the first translation of - > the search list).  Can anyone confirm this?  >    ------------------------------  % Date: Wed, 17 May 2006 10:27:50 -0400 # From: sol gongola <sol@adldata.com> 8 Subject: Re: what should Bad-Clients: in SMTP.CONFIG do?/ Message-ID: <1147876098.23780@nntp.acecape.com>   / Phillip Helbig---remove CLOTHES to reply wrote: > > In article <1147872767.364800@nntp.acecape.com>, sol gongola > <sol@adldata.com> writes:  > H >> The client is the relay/system that makes a connection to your systemB >> via port 25. It is the machine that passes the email on to your( >> system, possibly from somewhere else. >  > Right. > A >> You want to put a client in your bad client list when you know D >> spam comes to you through that system. However, it may not be theJ >> source of the spam, just the machine that actually delivered it to you. >  > Right. > B >> This works only when the email is coming directly into your vms >> smtp server.  >  > Yes, that's the case here. > D >> If you look at your email headers, the latest "Received:" line is/ >> the one you would put into your client list.  >  > Right. > J >> You probably want to use reject-mail-from to drop the offending emails. > > > The (alleged) sender is always different, but the IP addressG > (211.213.103.30) is always the same.  I get 50--100 spams a day from  H > this address, most to a domain I don't even accept mail for.  I don't I > mind rejecting all TCPIP traffic from this address as soon as possible.  >  . . Putting in to the bad client list should work.0 try 211.213.103.0/24 (you never know what works)S If you want to block everything from there you could do it in your router/firewall.   " Sorry for the usual obvious stuff:$ Are you sure the file is being used?: Are you sure the syntax is right for the smtp.config file? Is the file in the right place? ' Did you check your logs for any errors? / There are vms logicals to expand your smtp logs A to see if there are any errors when reading the smtp.config file.    other than that:6 http://whois.nic.or.kr/ reports 211.213.103.30 (korea). has a place to report abuse:  abuse@hanaro.com/ and a few other email addresses you can get to.    ------------------------------   End of INFO-VAX 2006.273 ************************