1 INFO-VAX	Tue, 19 Oct 2004	Volume 2004 : Issue 581       Contents:" Re: Alpha COBOL Debugger anomolies" Re: Alpha COBOL Debugger anomolies Re: AlphaStation 200 4/2331 Re: CC /MMS_DEPENDENCIES and related difficulties * Re: Cluster Ethernet Interconnect question* Re: Cluster Ethernet Interconnect question* Re: Cluster Ethernet Interconnect question DIAGNOSE failing on VMS Alpha % Re: DS10L SCSI card - recommendations  Re: ECU disk Re: ECU disk Re: ECU disk- How can I send an HTTP get request using DCL? 1 Re: How can I send an HTTP get request using DCL? 1 Re: How can I send an HTTP get request using DCL? 1 Re: How can I send an HTTP get request using DCL? 1 Re: How can I send an HTTP get request using DCL? % Re: Humor   was Re: VMS for the blind  installing Oracle on openVMS  Re: installing Oracle on openVMS  Re: installing Oracle on openVMS Re: J F on why USA is a regime Re: J F on why USA is a regime# Re: Latest on Windoze Navy software # Re: Latest on Windoze Navy software # Re: Latest on Windoze Navy software # Re: Latest on Windoze Navy software # Re: Latest on Windoze Navy software # Re: Latest on Windoze Navy software # Re: Latest on Windoze Navy software # Re: Latest on Windoze Navy software # Re: Latest on Windoze Navy software * Re: Need LK461 keyboard and 3 button mouse OT: Short Article about Sun  Re: pthreads and scheduling  Re: pthreads and scheduling  Re: pthreads and scheduling  Re: Question about Descriptors Re: Question about Descriptors Re: Question about Descriptors Re: Question about Descriptors Re: Question about Descriptors Re: Question about Descriptors Re: Question about Descriptors Re: Question about Descriptors Re: Question about Descriptors Re: Question about Descriptors Re: Question about Descriptors Re: Question about Descriptors Re: Question about Descriptors Re: Question about Descriptors Re: Question about Descriptors> Re: Routine to Resubmit print/batch entries from queue listing Re: RZ28B-VA drives wanted0 Re: Simulating monochrome to test Xwindows app ?" Re: SYSUAF => export passwords ???" Re: SYSUAF => export passwords ???" Re: SYSUAF => export passwords ???& Re: V6 OpenVMS Freeware CD's questions; Re: VTfm - OpenVMS file manager for VT-compatible terminals   F ----------------------------------------------------------------------  % Date: Tue, 19 Oct 2004 10:57:02 +0200 ' From: "Usenet" <luc.beaumet@banksys.be> + Subject: Re: Alpha COBOL Debugger anomolies 6 Message-ID: <4174d6e9$0$15727$ba620e4c@news.skynet.be>   Please use the new version :  Image Identification Information   image name: "DCOBOL"  , image file identification: "COBOL T2.8-1373"  # image file build identification: ""   ' link date/time: 17-SEP-2004 17:21:19.59    linker identification: "A11-20"     ; "Tom McWilliam" <mcwilliamt@aforbes.co.za> wrote in message 7 news:34f95bea.0410130048.653b4c48@posting.google.com... E > It was with great interest that I read a posting to this group from F > 1996 entitled "AXP program crashes vs. VAX crashes: Traceback info".: > We are currently experiencing the very same problem! :-) > C >     We are running a large COBOL based legacy application and are D > having many problems with the COBOL Debugger. We are running underH > 7.3-2 of the operating system, Compaq COBOL V2.8-1286 for the compilerF > and OpenVMS Alpha Debug64 version 7.3-200 for the debugger. The mostD > persistent problem is that programs do not always execute with the> > same results in debug as they do when run without the debug.D >     The problems currently encountered, started with the migrationF > from a VAX to the new hardware platform (ES47), which also coincidedD > with upgrading to the newer, supported, version of COBOL. I am notF > aware of any program or compile script changes that they should have4 > made when migrating from the Vax to the Alpha box.H >     As with the original posting, "The problem is that when crashes ofD > this kind occur on the VAX I get the exact line number and routineE > name of the error in the traceback info so it's easy to correct. On H > the AXP this is not as simple, I get traceback info, routine names andG > line numbers but then find that the error didn't occur there but at a F > completely different location. The only way to find that location isC > to step through the program in the debugger ......... and when it G > suddenly jumps to a totally unexpected location somewhere else in the 3 > program I know I have found what I'm looking for.  > G > This is not exactly a lot of fun to do and takes an awful lot of time F > compared to the VAX where I was always given the correct location ofA > the error and could correct it immediately instead of going and H > stepping through dozens, hundreds or, in the worst cases, thousands of! > lines of code in the debugger."  > / > I have already checked to see that all of the = > responses/recommendations to the original posting have been . > implemented, yet the problem still persists. > B > Surely the fixes applied to the comilers for the debug/tracebackF > information in the late 90's would have been applied the the current > versions of the compilers? > 0 > Any recommendations would be well appreciated. > 
 > Regards, >  > Tom.   ------------------------------  % Date: Tue, 19 Oct 2004 11:09:43 +0200 ' From: "Usenet" <luc.beaumet@banksys.be> + Subject: Re: Alpha COBOL Debugger anomolies 6 Message-ID: <4174da0a$0$19549$ba620e4c@news.skynet.be>  
 More info:  L there is a bug in the "jam". Use optimizer at max level 2 ( not DEFAULT ) or ask HP for the fix.    Luc Beaumet   2 "Usenet" <luc.beaumet@banksys.be> wrote in message0 news:4174d6e9$0$15727$ba620e4c@news.skynet.be... > Please use the new version :" > Image Identification Information >  > image name: "DCOBOL" > . > image file identification: "COBOL T2.8-1373" > % > image file build identification: ""  > ) > link date/time: 17-SEP-2004 17:21:19.59  > ! > linker identification: "A11-20"  >  > = > "Tom McWilliam" <mcwilliamt@aforbes.co.za> wrote in message 9 > news:34f95bea.0410130048.653b4c48@posting.google.com... G > > It was with great interest that I read a posting to this group from H > > 1996 entitled "AXP program crashes vs. VAX crashes: Traceback info".< > > We are currently experiencing the very same problem! :-) > > E > >     We are running a large COBOL based legacy application and are F > > having many problems with the COBOL Debugger. We are running underJ > > 7.3-2 of the operating system, Compaq COBOL V2.8-1286 for the compilerH > > and OpenVMS Alpha Debug64 version 7.3-200 for the debugger. The mostF > > persistent problem is that programs do not always execute with the@ > > same results in debug as they do when run without the debug.F > >     The problems currently encountered, started with the migrationH > > from a VAX to the new hardware platform (ES47), which also coincidedF > > with upgrading to the newer, supported, version of COBOL. I am notH > > aware of any program or compile script changes that they should have6 > > made when migrating from the Vax to the Alpha box.J > >     As with the original posting, "The problem is that when crashes ofF > > this kind occur on the VAX I get the exact line number and routineG > > name of the error in the traceback info so it's easy to correct. On J > > the AXP this is not as simple, I get traceback info, routine names andI > > line numbers but then find that the error didn't occur there but at a H > > completely different location. The only way to find that location isE > > to step through the program in the debugger ......... and when it I > > suddenly jumps to a totally unexpected location somewhere else in the 5 > > program I know I have found what I'm looking for.  > > I > > This is not exactly a lot of fun to do and takes an awful lot of time H > > compared to the VAX where I was always given the correct location ofC > > the error and could correct it immediately instead of going and J > > stepping through dozens, hundreds or, in the worst cases, thousands of# > > lines of code in the debugger."  > > 1 > > I have already checked to see that all of the ? > > responses/recommendations to the original posting have been 0 > > implemented, yet the problem still persists. > > D > > Surely the fixes applied to the comilers for the debug/tracebackH > > information in the late 90's would have been applied the the current > > versions of the compilers? > > 2 > > Any recommendations would be well appreciated. > >  > > Regards, > >  > > Tom. >  >    ------------------------------  % Date: Tue, 19 Oct 2004 05:35:51 -0700 , From: James Nykiel <jnykiel@access4less.net># Subject: Re: AlphaStation 200 4/233 < Message-ID: <pan.2004.10.19.12.35.50.696227@access4less.net>  
 Greetings:    G Thank you all for your assistance, I have obtained the AlphaStation 200 8 4/233 that I was looking for.  Please close this thread.     Thanks - Jim  7 On Thu, 30 Sep 2004 11:11:50 -0700, James Nykiel wrote:    > Greetings: >  > G > I am interested in purchasing a AlphaStation 200 4/233.  I would like E > the unit to have both a floppy and cdrom drive if possible but will C > settle for unit with a cdrom drive only.  I am also interested in H > obtaining DEC keyboard (LK461-AA) and 3 button mouse for use with this > unit.  > I > I would like a unit in fairly decent cosmetic condition if possible.  I D > have checked eBay but have not seen what I am looking for to date. > E > Let me know what you have and what you are asking.  I am located in  > central Oregon.  >  >  > Thank you... - James Nykiel    ------------------------------   Date: 19 Oct 2004 06:08:50 GMT2 From: "Dave Weatherall" <djw-nothere@nospam.nohow>: Subject: Re: CC /MMS_DEPENDENCIES and related difficulties? Message-ID: <DTiotGxQ0bj6-pn2-OlmZq2Y2HEFl@dave2_os2.home.ours>   9 On Mon, 18 Oct 2004 16:07:36 UTC, sms@antinode.org wrote:   = > From: koehler@eisner.nospam.encompasserve.org (Bob Koehler)  > G > >    I never did have a good reason to put object files in a separate H > >    directory.  I've seen lots of folks do this, but I think its justI > >    because they saw someone else do it.  For me it's good enough that / > >    I can tell them appart by the file type.  > G >    I don't recall seeing anyone else do it.  (I thought it was my own G > brilliant idea, but I don't claim priority.)  My separate directories J > are named [.ALPHA], [.IA64], [.VAX], and so on.  (You can probably guessD > why.)  I find this more reliable than distinguishing the differentF > object files by version numbers, and more convenient than using fileH > types like .AXP_OBJ, and the like.  (Not that CC /MMS is currently anyD > more or less friendly to that naming scheme than it is to separate > object file directories...)   F I do it too! [.VAX] amd [.AXP]. I'm not likely to get to [.I64]. OTOH @ I rarely  link to individual object modules anymore. In all our 8 projects, they get placed in an object library, also in E [.'host_architecture'], and get linked from there. Unlike Bob, I get   VAX running properly first.    --   Cheers - Dave W.   ------------------------------  % Date: Tue, 19 Oct 2004 08:54:00 +0200 2 From: Karl Rohwedder <emil.mustermann@t-online.de>3 Subject: Re: Cluster Ethernet Interconnect question ) Message-ID: <cl2dlq$etu4@doiweb4.b2x.vwg>    James Nykiel wrote:  > Greetings: >  > H > If I want to create a cluster with two AlphaStations using an Ethernet? > Interconnect do I need to have two Ethernet adapters in each?  > J > I have been looking through the documents ovms_73_cluster_config.pdf andL > ovms_731_cluster_systems.pdf but still do not understand alot of what I amF > reading so please excuse me if I have looked over something obvious. >  > Thanks - Jim  # One adapter per node is sufficient.    ------------------------------    Date: 19 Oct 2004 07:34:58 -0500; From: koehler@eisner.nospam.encompasserve.org (Bob Koehler) 3 Subject: Re: Cluster Ethernet Interconnect question 3 Message-ID: <SjbiQOlsc3Dw@eisner.encompasserve.org>   k In article <pan.2004.10.19.05.13.31.544140@access4less.net>, James Nykiel <jnykiel@access4less.net> writes:  > Greetings: >  > H > If I want to create a cluster with two AlphaStations using an Ethernet? > Interconnect do I need to have two Ethernet adapters in each?   ?    No.  You can run many protocols simultaneously over a single C    ethernet.  There is no need for second adapters in such a simple 	    setup.    ------------------------------  # Date: Tue, 19 Oct 2004 14:05:09 GMT 1 From: Keith Parris <keithparris_NOSPAM@yahoo.com> 3 Subject: Re: Cluster Ethernet Interconnect question 1 Message-ID: <pa9dd.1121$Q65.901@news.cpqcorp.net>    James Nykiel wrote: H > If I want to create a cluster with two AlphaStations using an Ethernet? > Interconnect do I need to have two Ethernet adapters in each?   H The only reason for multiple Ethernet adapters per node would be if you B wish the cluster to continue operating despite the failure of one.   ------------------------------  % Date: Tue, 19 Oct 2004 12:47:24 -0400 $ From: Mike Duffy <Duffy@process.com>& Subject: DIAGNOSE failing on VMS AlphaJ Message-ID: <63D30D6E10CFD11190A90000F805FE8605AED22B@lespaul.process.com>  " OpenVMS V7.3 on AlphaServer DS10L.  < Every now and then, DIAGNOSE fails in the following fashion:  2 $ diagnose sys$errorlog:errlog.sys /out=errlog.out  
 DECevent V3.4 ; %CMA-F-EXCCOP, exception raised; VMS condition code follows B -SYSTEM-F-STKOVF, stack overflow, PC=00000000002D9D70, PS=0000001BH %CMA-F-EXCEPTION, exception raised; address of exception object 003B6800  ; Starting with a new errlog.sys fixes it for a month or two.   B I suppose what I want to do is increase KSTACKPAGES (currently 2.); But since (if all goes well) I won't be rebooting for about > six months, I'd like to ask here before finding out next year  that I changed the wrong thing.   ; Is KSTACKPAGES the appropriate thing to fix?  If so, by how  much?    Thanks, 
 Mike Duffy   ------------------------------  % Date: Tue, 19 Oct 2004 11:50:07 -0400 , From: "warren sander" <warren.sander@hp.com>. Subject: Re: DS10L SCSI card - recommendations, Message-ID: <41753913$1@usenet01.boi.hp.com>  ; how about going to the supported options page for the ds10l J http://h18002.www1.hp.com/alphaserver/options/asds10l/asds10l_options.htmlH search for 'storage: Adapter - SCSI' in the subcatagory list and you get this:     G       AlphaServer DS10L Supported Options List Minimum Operating System  Version        VMS UNIX LINUX6       Part No. Description Rst Ver Max Ver Max Ver Max        Storage: Adapter - SCSIL       3X-KZPBA-CC PCI-X TO SCSI UWD (FAST20 DIFFERENTIAL) ADAPTER  Yes 7.2-1 1 4.0F 1 SPT CD L       3X-KZPCA-AA MODULE, ADAPTER, ULTRA 2(LVD), 1 CHANNEL  Yes 7.2-1 1 4.0F 1 SPT CDJ       3X-KZPEA-DB LVD Multimode U3 PCI SCSI Adapter, 2 Channel, VHDCI  Yes 7.2-2 1 5.1 1 N/S 0 L       KZPBA-CA SINGLE CHNL. PCI TO ULTRASCSI ADAPTER, UWSE  Yes 7.2-1 1 4.0F 1 SPT CD       KZPBA-CBK       (400577-001) PCI TO SCSI UWD (FAST20 DIFFERENTIAL) ADAPTER  Yes 7.2-1  1 4.0F 1 SPT CD K       SN-KZPBA-CA QLOGIC PCI TO FASTWIDE SCSI HOST BUS ADAPTER  Yes 7.2-1 1 
 4.0F 1 SPT CD               . "Gremlin" <not-here@all.mate> wrote in message' news:4170928a@duster.adelaide.on.net... L > Rather than using IDE (perhaps slow as in an earlier thread), could I haveE > recommendations for a SCSI card for a 36/72Gb disk - third party if  possibleK > as I want to get one in the next day or so and can't access "real" Compaq  > cards.  Adaptec? >  > Thanks >  >    ------------------------------    Date: 19 Oct 2004 10:04:42 -0500- From: Kilgallen@SpamCop.net (Larry Kilgallen)  Subject: Re: ECU disk 3 Message-ID: <UWGjAa7xXsCi@eisner.encompasserve.org>   ` In article <cl1i0j$n74$2@hercules.btinternet.com>, "Rob H" <robert.heyes@btinternet.com> writes:M > Where can I obtain an EISA (ECU) disk for an Alpha Server 1000? Can anyone   > send me a link or the files?  0 That last I knew, that was copyrighted software.  G Those of us who make a living from copyrighted software are unlikely to  be giving away illicit copies.   ------------------------------  + Date: Tue, 19 Oct 2004 15:24:57 +0000 (UTC) , From: lewis@PROBE.MITRE.ORG (Keith A. Lewis) Subject: Re: ECU disk . Message-ID: <cl3bk9$eis$1@newslocal.mitre.org>   "Rob H" <robert.heyes@btinternet.com> writes in article <cl1i0j$n74$2@hercules.btinternet.com> dated Mon, 18 Oct 2004 23:01:39 +0000 (UTC): L >Where can I obtain an EISA (ECU) disk for an Alpha Server 1000? Can anyone  >send me a link or the files?   H Me too!  Except I need one for an Alphastation 500/400, specifically the audio device file.  0 --Keith Lewis              klewis {at} mitre.org> The above may not (yet) represent the opinions of my employer.   ------------------------------  + Date: Tue, 19 Oct 2004 15:28:31 +0000 (UTC) , From: lewis@PROBE.MITRE.ORG (Keith A. Lewis) Subject: Re: ECU disk . Message-ID: <cl3bqv$eis$2@newslocal.mitre.org>   Kilgallen@SpamCop.net (Larry Kilgallen) writes in article <UWGjAa7xXsCi@eisner.encompasserve.org> dated 19 Oct 2004 10:04:42 -0500: a >In article <cl1i0j$n74$2@hercules.btinternet.com>, "Rob H" <robert.heyes@btinternet.com> writes: N >> Where can I obtain an EISA (ECU) disk for an Alpha Server 1000? Can anyone  >> send me a link or the files?  > 1 >That last I knew, that was copyrighted software.  > H >Those of us who make a living from copyrighted software are unlikely to >be giving away illicit copies.   F Larry, didn't each and every Alpha with EISA come with an ECU disk?  IL don't think it's nice to expect somebody to re-pay a license fee because the5 disk was lost (or wore out) somewhere along the line.   0 --Keith Lewis              klewis {at} mitre.org> The above may not (yet) represent the opinions of my employer.   ------------------------------    Date: 19 Oct 2004 04:31:06 -07005 From: stuart.symonds@commerzbank.com (Stuart Symonds) 6 Subject: How can I send an HTTP get request using DCL?= Message-ID: <90d46f26.0410190331.7b679a9a@posting.google.com>   E I wish to send an HTTP type request to a web server via port 80 using A DCL, I have been advised by our web server administrator that the D simplest way to do this is by using telnet. When I telnet to the webA server via port 80 and send the command interactively it works as @ expected, if a try to do the same in a DCL procedure it hangs orF aborts when making the telnet connection. The hang or abort depends on whether I redefine sys$command.   B Probably very trivial for most of you but would greatly appreciate some assistance.   Thanks,  Stuart.    ------------------------------    Date: 19 Oct 2004 06:57:18 -0500B From: clubley@remove_me.eisner.decus.org-Earth.UFP (Simon Clubley): Subject: Re: How can I send an HTTP get request using DCL?3 Message-ID: <5l$mrthqkFRY@eisner.encompasserve.org>   u In article <90d46f26.0410190331.7b679a9a@posting.google.com>, stuart.symonds@commerzbank.com (Stuart Symonds) writes: G > I wish to send an HTTP type request to a web server via port 80 using C > DCL, I have been advised by our web server administrator that the F > simplest way to do this is by using telnet. When I telnet to the webC > server via port 80 and send the command interactively it works as B > expected, if a try to do the same in a DCL procedure it hangs orH > aborts when making the telnet connection. The hang or abort depends on! > whether I redefine sys$command.  >   6 You cannot take telnet input from a command procedure.  H I believe that wget is available for VMS. Have a look at that and see if it does what you want.   Which stack are you using ?    Simon.   --  B Simon Clubley, clubley@remove_me.eisner.decus.org-Earth.UFP       7 Microsoft: The Standard Oil Company of the 21st century    ------------------------------    Date: 19 Oct 2004 07:17:52 -0500 From: briggs@encompasserve.org: Subject: Re: How can I send an HTTP get request using DCL?3 Message-ID: <f7Z20kgl9AeB@eisner.encompasserve.org>   u In article <90d46f26.0410190331.7b679a9a@posting.google.com>, stuart.symonds@commerzbank.com (Stuart Symonds) writes: G > I wish to send an HTTP type request to a web server via port 80 using C > DCL, I have been advised by our web server administrator that the F > simplest way to do this is by using telnet. When I telnet to the webC > server via port 80 and send the command interactively it works as B > expected, if a try to do the same in a DCL procedure it hangs orH > aborts when making the telnet connection. The hang or abort depends on! > whether I redefine sys$command.   M I use C-Kermit for this kind of thing.  You can script it with, for instance:   3 output {AT\13}			(Send "AT" command to Hayes modem) 7 input 5 OK			(Wait 5 seconds looking for "OK" response)  if SUCCESS goto wherever) echo { Failed to communicate with modem }  goto hangup   L It works with telnet as well as serial ports.  Get the version that supports your TCP stack.      	John Briggs   ------------------------------  + Date: Tue, 19 Oct 2004 12:17:44 +0000 (UTC) = From: jlw@psulias.psu.edu (j.lance wilkinson, (814) 865-1818) : Subject: Re: How can I send an HTTP get request using DCL?- Message-ID: <cl30l8$fq2$1@f04n12.cac.psu.edu>   u In article <90d46f26.0410190331.7b679a9a@posting.google.com>, stuart.symonds@commerzbank.com (Stuart Symonds) writes: F >I wish to send an HTTP type request to a web server via port 80 usingB >DCL, I have been advised by our web server administrator that theE >simplest way to do this is by using telnet. When I telnet to the web B >server via port 80 and send the command interactively it works asA >expected, if a try to do the same in a DCL procedure it hangs or G >aborts when making the telnet connection. The hang or abort depends on   >whether I redefine sys$command. > C >Probably very trivial for most of you but would greatly appreciate  >some assistance.   D 	Use cURL (found at http://curl.haxx.se/ ).  A VMS port exists and I? 	use it ALL THE TIME to do just what you're asking in Batch DCL D 	procedures and DCL CGI scripts.  Works like a charm, very powerful,> 	deals with SSL, authentication, everything I've thrown at it.  P +----"Never Underestimate the bandwidth of a station wagon full of mag tapes"--+B | J.Lance Wilkinson ("Lance")		InterNet:  Lance.Wilkinson@psu.edu < | Systems Design Specialist - Lead	AT&T:      (814) 865-18189 | Digital Library Technologies		FAX:       (814) 863-3560 2 | 3 Paterno Library				"I'd rather be dancing..." B | Penn State University		    A host is a host from coast to coast,K | University Park, PA 16802	    And no one will talk to a host that's close C | <postmaster@psulias.psu.edu>	    Unless the host that isn't close : | EMail Professional since 1978	    Is busy, hung or dead.P +---------"He's dead, Jim. I'll get his tricorder. You take his wallet."-------+9                 [apologies to DeForest Kelley, 1920-1999] 3 <A Href="http://perdita.lcs.psu.edu">home page</a>  J <a Href="http://perdita.lcs.psu.edu/junkdec.htm">junk mail declaration</a> --	/"\ 	\ /	ASCII RIBBON CAMPAIGN 	 X	AGAINST HTML MAIL  	/ \   ------------------------------  % Date: Tue, 19 Oct 2004 08:46:40 -0400 - From: JF Mezei <jfmezei.spamnot@teksavvy.com> : Subject: Re: How can I send an HTTP get request using DCL?, Message-ID: <41750CAF.FB1CBEDE@teksavvy.com>   Stuart Symonds wrote:  > G > I wish to send an HTTP type request to a web server via port 80 using  > DCL   L Telnet won't work. but Kermit will work. Kermit can access logical names and DCL symbols.  I Here is a sample kermit script I used once: (reads stuff from a text file K (logical name $TRANSACT) and sends transactions to a remote (very remote in N this case) server and logs answer to a file. Note that instead of CONNECT, youE can use an INPUT statement with some option to end the input when the = connection is closed, but I don't recall off hand what it is.    SET TERMINAL BYTESIZE 8  SET COMMAND  BYTESIZE 8  SET FILE CHARACTER-SET LATIN1 & SET TRANSFER CHARACTER-SET TRANSPARENT( SET TERMINAL CHARACTER-SET LATIN1 LATIN1 SET BUFFER 20000 20000 SET NETWORK TCP  LOG SESSION \$(TRANSLOG) open read \$(TRANSACT) ;  ;  ;************* :loop  ;************* READ \%a ECHO ****\%a*****\13 if fail goto endloop SET HOST www.erin.gov.au:80  if failure end 1 cannot connect  SET TERMINAL ECHO ON ; ( OUTPUT POST /cgi-bin/ERIS.pl HTTP/1.0\13? OUTPUT Referer: http://www.erin.gov.au/database/MAN200R.html\13   OUTPUT Connection: Keep-Alive\135 OUTPUT User-Agent: Mozilla/3.01(Macintosh; I; PPC)\13  OUTPUT Host: www.erin.gov.au\13 J OUTPUT Accept: image/gif, image/x-xbitmap, image/jpeg, image/pjpeg, */*\13' OUTPUT Accept-Language: en-GB, fr-CA\13 9 OUTPUT Content-type: application/x-www-form-urlencoded\13 ' OUTPUT Content-length: \Flength(\%a)\13 
 OUTPUT \13
 OUTPUT \%a\13 4 ;OUTPUT report=MAN200R&name=ab&feature=ALL&region=WA CONNECT 	 GOTO loop  ;  :endloop
 CLOSE read HANGUP EXIT        L Other options are to use simple utilities such as FETCH_HTTP (comes with OSUN web server), or CURL  (use google to search this newsgroup for CURL and you'llG find where the latest version is located). CURL is a much more powerful  version of FETCH_HTTP.   ------------------------------  % Date: Tue, 19 Oct 2004 08:38:56 -0400 # From: "John Smith" <a@nonymous.com> . Subject: Re: Humor   was Re: VMS for the blind, Message-ID: <XcudnZu4-qVHl-jcRVn-qg@igs.net>   David J Dachtera wrote:  > John Smith wrote:  >>E >> When I saw the title of this thread "VMS for the blind", the first  >> thing that came to mind was:  >>F >> Yes!!! Somebody has finally created a presentation for HP executiveD >> management about VMS at a simple enough level that even carly(tm)H >> could understand - music, sing-along, follow the bouncing ball, ..... > * > Purple dinosaur? ...or is that too much?  J I recall that there was a time about 10 years ago when tourists were beingI car-jacked and shot at an alarming rate in Florida that the standing joke  was:  ( Q. What do you get Barney for Christmas?F A. One-way ticket to Florida and a t-shirt that says "I am a tourist'.  2 Perhaps carly(tm) would use the same gift package.              A >> Imagine my chagrin when I discovered that while the thread was D >> actually about a worthy and important use of VMS, it wasn't about, >> changing HP management perception of VMS. > E > I'm certain it's a deeply engrained culture and will not die except ! > with those who perpetuate it...     / Ostriches, but with their heads up their 'six'.    ------------------------------  # Date: Tue, 19 Oct 2004 13:12:47 GMT ) From: simon@stirley.me.uk (Simon Stirley) % Subject: installing Oracle on openVMS 2 Message-ID: <417512bf.1822545441@news.demon.co.uk>   Hi,   ? I've got an old DEC Alpha that I've installed openVMS (Hobbyist F license) onto, and am now trying to get Oracle setup. Everytime I comeF to run the installer (@runInstaller off Disk1) it tells me I must haveD certain permissions for the current user - but from what I can see IF have those permissions. Is this a common problem that anyone knows how to solve ? thanks.  C PS. the user I created to install under is 'sa' and has permissions ! added with AUTHORIZE as follows -  CMKRNL GROUP  WORLD  GRPNAM SYSNAM  C The 'system' account itself doesn't work as the installer won't let ; you install under that, or any account in the system group.    ------------------------------  % Date: Tue, 19 Oct 2004 16:46:24 +0200 9 From: Jan-Erik =?iso-8859-1?Q?S=F6derholm?= <aaa@aaa.com> ) Subject: Re: installing Oracle on openVMS ' Message-ID: <417528C0.F4AB2A1B@aaa.com>    Hi. ? You don't say if you have a specific need for Oracle (Classic), @ but note that you are allowed to run Oralce Rdb (which is a much: better database mgmt system, of course :-) ) for free in a= non-profit environment. Well, even fr "development-only", if  I'm not wrong.  @ B.t.w, "sa" (for "system administator"), isn't that the usernameB that was used in Sybase (and of course now in MS SQL Server, since0 that's nothing but the old "Sybase for NT"...) ?  	 Jan-Erik.    Simon Stirley wrote: >  > Hi,  > A > I've got an old DEC Alpha that I've installed openVMS (Hobbyist H > license) onto, and am now trying to get Oracle setup. Everytime I comeH > to run the installer (@runInstaller off Disk1) it tells me I must haveF > certain permissions for the current user - but from what I can see IH > have those permissions. Is this a common problem that anyone knows how > to solve ? thanks. > E > PS. the user I created to install under is 'sa' and has permissions # > added with AUTHORIZE as follows -  > CMKRNL > GROUP  > WORLD  > GRPNAM > SYSNAM > E > The 'system' account itself doesn't work as the installer won't let = > you install under that, or any account in the system group.    ------------------------------    Date: 19 Oct 2004 08:04:28 -07001 From: nothome@spammers.are.scum (Malcolm Dunnett) ) Subject: Re: installing Oracle on openVMS - Message-ID: <b71AtGHpCTwE@malvm7.mala.bc.ca.>   2 In article <417512bf.1822545441@news.demon.co.uk>,,  simon@stirley.me.uk (Simon Stirley) writes:   > A > I've got an old DEC Alpha that I've installed openVMS (Hobbyist H > license) onto, and am now trying to get Oracle setup. Everytime I comeH > to run the installer (@runInstaller off Disk1) it tells me I must haveF > certain permissions for the current user - but from what I can see IH > have those permissions. Is this a common problem that anyone knows how > to solve ? thanks. > E > PS. the user I created to install under is 'sa' and has permissions # > added with AUTHORIZE as follows -  > CMKRNL > GROUP  > WORLD  > GRPNAM > SYSNAM > H   Did you make them default privileges as well as authorized privileges?  I   If you do a "SHOW PROC/PRIV" at the $ prompt do you see the privileges?   I   if not do a "$SET PROC/PRIV=ALL" and repeat. If they show up then, they E aren't default privileges, use the MOD user/DEFPRIV=(xxx,...) command  within authorize to fix that.       ------------------------------  % Date: Tue, 19 Oct 2004 11:30:06 -0400 ( From: David Froble <davef@tsoft-inc.com>' Subject: Re: J F on why USA is a regime , Message-ID: <417532FE.9060106@tsoft-inc.com>   Bill Todd wrote:  7 > "David Froble" <davef@tsoft-inc.com> wrote in message $ > news:4172160D.503@tsoft-inc.com... >  >>Bill Todd wrote: >>/ >>It's always nice to hear fromt he far left...  >> > M > Not that one ever does around here.  I'm strictly a centrist, myself:  it's K > just that a lot of people have been listening to rabid ideologues so long / > that they have forgotten where the center is.     O Location can be based upon perspective.  So even while you are far 'left', you  ? look around and determine that you are at the center of things.     8 >>>"David Froble" <davef@tsoft-inc.com> wrote in message) >>>news:4170088A.6070901@tsoft-inc.com...  >>>  >>>... >>>  >>>  >>> E >>>>>>The day the USA began its illegal invasion of Iraq, it became a  >>>>>> >>>>>> >>>"regime". >>>  >>> J >>>>It's even more illegal and a crime against humanity to allow something >>>> >>>> >>>like  >>>  >>> I >>>>Saddam to remain in power and impose harm on those within and outside  >>>> >>>> >>>Iraq. >>> F >>>Actually, the latter is not technically illegal at all, whereas the >>> 
 > invasion >  >>>most assuredly was. >>>  >>K >>I guess it depends who is defining 'illegal'.  In "Dave's World", someone  >> > like > B >>Saddam is definitely 'illegal'.  Possible not in "Bill's World". >> > L > I hadn't realized that you had your own private fiefdom wherein you got toN > define legality.  In the world that most of the rest of us inhabit, legalityM > is defined by governmental entities, and that's the definition I was using.     P Strange.  I thought that the people, as a group, defined the government.  Well, O in a democracy and such.  With his 'lists' and 'safety' I fear a bit for those  M 'rights' and 'priviledges' you like to assume.  Regardless, when people in a  G democracy see something wrong, they have the capability of changing it.     N > But in matters like this one you don't seem particularly concerned with muchK > beyond your own prejudices, so I doubt that factual issues matter much to L > you.  Which is why people like you should be prevented (by 'whatever means< > necessary') from imposing their uninformed will on others.    O Now that's really humorous.  When I propose "whatever means necessary" you get  Q bent out of shape.  But it seems that "whatever means necessary" is quite Ok for X you, and youe 'Elk'.  :-)0    G >>>Perhaps you need to brush up on your international and United States  >>>  > law, > F >>>rather than speak from such abject ignorance while acting as if you >>>i
 > actually > I >>>knew something about the subject.  Kofi Annan has (unfortunately, overs >>>  > 18 > K >>>months too late) finally recently admitted the reality that the invasionC >>>r > waso > J >>>indeed illegal:  not only was it a violation of the U.N. charter, whichK >>>restricts military action to self-defense unless authorized by the U.N.,e >>>  > bute > C >>>by extension it was also a violation of U.S. law, since the U.S.t >>>g
 > ratified > G >>>said charter as a treaty and the Constitution (Article VI, clause 2)hE >>>specifies that treaties are part of 'the supreme law of the land'.S >>>  > N > (I just thought that bore repeating, should you care to rejoin the civilized" > world's definition of legality.) >  > ...     M Some have maintained that 'legality comes from the barrel of a gun'.  In the eO end, sadly it's true.  Where would the first ammendment to the US constitution  O be if the people of the US were not ready to defend it, with guns if necessary?z    I >>>Attitudes like yours are a significant part of the reason that a large J >>>portion of the world would stand up and cheer if Washington, D.C., wereI >>>wiped off the face of the earth - not as vengeance so much as to knockn >>>t > us > F >>>down enough pegs to make us cease being a threat to the rest of the >>>i	 > planet.s > G >>>Until we change our attitude significantly, you can count me in that  >>>i	 > number.s > 	 >>>- billu >>>n >>C >>So you would actually cheer if hundreds of thousands of people ins >> > Washington >  >>D.C. were to die?e >> > L > Yes:  the currently-popular phrase for that is 'collateral damage', and ifN > that's what it takes to stop our insanity (the condition I specified above),L > so be it.  My guess is that it might well actually minimize the total loss< > of life compared to the situation if we are *not* stopped.    N In 'Dave's World', there is no such thing as 'collateral damage'.  If 'Bill's I World' supports the idea of 'collateral damage', then I suggest that you  P practice what you preach, and be the first to experience first hand 'collateral P damage'.  No?  I didn't think so.  People who think thus, always deem it proper L for others, but not themselves.  A bit like the hyprocrisy of the terrorist O leaders.  Those cowards will recruit others for suicide attacks, but strangly, n0 they themselves never seem to make such attacks.    8 >   I guess I just have to ask, did you cheer as the WTC >  >>buildings fell?t >> > M > No:  it only became clear to me in retrospect just how justified the attackm> > had been, primarily due to the nature of our response to it.    Q So, if I punch you in the mouth, you're going to just accept the damage?  You're sL not going to attempt to stop me from doing so again?  Forgive me if I don't 
 believe that.o  P Let's declare an experiment.  Anyone reading this, who is locationally close to I Bill, go over and punch him in the mouth.  Then report back his reaction.s    @ >   They too were a symbol.  Be careful to avoid being on one of >  >>dubya's lists. >> > N > The only thing I'm at all careful about is running afoul of actual law (realK > law, not "Dave's law"...).  Using protected free speech to give the thugshJ > running our country something to fume about is not only a pleasure but aM > civic duty - and even the current reactionary Supreme Court is having a bit N > of difficulty stomaching the idea that free speech is now something that can" > only occur in specified 'zones'.    J Well, there we seem to be in agreement.  Actually, anyone who cares about P freedom should be very afraid.  Very very afraid.  I have to wonder whether any O of the world's tyrants throughout history were ever deemed to be a tyrant when t they were just getting started.e  K Freedom is more important than safety.  Anyone who wanted safety more than  Q freedom would not put themselves in harm's way to defend freedom.  Dubya's daddy bM values freedom.  Takes some balls and such to fly off a WWII carrier.  Seems tM dubya was of a different opinion, getting into the ANG and then possibly not  J fulfilling his duties there.  (Let's not get into the argument of whether # freedom was at stake at that time.)      So summing up:    L 1) Bill seems to believe in collateral damage, but possibly only for others.  : 2) Bill seems to think that if it's 'legal', then it's Ok.  I 3) Bill seems to think that the US has earned the 'right' to be attacked.d  @ Please correct me if any of the above conclusions are incorrect.     Dave   -- s4 David Froble                       Tel: 724-529-04504 Dave Froble Enterprises, Inc.      Fax: 724-529-0596> DFE Ultralights, Inc.              E-Mail: davef@tsoft-inc.com 170 Grimplin Roadt Vanderbilt, PA  15486t   ------------------------------  % Date: Tue, 19 Oct 2004 12:48:54 -0400 - From: JF Mezei <jfmezei.spamnot@teksavvy.com>a' Subject: Re: J F on why USA is a regime , Message-ID: <41754567.4736AE39@teksavvy.com>   David Froble wrote: N > Some have maintained that 'legality comes from the barrel of a gun'.  In theP > end, sadly it's true.  Where would the first ammendment to the US constitutionQ > be if the people of the US were not ready to defend it, with guns if necessary?   K In Canada, they have those not-for-profit ads on TV about famous canadians.EM One of those ads features discussions at the UN about the introduction of they? charter of human rights, which was done by a canadian diplomat.t  H One of the aspects which is outlined by this ad, and which was extremelyK important at the time is that all signataries, (and this includes the USA),"< agreed that the UN charter of rights was ABOVE any country'sM constitution/laws. It was a big deal because every country agreed that the UN K charter was sovereign over their own territory and could not be overturned.e  H This was to prevent any country from thinking it was legal to torture orB suspend due legal process by simply passing some domestic law/act.  K At the time of the founding of the UN and signing of the charter of rights,cL certain countries were considered to be mature enough to be given a right ofJ veto.  It was never considered that one of those "founding" members  couldK turn into a rogue government that totally dismisses its obligatiosn towards2M the UN charter and has suspended due legal process for a lot of people inside N its own jurisdiction, tortured people domestically and abroad etc etc etc etc.  H You should recall that the current regime in Washington has often statedN categorically that they would never allow the UN or anyone else to prevent the( USA from doing anything it wants to do. K This is absolutely contrary to the UN treaties the USA has agreed to uphold 
 and abide by.n  L So while they may not have broken the USA constitution, they have broken theL law of the land because they have broken many international alws which applyJ to the USA by the fact that the USA agreed to have them supercede any lawsB passed domestically in the USA. (same applies to all signataries).  M Therefore, the current regime has broken the law which applies inside the USAsN and thus should be impeached. This is far more serious than getting a blow job= in the oval office while on the phone with some world leader.e  N You should note that the Parti Qubcois had passed stringent language laws inM Qubec in the late 1970s, and some of those clauses went as far as the UnitedxJ Nations which struck them down and forced Canada to force Qubec to revokeG those clauses because they went against the UN charter of human rights.t   ------------------------------    Date: 19 Oct 2004 07:42:38 -0500; From: koehler@eisner.nospam.encompasserve.org (Bob Koehler)c, Subject: Re: Latest on Windoze Navy software3 Message-ID: <byLQ03yAI55o@eisner.encompasserve.org>s  Q In article <417020A2.D63B054B@adldata.com>, sol gongola <sol@adldata.com> writes:d > H > That is correct. However, of the languages I know about, none specify  > internal activity.  @    C is the exception.  While C does not demand that pointers beG    addresses, it does specify argument passing methods (relevant here: lC    arrays are passed by pointers).  C also specifies some other low ;    level details, such as the representation of true/false.c  B    Other high level languages have the good sense to be high level=    languages:  behaviour is specified but not implementation.   C    And C is not the only language that tries to replace traditional/<    assembly language needs, the others don't seem to requireG    implementation details, so it's not a requirement of the environmentd    C was designed for.   ------------------------------    Date: 19 Oct 2004 07:48:35 -0500; From: koehler@eisner.nospam.encompasserve.org (Bob Koehler)e, Subject: Re: Latest on Windoze Navy software3 Message-ID: <zK3cK$50RF2I@eisner.encompasserve.org>   Z In article <qsdcd.1004$aJ1.266@news.cpqcorp.net>, John Reagan <john.reagan@hp.com> writes: > Bob Koehler wrote:I >>    I've never really understood why DEC decided to use counted stringsnJ >>    instead of descriptors in PASCAL, other than I think that's what theG >>    reference model uses.  I think all PASCAL requirements could havesF >>    been met if strings were passed by descriptor, and it would haveG >>    made the interface to VMS routines as simple as it is in Fortran.n > G > What "reference model"?  Classic Pascal doesn't have variable length 0: > strings.  They are only fixed length, just like Fortran.  @    The original PASCAL language prohibited one from constructingA    functions that could take different length strings for a given C    argument.  Many vendors added this as an extension, and therefor B    were not writing PASCAL compilers according to the authors, who    disallowed all extensions.o  G    Pascal now has an ANSI (and I assume ISO) standard.  Most high levelsE    language standards specify behaviour, but are often accompanied by >    reference models which describe how such behaviour might beA    accomplished.  Reference models are often discussed during the =    writing of new or updated standards.  I assume there is a l+    reference model for the Pascal standard.r   ------------------------------  # Date: Tue, 19 Oct 2004 13:12:22 GMT 5 From: "Ed Vogel" <edward.vogel_stop_the_spam.@hp.com>e, Subject: Re: Latest on Windoze Navy software1 Message-ID: <Wo8dd.1118$OY5.717@news.cpqcorp.net>.  , "reb" <natron@ntlworld.com> wrote in message* news:opsf27hwxeaqvj2s@news.ntlworld.com...: > Unlike C, RTL/2 provides runtime array bound checking as9 > a basic feature, so at various points in their code thet > programmer can select to:e > - check all array accesses; > - check "unsafe" array accesses (i.e. writes & reads from 4 >    arrays of REFs [typed pointers])  {the default} > - suppress all checksl  9     While most C compilers do not implement runtime arrays:     bound checking, recent versions of Compaq/HP C do have<     this feature if a command line qualifier (/CHECK=BOUNDS)     is specified.u  H     Exactly what is checked is rather complicated...see the documetation     of HELP text for more info.l       Ed Vogel     HP/Compaq C Engineering.   ------------------------------    Date: 19 Oct 2004 10:01:59 -0500- From: Kilgallen@SpamCop.net (Larry Kilgallen) , Subject: Re: Latest on Windoze Navy software3 Message-ID: <bZrhoIB6K1CA@eisner.encompasserve.org>c  V In article <2tj4roF1vr800U1@uni-berlin.de>, bill@cs.uofs.edu (Bill Gunshannon) writes:3 > In article <L8Tcd.1055$DV4.151@news.cpqcorp.net>, + > 	John Reagan <john.reagan@hp.com> writes:c >> Bill Gunshannon wrote:- >> - >>> G >>> Hmmmm....  I remember "pass-by-value" and "pass-by-reference" but IsF >>> sure don't remember "pass-by-descriptor" in any Pascal text I ever	 >>> read.s >>>  >> b> >> Compaq Pascal User Manual for OpenVMS Systems, section 5.3  >> Parameter-Passing Mechanismst >> mJ >> "By default, Compaq Pascal uses the by reference mechanism to pass all K >> actual parameters except those that correspond to conformant parameters eJ >> and undiscriminated schema parameters, in which case the by descriptor  >> mechanism is used." >>  I >> Then in section 5.3.3, By Descriptor, table 5-3, lists the descriptor aK >> classes chosen for the various conformant array types (the compiler can 0* >> generate CLASS_A, _NCA, _VS, and _VSA). >> _ > B > That's not Pascal, that's Compaq Pascal.  An implementation, not
 > a language.p  F As has been explained to you before, descriptors are an implementation issue, not a language issue.   ------------------------------  % Date: Tue, 19 Oct 2004 11:13:40 -0400e# From: sol gongola <sol@adldata.com>o, Subject: Re: Latest on Windoze Navy software+ Message-ID: <41752F24.87362403@adldata.com>e   Bill Gunshannon wrote: > 3 > In article <L8Tcd.1055$DV4.151@news.cpqcorp.net>,12 >         John Reagan <john.reagan@hp.com> writes: > > Bill Gunshannon wrote: > >a > >>H > >> Hmmmm....  I remember "pass-by-value" and "pass-by-reference" but IG > >> sure don't remember "pass-by-descriptor" in any Pascal text I evern
 > >> read. > >> > > > > > Compaq Pascal User Manual for OpenVMS Systems, section 5.3  > > Parameter-Passing Mechanisms > >OJ > > "By default, Compaq Pascal uses the by reference mechanism to pass allK > > actual parameters except those that correspond to conformant parameters J > > and undiscriminated schema parameters, in which case the by descriptor > > mechanism is used."  > >uI > > Then in section 5.3.3, By Descriptor, table 5-3, lists the descriptoriK > > classes chosen for the various conformant array types (the compiler can + > > generate CLASS_A, _NCA, _VS, and _VSA).a > >t > B > That's not Pascal, that's Compaq Pascal.  An implementation, not@ > a language.  So, if I decide to write a modifed C Library thatD > uses some format other than the dreaded "null terminated characterC > array" and distribute it with a some compiler does that mean thatt= > the C language would be changed?  Or just my quirky system?  >  > bill >  > --L > Bill Gunshannon          |  de-moc-ra-cy (di mok' ra see) n.  Three wolvesF > bill@cs.scranton.edu     |  and a sheep voting on what's for dinner. > University of Scranton   |@ > Scranton, Pennsylvania   |         #include <std.disclaimer.h>  . It would not change the language definition.  5 It would mean that you have a C support library that x2 is incompatible with the standard library and the " standard use of character strings.  < Now if you had a C like compiler where strings would not be 7 generated with a trailing null as would be expected in   the following:' 	char    strnam[] = "This is a string";e 	printf("Here is some data\n");   5 Then 'printf' and other functions would have to have I8 another way of knowing when the data in the string ends. and it wouldn't be C anymore.I   solA   ------------------------------  # Date: Tue, 19 Oct 2004 15:28:25 GMTC& From: John Reagan <john.reagan@hp.com>, Subject: Re: Latest on Windoze Navy software2 Message-ID: <toadd.1136$dY5.1052@news.cpqcorp.net>   Bob Koehler wrote:   >  > B >    The original PASCAL language prohibited one from constructingC >    functions that could take different length strings for a given E >    argument.  Many vendors added this as an extension, and thereforyD >    were not writing PASCAL compilers according to the authors, who >    disallowed all extensions.- > I >    Pascal now has an ANSI (and I assume ISO) standard.  Most high levelrG >    language standards specify behaviour, but are often accompanied bye@ >    reference models which describe how such behaviour might beC >    accomplished.  Reference models are often discussed during theb? >    writing of new or updated standards.  I assume there is a  - >    reference model for the Pascal standard.w  C No, there have never been approved reference models for any of the c Pascal standards.   I The original Pascal ANSI standard did not include the "conformant array" -G feature, but the original ISO 7185 Pascal standard did.  Subsequently, IF there is now an Extended Pascal standard (with run-sized times, etc.)  that was approved in 1991.     -- r John Reagant/ HP Pascal/{A|I}MACRO for OpenVMS Project Leaderc Hewlett-Packard Companyd   ------------------------------  % Date: Tue, 19 Oct 2004 08:39:05 -0700 # From: "Tom Linden" <tom@kednos.com>y, Subject: Re: Latest on Windoze Navy software( Message-ID: <opsf4i3fjezgicya@hyrrokkin>  H On 19 Oct 2004 10:01:59 -0500, Larry Kilgallen <Kilgallen@SpamCop.net>   wrote:  F > In article <2tj4roF1vr800U1@uni-berlin.de>, bill@cs.uofs.edu (Bill   > Gunshannon) writes:-4 >> In article <L8Tcd.1055$DV4.151@news.cpqcorp.net>,, >> 	John Reagan <john.reagan@hp.com> writes: >>> Bill Gunshannon wrote: >>>e >>>>H >>>> Hmmmm....  I remember "pass-by-value" and "pass-by-reference" but IG >>>> sure don't remember "pass-by-descriptor" in any Pascal text I ever9
 >>>> read. >>>> >>>0> >>> Compaq Pascal User Manual for OpenVMS Systems, section 5.3  >>> Parameter-Passing Mechanisms >>> J >>> "By default, Compaq Pascal uses the by reference mechanism to pass allK >>> actual parameters except those that correspond to conformant parametershJ >>> and undiscriminated schema parameters, in which case the by descriptor >>> mechanism is used."i >>>aI >>> Then in section 5.3.3, By Descriptor, table 5-3, lists the descriptorhK >>> classes chosen for the various conformant array types (the compiler cane+ >>> generate CLASS_A, _NCA, _VS, and _VSA).  >>>h >>C >> That's not Pascal, that's Compaq Pascal.  An implementation, not  >> a language. >yH > As has been explained to you before, descriptors are an implementation > issue, not a language issue.  I Well, %DESCR is an extension, which would make a source program employingcJ it non-portable.  So, from that point of view it is a language issue, evenI though it appears to be a preprocessor function.  BTW, I noted that the  b online! docs for Pascal were only in pdf.n  K PL/I does the same, BTW, but goes a step further and makes them attributes.bM http://www.kednos.com/pli/docs/REFERENCE_MANUAL/6291pro_021.html#index_x_1030n       -- vC Using Opera's revolutionary e-mail client: http://www.opera.com/m2/e   ------------------------------  % Date: Tue, 19 Oct 2004 22:15:26 +0800e From: prep@prep.synonet.coms, Subject: Re: Latest on Windoze Navy software- Message-ID: <87y8i2n69t.fsf@prep.synonet.com>s  * bill@cs.uofs.edu (Bill Gunshannon) writes:  D > That's not Pascal, that's Compaq Pascal.  An implementation, not aC > language.  So, if I decide to write a modifed C Library that usesnF > some format other than the dreaded "null terminated character array"B > and distribute it with a some compiler does that mean that the C7 > language would be changed?  Or just my quirky system?S  A Yes, the definintion of C includes the `standard library'. Changee" that, and you change the language.   -- v< Paul Repacholi                               1 Crescent Rd.,7 +61 (08) 9257-1001                           Kalamunda. @                                              West Australia 6076* comp.os.vms,- The Older, Grumpier Slashdot. Raw, Cooked or Well-done, it's all half baked.F EPIC, The Architecture of the future, always has been, always will be.   ------------------------------    Date: 19 Oct 2004 12:20:04 -0500 From: briggs@encompasserve.org, Subject: Re: Latest on Windoze Navy software3 Message-ID: <BpdHba6zwVab@eisner.encompasserve.org>   N In article <opsf4i3fjezgicya@hyrrokkin>, "Tom Linden" <tom@kednos.com> writes:J > On 19 Oct 2004 10:01:59 -0500, Larry Kilgallen <Kilgallen@SpamCop.net>   > wrote:I >> As has been explained to you before, descriptors are an implementationa >> issue, not a language issue.I > K > Well, %DESCR is an extension, which would make a source program employingrL > it non-portable.  So, from that point of view it is a language issue, even2 > though it appears to be a preprocessor function.  B If you use a non-portable extension then the behavior of a programC that uses that extension is an implementation issue, not a language  issue.  B If you need to use a non-portable extension then the behavior thatB you are trying to influence must be implementation-specific rather than language-defined.  C The language definition is silent on both the question of parametertC passing mechanism (reference versus descriptor) and on the questionsC of interpretation of the %DESCR extension.  Accordingly, neither ist2 a language issue.  Both are implementation issues.  A One should distinguish between the semantics of parameter passingr' and the mechanism of parameter passing.r  A Call by value, by reference or by copy-in, copy-back -- those ares; semantics.  You can write a program to tell which semanticsuA are in use.  And the language definition may prescribe particular-< semantics.  Or it may be left as an implementation decision.  E Call by value, by reference or by descriptor -- those are mechanisms.HE You cannot, in general, write a program to tell which mechanism is in@E use.   Parameter passing mechanism is almost always an implementationg	 decision.>  < One popular choice used by language definers is to prescribeA copy-in, copy-back semantics by default but to allow by-reference ! semantics for structured objects.b   On VAX:i  J You can implement by-value semantics using value, reference or descriptor.  F You can implement by-reference semantics using reference or descriptor  L You can implement copy-in, copy-back semantics using reference or descriptor       	John Briggs   ------------------------------  % Date: Tue, 19 Oct 2004 05:31:56 -0700o, From: James Nykiel <jnykiel@access4less.net>3 Subject: Re: Need LK461 keyboard and 3 button mouse ; Message-ID: <pan.2004.10.19.12.31.54.64010@access4less.net>x  
 Greetings:    I Thank you all for your assistance, I have obtained the LK461 keyboard and A 3 button mouse that I was looking for.  Please close this thread.S     Thanks - Jim  7 On Sat, 02 Oct 2004 17:23:39 -0700, James Nykiel wrote:    > Greetings: >  > G > I am in the process of putting together an AlphaStation 200 system on H > which to learn OpenVMS.  I would like to obtain the following items to+ > help complete my AlphaStation 200 system.	 > I > (1) LK461-A2 - 108 Key OpenVMS Keyboard. (1) PCXAS-AB - 3 Button Mouse.  > E > I have recently checked eBay but have had no luck in locating theserH > items. If you have either or both of these items and the are in decentJ > condition I would be interested in purchasing them if the price is right0 > as my budget is somewhat limited at this time. >  > Note:' > I > I am not sure if the PCXAS-AB 3 Button Mouse is the best choice for use J > with an AlphaStation 200 running OpenVMS and DECWindows.  I know I couldH > probably use almost any brand PS/2 mouse but would like to stay with a > DEC mouse. >  >  > Thank you... - James Nykiel    ------------------------------  % Date: Tue, 19 Oct 2004 13:16:20 -0400 # From: "John Smith" <a@nonymous.com> $ Subject: OT: Short Article about Sun, Message-ID: <K82dnYh-wOJC1ujcRVn-rQ@igs.net>  5 From www.Forbes.com (free membership signup required)t   also at>  L http://story.news.yahoo.com/news?tmpl=story&cid=64&ncid=1208&e=10&u=/fo/2004+ 1018/bs_fo/8862d235842307a7248f33ab0aee5209n    	 Sun Shiner Mon Oct 18, 3:46 PM ET   By Chana R. Schoenberger  C Would an initial public offering like this interest you? You have afK well-established high-tech company with a cash hoard. Add debt and subtracttK cash from the market capitalization and the business is valued at 1.1 timessK annual revenue. Despite some rough patches, it's a growth company: Revenuesi' are double what they were a decade ago.   C This is no new issue but an ancient (by silicon standards) computerOJ manufacturer: SUN MICROSYSTEMS (4, SUNW). Analyst Mark Stahlman of Caris &I Co. says the shares are a bargain, and not only because they are down 94%28 from their 2000 high. He thinks its shares are worth $8.  G Sun's recent results aren't much to write bullish reports about: a $376CD million loss in the June 30 fiscal year on $11 billion in sales amidI shrinking margins. That included a final quarter with 24 cents a share ofcI earnings. (Chairman Scott McNealy's self-deprecating take: "We've had one/D good quarter in a row.") The company was due to report first-quarterK earnings after this magazine went to press, but analysts were expecting SunT to lose 3 cents a share.  C Stahlman sees salvation in Sun's new 64-bit servers. These powerfulGG machines, an upgrade from the now-standard 32-bit servers, use Advanced J Micro Devices' Opteron chips to run bigger databases with fancier graphicsL and better security. Close links with AMD may enable Sun to upstage Dell andL Hewlett-Packard, both committed to Intel's less than dazzling Itanium 64-bit chip.h     -----------------------------   L What I find interesting about this article is that more and more Wall StreetL analysts are denigrating Itanic. You may want to call them 'casino analysts'B but my money is on the market (both stock and business consumer ofJ technology market, read 'datacenter') gravitating to AMD64 long before and! in far higher volume than Itanic.e  D From an executive suite view, it is far more likely that a senior ITH decision maker is going to make a purchasing decision based more on whatK his/her boss would support than on almost anything else -- having seen thisoJ occur more times than I can count. The thing of it is that the CEO is moreJ likely to read the Morgan Stanley or Merrill research report viewing AMD64J more positively than Itanic vs. reading a Gartner report which may/may not effuse about Itanic.  / Like they say, there's no press like bad press.h  A And recently all the market has been hearing about HP is that theoC Enterprise, PC, and Storage divisions are having problems, HP can'thI implement its own order management system properly, and that carly(tm) isi> pushing the company towards selling TV's to compete with Sony.  > Draw your own conclusions about the market's perception of HP.   ------------------------------   Date: 19 Oct 2004 06:08:52 GMT2 From: "Dave Weatherall" <djw-nothere@nospam.nohow>$ Subject: Re: pthreads and scheduling? Message-ID: <DTiotGxQ0bj6-pn2-r1etuI6GdCNE@dave2_os2.home.ours>r  2 On Mon, 18 Oct 2004 13:35:34 UTC, "Johan Nilsson" ) <johan.nilsson@---.esrange.ssc.se> wrote:-   > Hi,a > D > I'm having trouble figuring out how the thread scheduling actuallyC > behaves/should behave under OpenVMS Alpha (7.3-1). Here's a brief8- > description of the application in question:a >  > - Implemented in C++ > - Using POSIX threads * > - Process runs at realtime priority (18)N > - Main thread basically kicks off an I/O thread, and then waits for a common% > event flag that signals terminationnG > - The I/O thread reads UDP multicast data using QIO + uses a timer insI > combination with sys$wflor to implement read timeouts. The thread never I > waits for I/O completion for more than 100ms. Incoming data varies fromh > ~10-100Hz.J > - A logging thread is also started at ~main entry (writes log entries to > disc).? > - Application's main is compiled with /reentrancy=multithreadaB > - Application is linked using /threads_enable=(multiple,upcalls) > H > Using the following thread priorities/policies the main thread _never_N > returns from the $waitfr system call (I know the event flag is getting set): > - > 1. Main thread runs at SCHED_RR, PRI_RR_MIN H > 2. I/O thread runs at SCHED_RR, PRI_RR_MIN+((PRI_RR_MAX-PRI_RR_MIN)/2)5 > 3. Logging thread runs at SCHED_OTHER+PRI_OTHER_MINb > I > I tried to set the main thread to SCHED_RR,PRI_RR_MAX: but then the I/OvJ > thread _never_ runs, even as the main thread simply waits for the common > event flag to become set.   B That's consistent with it (main) never giving up control. I would E double-check the $WAITFR setup. OTOH wouldn't the main be better off  A waiting for pthread condition rather than a VMS event flag?  The  A UPCALLS-enable should actually make the two things work together   but...   N > What's going on here - am I doing something fundamentally wrong, or is thereG > a problem with the pthreads implementation (highly unlikely I guess)?bL > Another thing that surprised me was that threads created by default from aK > realtime process does not seem to inherit the main thread's priority - itnD > looks like they are started at SCHED_OTHER + some medium priority.  E My memory is that each thread takes the priority set by the explicit sF call or defaults to ? I don't remember it inheriting from the parent. , Need to look in the book and that's at work.   --   Cheers - Dave W.   ------------------------------  % Date: Tue, 19 Oct 2004 08:39:04 +0200a8 From: "Johan Nilsson" <johan.nilsson@---.esrange.ssc.se>$ Subject: Re: pthreads and scheduling8 Message-ID: <1098167943.r3DpxkQ0sclGwSHJRhH5ow@teranews>  = "Dave Weatherall" <djw-nothere@nospam.nohow> wrote in message 9 news:DTiotGxQ0bj6-pn2-r1etuI6GdCNE@dave2_os2.home.ours...M3 > On Mon, 18 Oct 2004 13:35:34 UTC, "Johan Nilsson"e+ > <johan.nilsson@---.esrange.ssc.se> wrote:e >s > > Hi,r > >tF > > I'm having trouble figuring out how the thread scheduling actuallyE > > behaves/should behave under OpenVMS Alpha (7.3-1). Here's a brieft/ > > description of the application in question:s > >?   [details snipped]h   > >IJ > > Using the following thread priorities/policies the main thread _never_J > > returns from the $waitfr system call (I know the event flag is getting set):a > >e/ > > 1. Main thread runs at SCHED_RR, PRI_RR_MINeJ > > 2. I/O thread runs at SCHED_RR, PRI_RR_MIN+((PRI_RR_MAX-PRI_RR_MIN)/2)7 > > 3. Logging thread runs at SCHED_OTHER+PRI_OTHER_MINI > >yK > > I tried to set the main thread to SCHED_RR,PRI_RR_MAX: but then the I/OxL > > thread _never_ runs, even as the main thread simply waits for the common > > event flag to become set.4 >:; > That's consistent with it (main) never giving up control.c  H Calling $waitfr should suspend the calling thread until the flag is set, thus giving up control - or?  	 > I would"! > double-check the $WAITFR setup.s   sys$waitfr(efn) ???0  % >OTOH wouldn't the main be better off:B > waiting for pthread condition rather than a VMS event flag?  TheB > UPCALLS-enable should actually make the two things work together > but... >t  C This is a new process in a legacy application consisting of several K processes. The event flag in question is used to signal process terminationp7 to all processes and I really can't change this design.   L When the main thread returns from $waitfr, it signals the I/O thread to exitI (internally using $cancel on the I/O thread's device, which tells the I/OoH thread to quit) and then waits for the thread to exit using pthread_join before returning from main.q  J > > What's going on here - am I doing something fundamentally wrong, or is thereaI > > a problem with the pthreads implementation (highly unlikely I guess)? L > > Another thing that surprised me was that threads created by default from a J > > realtime process does not seem to inherit the main thread's priority - itF > > looks like they are started at SCHED_OTHER + some medium priority. > F > My memory is that each thread takes the priority set by the explicitG > call or defaults to ? I don't remember it inheriting from the parent.". > Need to look in the book and that's at work. >u   That seems to be the case, yes.i   // Johan   ------------------------------  % Date: Tue, 19 Oct 2004 09:45:30 +0200a8 From: "Johan Nilsson" <johan.nilsson@---.esrange.ssc.se>$ Subject: Re: pthreads and scheduling8 Message-ID: <1098171929.T9+Adf+9uj5UA3KcQfRQ+w@teranews>  C "Johan Nilsson" <johan.nilsson@---.esrange.ssc.se> wrote in messagey2 news:1098167943.r3DpxkQ0sclGwSHJRhH5ow@teranews... >y? > "Dave Weatherall" <djw-nothere@nospam.nohow> wrote in message ; > news:DTiotGxQ0bj6-pn2-r1etuI6GdCNE@dave2_os2.home.ours... 5 > > On Mon, 18 Oct 2004 13:35:34 UTC, "Johan Nilsson"l- > > <johan.nilsson@---.esrange.ssc.se> wrote:j > >s	 > > > Hi,t > > >aH > > > I'm having trouble figuring out how the thread scheduling actuallyG > > > behaves/should behave under OpenVMS Alpha (7.3-1). Here's a briefo1 > > > description of the application in question:s > > >  >r > [details snipped]  >e > > >yL > > > Using the following thread priorities/policies the main thread _never_L > > > returns from the $waitfr system call (I know the event flag is getting > set):  > > >t1 > > > 1. Main thread runs at SCHED_RR, PRI_RR_MINeL > > > 2. I/O thread runs at SCHED_RR, PRI_RR_MIN+((PRI_RR_MAX-PRI_RR_MIN)/2)9 > > > 3. Logging thread runs at SCHED_OTHER+PRI_OTHER_MINs > > >tI > > > I tried to set the main thread to SCHED_RR,PRI_RR_MAX: but then thee I/O G > > > thread _never_ runs, even as the main thread simply waits for they common > > > event flag to become set.w > > = > > That's consistent with it (main) never giving up control.b >hJ > Calling $waitfr should suspend the calling thread until the flag is set, > thus giving up control - or?  K Talkin to myself, but I just found the following in the "Guide to the POSIXe Threads Library", B.12.8:a   --- Note ---E If a thread calls a system service that uses a common event flag, the L calling thread's virtual processor blocks until the wait is satisfied. (ThatH is, no upcall is made to the OpenVMS kernel to schedule another thread.)H On a uniprocessor, such a system service call will most likely cause all  threads in the process to block. ------  J Which explains what's happening in the second case. It doesn't explain whyL running the main and I/O threads at SCHED_RR as in the first case causes the7 main thread to _never_ return from the $waitfr, though.a  J Looking at the note above though, it should lock up the entire applicationE as well as soon as it (the main thread) calls into $waitfr - but thateG doesn't seem to happen. Could it possibly be so that the main thread isnK never scheduled until the I/O thread wakes up - doesn't seem likely as this I behaviour occurs even with no incoming data - weird. Maybe it's the "mostn" likely" part that isn't satisfied.   // Johan   ------------------------------    Date: 19 Oct 2004 07:31:02 -0500; From: koehler@eisner.nospam.encompasserve.org (Bob Koehler)o' Subject: Re: Question about Descriptorso3 Message-ID: <bXR3cF4FaZ2$@eisner.encompasserve.org>u  \ In article <41742D41.D46B5DB7@teksavvy.com>, JF Mezei <jfmezei.spamnot@teksavvy.com> writes:N > When descriptors were originally designed, why did they not put 2 lengths in > the structure ?  >  > 	dsc$w_alloc
 > 	dsc$w_size   #    Ask the guy who did item_list_1.r   ------------------------------  % Date: Tue, 19 Oct 2004 09:21:55 -0400 3 From: Steve Lionel <Steve.Lionel@REMOVEintelME.com>e' Subject: Re: Question about Descriptorsv8 Message-ID: <lm4an0d7uma14a35r5eu7qq2nko0pn2tpm@4ax.com>  K On Mon, 18 Oct 2004 16:53:34 -0400, JF Mezei <jfmezei.spamnot@teksavvy.com>t wrote:  M >When descriptors were originally designed, why did they not put 2 lengths inh >the structure ?  L The original design of descriptors was such that the most common descriptorsM (CLASS_S in particular) could be loaded with a simple MOVQ instruction.  ThissN is also why the length field was 16 bits (as was the maximum length of the VAX& character instructions such as MOVC3.)  M As mentioned by others, the effect you want is met by the CLASS_VS descriptor'G with DTYPE_VT datatype. For non-string datatypes, specifying a separatetD allocation field seems, to me, outside of what descriptors were for.   Steve    ------------------------------  % Date: Tue, 19 Oct 2004 11:19:10 -0400>- From: JF Mezei <jfmezei.spamnot@teksavvy.com> ' Subject: Re: Question about Descriptorsu, Message-ID: <41753065.9C01C072@teksavvy.com>   Steve Lionel wrote:cO > As mentioned by others, the effect you want is met by the CLASS_VS descriptor I > with DTYPE_VT datatype. For non-string datatypes, specifying a separaterF > allocation field seems, to me, outside of what descriptors were for.  J Ok, now the tough question. If I were to start using this descriptor classK instead of the "default" one, how many system services and library routinesu, would actually handle this format properly ?   ------------------------------  % Date: Tue, 19 Oct 2004 11:27:45 -0400 3 From: Steve Lionel <Steve.Lionel@REMOVEintelME.com>s' Subject: Re: Question about Descriptorsn8 Message-ID: <macan0lkajoj9ghi75opt3bq1v6keits15@4ax.com>  K On Tue, 19 Oct 2004 11:19:10 -0400, JF Mezei <jfmezei.spamnot@teksavvy.com>s wrote:   >Steve Lionel wrote:P >> As mentioned by others, the effect you want is met by the CLASS_VS descriptorJ >> with DTYPE_VT datatype. For non-string datatypes, specifying a separateG >> allocation field seems, to me, outside of what descriptors were for.. > K >Ok, now the tough question. If I were to start using this descriptor classsL >instead of the "default" one, how many system services and library routines- >would actually handle this format properly ?s  G None of the system services would, last I checked (which admittedly wasvE several years ago).  All of the LIB$ and STR$ routines should, unless N documented to the contrary.  I will note that the only compiler I know of withJ built-in support for CLASS_VS descriptors, to the extent that it generatesL them automatically, is Pascal. It also knows how to pass CLASS_S descriptors to routines that need them.0  N I could not get from your earlier posts exactly what problem you are trying to solve.   Steveg   ------------------------------  % Date: Tue, 19 Oct 2004 12:19:32 -0400 - From: JF Mezei <jfmezei.spamnot@teksavvy.com> ' Subject: Re: Question about Descriptorse, Message-ID: <41753E87.760B9013@teksavvy.com>   Steve Lionel wrote:hP > I could not get from your earlier posts exactly what problem you are trying to > solve.  K not trying to solve any specific problem. But in the context of dicsussionsfJ about the "dangers" of null terminated strings where "descriptors" are theJ answer to alkl evil microsoft problems, I started to wonder about how mustM manually differentiate between allocated and used values in a descriptor, (ataK least in C) and hence my question whether the engineers had ever consideredoC having both values in the structure when they initially devised it.l  M But I guess one answer about the ability to fit the descriptor structure in atL small enough space to be handled by a single vax instruction may have been a driving force.   ------------------------------  % Date: Tue, 19 Oct 2004 11:55:52 -0400s( From: David Froble <davef@tsoft-inc.com>' Subject: Re: Question about Descriptors , Message-ID: <41753908.6080304@tsoft-inc.com>   JF Mezei wrote:o   > Tim Sneddon wrote: > L >>What is wrong with using a DSC$K_CLASS_S string and passing the address ofL >>DSC$W_LENGTH for the length argument too? The following is some BLISS code >>that explains what I mean: >> > M > Yeah, but still requires that you hold the allocated size separately so youcJ > can reset the Ddsc$w_length field after you'd used the field to indicate > current string length. > L > I know how to work with the current design. But just wondered if there wasO > some reason they didn't put in the 2 lengts (used, allocated) in the originaleE > description definitions which would have made for cleaner programs.t >   M Just guessing here.  The use of dynamic strings seems to be in BASIC.  (Note eP that I've not used all the languages that are/were available on VMS.)  Prior to K the development of dynamic strings for BASIC, it's my impression that most tO string usage (with descriptors) was static (fixed length) strings.  Being more rG specific, the allocated memory for the actual data was fixed in length.   Q If there's any reality to my guess, then the development of dynamic strings used n> the available descriptors rather than develop new descriptors.  O I've heard, somewhere in the past, that when re-setting the value of a dynamic  Q string, that if the new data is larger, then the existing memory is released and  O a new block of memory of sufficient size is acquired.  If smaller, the size of cQ the data is set and the rest of the memory ignored.  I'm guessing that if that's 4O so, that the memory would be in a location that is all released when the image :N is run down.  Still, a long running program would be 'leaking' memory.  Never : got into reading the source for the STR$ libraby routines.  M I can see the reasoning for 2 length values.  Such would keep track of total uG allocated memory for the string, thus never 'losing' any of the memory.e  L It also is a bit disconcerting when thinking about changing the length of a P string.  Even if the current memory is released, just what is a system/OS/image Q going to do with many memory fragments each consisting of a few bytes?  Consider eK changing the value (and length) of a string variable several million times.l   Dave   --  4 David Froble                       Tel: 724-529-04504 Dave Froble Enterprises, Inc.      Fax: 724-529-0596> DFE Ultralights, Inc.              E-Mail: davef@tsoft-inc.com 170 Grimplin Roade Vanderbilt, PA  15486t   ------------------------------  # Date: Tue, 19 Oct 2004 16:51:05 GMT-& From: John Reagan <john.reagan@hp.com>' Subject: Re: Question about Descriptorsr1 Message-ID: <ZBbdd.1145$9f6.518@news.cpqcorp.net>R   JF Mezei wrote:e   > L > Ok, now the tough question. If I were to start using this descriptor classM > instead of the "default" one, how many system services and library routines . > would actually handle this format properly ?  # I was just going to mention this...   G In general (and I have to check to be sure), most SYS$ system services  H do not handle CLASS_VS very well.  They just expect CLASS_S descriptors  for output buffers.m  H However, as others have mentioned, you can can pass a "full" descriptor I to the OUTBUF parameter and the address of the descriptor's DSC$W_LENGTH e field to the OUTLEN parameter.  E You can use varying strings with system services.  You would have to lE pass a descriptor to the "body" as the OUTBUF and the address of the s length work as the OUTLEN. For example,  > OUTBUF := myvarying.body, OUTLEN := iaddress(myvarying.length)  H so the system service sees a CLASS_S that describes the whole body (ie, @ the maximum length) and then stores the resultant length in the L varying's length word.  When you are done, you have a valid varying of char.    F For other RTL routines (LIB$, STR$, etc.), they then to accept all of I the valid string descriptors (CLASS_S, CLASS_D, CLASS_VS, CLASS_SB, etc.)e   --   John Reagant/ HP Pascal/{A|I}MACRO for OpenVMS Project Leaderd Hewlett-Packard Company    ------------------------------  % Date: Tue, 19 Oct 2004 13:03:36 -0400-# From: sol gongola <sol@adldata.com> ' Subject: Re: Question about Descriptorsa+ Message-ID: <417548E8.E5B8CC1D@adldata.com>:   JF Mezei wrote:N >  > Steve Lionel wrote:WR > > I could not get from your earlier posts exactly what problem you are trying to
 > > solve. > M > not trying to solve any specific problem. But in the context of dicsussionseL > about the "dangers" of null terminated strings where "descriptors" are theL > answer to alkl evil microsoft problems, I started to wonder about how mustO > manually differentiate between allocated and used values in a descriptor, (ateM > least in C) and hence my question whether the engineers had ever considered E > having both values in the structure when they initially devised it.a > O > But I guess one answer about the ability to fit the descriptor structure in a N > small enough space to be handled by a single vax instruction may have been a > driving force.  B Descriptors are not "The Solution" to the problem. They are a way B to manage a program. Languages that supported strings worked fine A before DEC had descriptors. Even null terminated strings are not e@ the problem. The problem is in the coding practice where simple C checks are not being performed to prevent undesirable data overlaysi' including buffer overrun type problems.c  B A programmer needs to maintain control of his/her program, should A anticipate the possibility that a string length may not be valid IB regardless of the length being in a descriptor, a separate length @ field, in the first 2 or 4 bytes at the actual address a string,= or a string terminated with a specific character or trailing s characters.    sol>   ------------------------------  % Date: Tue, 19 Oct 2004 10:01:01 -0700e# From: "Tom Linden" <tom@kednos.com> ' Subject: Re: Question about Descriptorsr( Message-ID: <opsf4mvzwezgicya@hyrrokkin>  H On Tue, 19 Oct 2004 13:03:36 -0400, sol gongola <sol@adldata.com> wrote:   > JF Mezei wrote:  >> >> Steve Lionel wrote:K >> > I could not get from your earlier posts exactly what problem you are  a >> trying to >> > solve.s >>D >> not trying to solve any specific problem. But in the context of   >> dicsussionsK >> about the "dangers" of null terminated strings where "descriptors" are  s >> theJ >> answer to alkl evil microsoft problems, I started to wonder about how   >> must.B >> manually differentiate between allocated and used values in a   >> descriptor, (atE >> least in C) and hence my question whether the engineers had ever  e
 >> considered4F >> having both values in the structure when they initially devised it. >>C >> But I guess one answer about the ability to fit the descriptor  e >> structure in a:J >> small enough space to be handled by a single vax instruction may have  	 >> been ad >> driving force." >tC > Descriptors are not "The Solution" to the problem. They are a wayIC > to manage a program. Languages that supported strings worked finenB > before DEC had descriptors. Even null terminated strings are notA > the problem. The problem is in the coding practice where simplesE > checks are not being performed to prevent undesirable data overlaysp) > including buffer overrun type problems.p >mC > A programmer needs to maintain control of his/her program, shoulduB > anticipate the possibility that a string length may not be validC > regardless of the length being in a descriptor, a separate lengthtB > field, in the first 2 or 4 bytes at the actual address a string,> > or a string terminated with a specific character or trailing
 > characters.aC I don't disagree with what you say; however, when you are trying tomF do this in C, you have to do a lot more work, and you still don't haveD on conditions or automatic bounds/range checking.  So guess what, it doesn't get done.e >  > sol        -- yC Using Opera's revolutionary e-mail client: http://www.opera.com/m2/-   ------------------------------    Date: 19 Oct 2004 12:12:31 -0500; From: koehler@eisner.nospam.encompasserve.org (Bob Koehler)t' Subject: Re: Question about Descriptors 3 Message-ID: <2ahqZiQ++kZU@eisner.encompasserve.org>   \ In article <41753E87.760B9013@teksavvy.com>, JF Mezei <jfmezei.spamnot@teksavvy.com> writes: > O > But I guess one answer about the ability to fit the descriptor structure in ayN > small enough space to be handled by a single vax instruction may have been a > driving force.  ?   Also remember at the time Fortran was king, and the upcomming"E   Fortran-77 standard was going to require tracking the data locationnA   and allocated length.  Nothing was driving a need to track used 	   length..  G   Meanwhile dynamic descriptors are your friend, when and where you can    use them.1   ------------------------------  % Date: Tue, 19 Oct 2004 13:14:59 -0400R# From: sol gongola <sol@adldata.com>e' Subject: Re: Question about Descriptorsr* Message-ID: <41754B93.15F8C9F@adldata.com>   David Froble wrote:n >  > JF Mezei wrote:o >  > > Tim Sneddon wrote: > >aN > >>What is wrong with using a DSC$K_CLASS_S string and passing the address ofN > >>DSC$W_LENGTH for the length argument too? The following is some BLISS code > >>that explains what I mean: > >> > >lO > > Yeah, but still requires that you hold the allocated size separately so yourL > > can reset the Ddsc$w_length field after you'd used the field to indicate > > current string length. > >mN > > I know how to work with the current design. But just wondered if there wasQ > > some reason they didn't put in the 2 lengts (used, allocated) in the originalIG > > description definitions which would have made for cleaner programs.o > >r > N > Just guessing here.  The use of dynamic strings seems to be in BASIC.  (NoteQ > that I've not used all the languages that are/were available on VMS.)  Prior to,L > the development of dynamic strings for BASIC, it's my impression that mostP > string usage (with descriptors) was static (fixed length) strings.  Being moreI > specific, the allocated memory for the actual data was fixed in length.e > R > If there's any reality to my guess, then the development of dynamic strings used@ > the available descriptors rather than develop new descriptors. > P > I've heard, somewhere in the past, that when re-setting the value of a dynamicR > string, that if the new data is larger, then the existing memory is released andP > a new block of memory of sufficient size is acquired.  If smaller, the size ofR > the data is set and the rest of the memory ignored.  I'm guessing that if that'sP > so, that the memory would be in a location that is all released when the imageO > is run down.  Still, a long running program would be 'leaking' memory.  Never"< > got into reading the source for the STR$ libraby routines. > N > I can see the reasoning for 2 length values.  Such would keep track of totalI > allocated memory for the string, thus never 'losing' any of the memory.f > M > It also is a bit disconcerting when thinking about changing the length of a Q > string.  Even if the current memory is released, just what is a system/OS/imageeR > going to do with many memory fragments each consisting of a few bytes?  ConsiderM > changing the value (and length) of a string variable several million times.o  nK IIRC, PDP-11 RSTS BASIC plus deallocated and reallocated strings each time  F an string assignment was made. There was also a function that allowed E a string to be modified "in place" without the deallocate/reallocate.s  G Hopefully this is being handled internally by the language runtime and S not by the system OS.    sol    ------------------------------  + Date: Tue, 19 Oct 2004 17:19:25 +0000 (UTC)t( From: m.kraemer@gsi.de (Michael Kraemer)' Subject: Re: Question about DescriptorsI5 Message-ID: <cl3iat$c5f$1@lnx107.hrz.tu-darmstadt.de>n  N In article <opsf4mvzwezgicya@hyrrokkin>, "Tom Linden" <tom@kednos.com> writes: > E > I don't disagree with what you say; however, when you are trying toxH > do this in C, you have to do a lot more work, and you still don't haveF > on conditions or automatic bounds/range checking.  So guess what, it > doesn't get done.  >   F One could stuff complex data types into structures holding address andF length (this is essentially what PL/I does) and invent a set of macrosF or functions to access the items. Put that into a personal toolbox and there we go.= PL/I gives this out of the box, howewer, it is not foolproof, @ since the concept of pointers (existing also in PL/I) allows you. to play tricks with those decriptors as well. H I remember that in my long gone PL/I days one could fool the compiler atK run time to accept a multi-dim array as a one-dimensional one or vice versaa  by manipulating the descriptors.^ Don't remember exactly the reason for doing so, could be performance or "ease of programming".   ------------------------------  # Date: Tue, 19 Oct 2004 17:32:32 GMT 4 From: "Fred Kleinsorge" <fred.nospam@nospam.dec.com>' Subject: Re: Question about Descriptors 1 Message-ID: <Qccdd.1150$6h6.126@news.cpqcorp.net>t  H When these things were designed (think PDP11's and VAX 780's) memory wasF expensive, so a lot of things in retrospect were dumb, like making theG length a word value instead of a longword.  Adding a second length onlybF makes sense if the descriptor is being dynamically resized/reused on aI frequent basis, and the allocation size itself might be itself a variable0H size.  Since that is not that common, you don't get what you would like.  G If you are, I would suggest that you allocate a block of memory 4 bytesaI longer than you need, and set the start if the descriptor to 4 bytes intocH the block of memory.  Store the initial allocation in the first 4 bytes.E You can fetch it as a negative offset to the start of the descriptor.t      : "JF Mezei" <jfmezei.spamnot@teksavvy.com> wrote in message& news:41747B97.93AD2602@teksavvy.com... > Tim Sneddon wrote:K > > What is wrong with using a DSC$K_CLASS_S string and passing the addresst ofI > > DSC$W_LENGTH for the length argument too? The following is some BLISSa code > > that explains what I mean: >aI > Yeah, but still requires that you hold the allocated size separately sos youfJ > can reset the Ddsc$w_length field after you'd used the field to indicate > current string length. >nL > I know how to work with the current design. But just wondered if there wasF > some reason they didn't put in the 2 lengts (used, allocated) in the originalE > description definitions which would have made for cleaner programs.-   ------------------------------  # Date: Tue, 19 Oct 2004 17:47:57 GMTu4 From: "Fred Kleinsorge" <fred.nospam@nospam.dec.com>' Subject: Re: Question about Descriptorsl1 Message-ID: <hrcdd.1151$Ei6.358@news.cpqcorp.net>,  L Consider usage patterns here.  The only real concern being expressed here isK where the string length changes frequently (more than once), and the memorycF for the string buffer itself is a variable length.  It's not the norm.: Strings are more likely to be static, or initialized once.  L In the case where someone wants to re-use a string descriptor and change theD string and/or it's length - it is easily dealt with by allocating anL additional longword to contain the allocation.  You could prepend or postfixH the descriptor itself, or the buffer the dsc$a_pointer points to.  Is itL non-standard?  Sure.  But I know I've done it at least for the last 15 yearsK when I've had the need to do something like this -- which is not too often.e        5 "David Froble" <davef@tsoft-inc.com> wrote in messageo& news:41753908.6080304@tsoft-inc.com... > JF Mezei wrote:4 >s > > Tim Sneddon wrote: > >nK > >>What is wrong with using a DSC$K_CLASS_S string and passing the addresst ofI > >>DSC$W_LENGTH for the length argument too? The following is some BLISS  code > >>that explains what I mean: > >> > >lK > > Yeah, but still requires that you hold the allocated size separately son youaL > > can reset the Ddsc$w_length field after you'd used the field to indicate > > current string length. > >tJ > > I know how to work with the current design. But just wondered if there was)H > > some reason they didn't put in the 2 lengts (used, allocated) in the originalG > > description definitions which would have made for cleaner programs.a > >e >eG > Just guessing here.  The use of dynamic strings seems to be in BASIC.r (NoterG > that I've not used all the languages that are/were available on VMS.)  Prior toL > the development of dynamic strings for BASIC, it's my impression that mostK > string usage (with descriptors) was static (fixed length) strings.  Beinge moreI > specific, the allocated memory for the actual data was fixed in length.  >hE > If there's any reality to my guess, then the development of dynamicu strings used@ > the available descriptors rather than develop new descriptors. >SH > I've heard, somewhere in the past, that when re-setting the value of a dynamicrE > string, that if the new data is larger, then the existing memory isn released andH > a new block of memory of sufficient size is acquired.  If smaller, the size ofaK > the data is set and the rest of the memory ignored.  I'm guessing that ifr that'sJ > so, that the memory would be in a location that is all released when the imagehH > is run down.  Still, a long running program would be 'leaking' memory. Neverc< > got into reading the source for the STR$ libraby routines. >vH > I can see the reasoning for 2 length values.  Such would keep track of totalaI > allocated memory for the string, thus never 'losing' any of the memory.o >tK > It also is a bit disconcerting when thinking about changing the length ofa aaA > string.  Even if the current memory is released, just what is as system/OS/image4H > going to do with many memory fragments each consisting of a few bytes? ConsiderF > changing the value (and length) of a string variable several million times. >i > Dave >n > -- l6 > David Froble                       Tel: 724-529-04506 > Dave Froble Enterprises, Inc.      Fax: 724-529-0596@ > DFE Ultralights, Inc.              E-Mail: davef@tsoft-inc.com > 170 Grimplin Roade > Vanderbilt, PA  15486t >r   ------------------------------  % Date: Tue, 19 Oct 2004 13:48:14 -0400x3 From: Steve Lionel <Steve.Lionel@REMOVEintelME.com>e' Subject: Re: Question about DescriptorsT8 Message-ID: <0ekan0tlgh375k9qgt1ea3ss409v2s7kss@4ax.com>  M On Tue, 19 Oct 2004 11:55:52 -0400, David Froble <davef@tsoft-inc.com> wrote:s  N >Just guessing here.  The use of dynamic strings seems to be in BASIC.  (Note Q >that I've not used all the languages that are/were available on VMS.)  Prior to  L >the development of dynamic strings for BASIC, it's my impression that most P >string usage (with descriptors) was static (fixed length) strings.  Being more H >specific, the allocated memory for the actual data was fixed in length. >tR >If there's any reality to my guess, then the development of dynamic strings used ? >the available descriptors rather than develop new descriptors.   J The great thing about the VAX Calling Standard was, and still is, that theL people who put it together thought ahead - an unusual quality, sadly lackingN in operating systems that came later. VAX BASIC was already being developed atM the time, and it was known that dynamic strings would be needed.  The VMS RTLyJ had dynamic string support in the guise of LIB$ routines from version 1.0.G When it was later released, VAX BASIC took advantage of dynamic stringsuK naturally. There was no need to create yet another descriptor class for it.e  M Later, when VAX Pascal came on the scene, the calling standard was amended to K provide for CLASS_VS descriptors.  The calling standard was revised severalaH times over the years to accomodate new requirements, but the fundamental structure was very sound.i  L The RTL kept its own dynamic string allocation size in "hidden" storage, andM it had an evolving strategy as to whether and how it would do reallocation int< an attempt to improve performance for the majority of cases.   Steveg   ------------------------------    Date: 19 Oct 2004 07:24:13 -0700" From: horn@shsu.edu (James T Horn)G Subject: Re: Routine to Resubmit print/batch entries from queue listing = Message-ID: <843706dc.0410190624.60bedb10@posting.google.com>u  h "Fred Hoenisch" <Fred.Hoenisch@gems9.gov.bc.ca> wrote in message news:<4173e87b$1@obsidian.gov.bc.ca>...K > Sorry, I see my wording wasn't very clear.  I specifically don't want thecL > routine to generate a queue listing file - I want to resubmit jobs from anI > older (historic) queue file (ie. I did the SHOW /QUEUE /FULL /ALL /OUT=n > earlier).  > $ > Thanks for the responses thus far. >  > Yours truly, > Fred.t  B You need to check out FIXQUE.COM. This is a command file that I'veC setup to run nightly, that will get a copy of the queues and createhC the files to recreate the queues. http://vms.tuwien.ac.at/freeware/w   ------------------------------  % Date: Tue, 19 Oct 2004 05:29:08 -0700e, From: James Nykiel <jnykiel@access4less.net># Subject: Re: RZ28B-VA drives wanteda< Message-ID: <pan.2004.10.19.12.29.05.339874@access4less.net>  
 Greetings:    J Thank you all for your assistance, I have obtained the (2) RZ28B-VA drives2 that I was looking for.  Please close this thread.     Thanks - Jim    7 On Sun, 17 Oct 2004 21:46:52 -0700, James Nykiel wrote:)   > Greetings: >  > I > I am assembling a BA353 StorageWorks enclosure for use with my hobbyistCH > system and would like to obtain (2) RZ28B-VA drives.  Unfortunately myA > budget is fairly limited at this time so try and keep the price : > affordable if you can.  Please email me with your offer. >  >  > Thanks - Jim   ------------------------------  # Date: Tue, 19 Oct 2004 12:30:10 GMTt4 From: "Fred Kleinsorge" <fred.nospam@nospam.dec.com>9 Subject: Re: Simulating monochrome to test Xwindows app ?e1 Message-ID: <mN7dd.1112$gX5.739@news.cpqcorp.net>o  : "JF Mezei" <jfmezei.spamnot@teksavvy.com> wrote in message& news:41742B1D.CDDACB38@teksavvy.com... >kI > So, is there some magic undocumented jumper that can be removed to makeb thea! > SPX into a 24 bit 3d card ? :-)  >   J This is so long ago, I can't tell you.  The answer is most likely no.  TheJ SPX was originally designed as a 24-plane 3D card.  It eventually was spunJ into the standard graphics offering.  Usually in those days, we would have* built two seperate boards for the purpose.   >iG > OK, I took a XPM generated pixmap converted to UIL file with a colour- table- > that uses colours such as :0 >0. > colour1: colour("#3F3F3F3F3F3F",foreground); >n) > Now, wouldn't those colours be 24 bit ?  >   J Pixels can be any number of bits.  But usually 1, 4, 8, 12, 16, 24, and 32I are the lengths.  The color format determines how to interpret them.  The-J visuals (use xdpyinfo to show you what you have) tell you the pixel formatJ for each.  For instance, a 8 bit pixel can be interpreted as pseudo-color,K or direct color, or grayscale, or truecolor - depending on the visual type.l  I > Also, I remember reading a bit about the issue of foreground/background  whichiJ > allows a bitonal server to decide whether a colour should be rendered as whiten > or black.n >h  F In the way-back machine, I remember something along those lines... butL frankly nobody has used monochrome monitors in sooo long, I can't even start to recall any of it.  J > I find it very odd that X would not have tackled colour depths properly,I > considering that Apple had colour support from teh start (even if earlyy macsG > didn't actually have colour, software supported colours and then, any  colour: > application ran fine on bitonal displays and vice versa.  L X11 simply opened up the entire toolbox to you, but provides no higher levelJ conversion.  For instance, if you have a image that is 24-bit data (packedI in 32-bit pixels - the norm) you need to A) make sure that the visual youoJ are using supports the format - and even make sure that the red/green/blueI data is in the right place in the pixel.  X11 won't convert a 8-bit imagen into a 24-bit one for example.  H I'm hazy on bitonal stuff (confusing 3 different window systems from theI past) in what is in X11 as a "standard" - I know there is device specificeH code for some old B&W flat panels that does it - but my hazy memory says this isn't standard.   ------------------------------  + Date: Tue, 19 Oct 2004 11:38:03 +0000 (UTC)h From: david20@alpha2.mdx.ac.uk+ Subject: Re: SYSUAF => export passwords ???h) Message-ID: <cl2uar$3ec$1@news.mdx.ac.uk>n  g In article <cl0mut$6sr$1@news.tiscali.fr>, "Alexandre Mongin" <NOSPAM_alexandre.mongin@csgv.fr> writes:-L > > >> I would like to export all user's passwords stored in the sysuaf from >aF >> >> first machine up to a second one, where the accounts are already >existing... >> >>h7 >> >> Is this possible through DCL or some free tools ?e >> > >>L >> There is a public domain tool called SYSUAF which has the ability to takeH >> encrypted passwords from one node and apply them to the corresponding	 >accountse >> on a remote node. >  >Thanks for all your answers,  > I >The one that best fits to my need is david20's... This 's a really greata8 >tool that I didn't still know (I'm still just a newbie) > 4 Did you use my copy or did you find a newer version.J The one problem with my copy is that the reporting by date options are not year 2k compliant.   >: >>< >> I've never used the password copying part of the program. >> >tM >It's already done and it was okay for me... I exported some passwords from ayK >first cluster to a second one, and everything went fine... interactively ! M >However, I'd wish to make it work automatically as a batch... but everythingCE >I tried in order to redefine the sys$command parameter needed by then# >SYSUAF.EXE just went wrong...  :-(r >e0 >Have you experienced to use this as a batch ??? >here's the procedure... >u >$SAVE_PRIV=F$SETPRV("BYPASS") >$ON ERROR THEN GOTO DONEb' >$! here I need to redefine SYS$COMMAND   
 Try insertinge  # $ define/user sys$command sys$inputn  ! >$ RUN TOOLS:TRANSFER_PWD_TO_NODE  >POLO # >$DONE: SET PROC/PRIV=('SAVE_PRIV')p >$DEASS/USER >s >e  
 David Webb Security Team Leader CCSS Middlesex University       >t$ >here's the source of the executable >t6 >        ! ----- READ NODENAME FROM SYS$COMMAND: -----) >        ERROR_LINE = "OPEN SYS$COMMAND:"M4 >        OPEN "SYS$COMMAND:" AS FILE #1%, SEQUENTIAL >u1 >        PRINT #1%, "Enter Node Name (as node::)"r > % >        ERROR_LINE = "READ NODENAME"R >        LINPUT #1%, NODE_NAME+ >        ERROR_LINE = "AFTER READ NODENAME"dE >        CLOSE #1%                               ! CLOSE SYS$COMMAND:  > G >Unfortunately, I haven't got any basic compiler so I can't even try to 2 >recode this ' ma propre sauce' (from the french) >gJ >Do I well understand that the exe's reading the nodename from SYS$COMMAND4 >??? And can it be modified in the VMS procedure ??? >r >Thank you all ! >  >e   ------------------------------  % Date: Tue, 19 Oct 2004 17:54:52 +0200 : From: "Alexandre Mongin" <NOSPAM_alexandre.mongin@csgv.fr>+ Subject: Re: SYSUAF => export passwords ???t* Message-ID: <cl3dcb$ba4$1@news.tiscali.fr>  6 > Did you use my copy or did you find a newer version.L > The one problem with my copy is that the reporting by date options are not > year 2k compliant. >sK I downloaded your copy from your ftp, but the reporting tool doesn't really_
 mind to me...rD I'm really surprised with the password exporting tool, which is very
 impressive...n   >t > Try insertingu > % > $ define/user sys$command sys$inputn >s already done... :-(t  I I've tried many, many things... and I can't make work the exe in a batch.   % $ CMD = "$TOOLS:TRANSFER_PWD_TO_NODE"n# $ DEFINE/USER SYS$COMMAND SYS$INPUTd $ CMDu NODE:: USER1m USER2  $t   it doesn't work.  , Unexpected error 136 in TRANSFER_PWD_TO_NODE# After Error Line: OPEN SYS$COMMAND:  Notify your VAX System Manager	 Username=- FLAGS(0%)= 0 UAF_SALT= 0t UAF_ENCRYPT= 0 ?Illegal or illogical access  ; every time I'm facing the same error after OPEN SYS$COMMANDu   any idea ???   thanks for your help   Alexandre M.   ------------------------------  + Date: Tue, 19 Oct 2004 17:38:37 +0000 (UTC)a From: david20@alpha2.mdx.ac.uk+ Subject: Re: SYSUAF => export passwords ???d) Message-ID: <cl3jet$ara$1@news.mdx.ac.uk>r  g In article <cl3dcb$ba4$1@news.tiscali.fr>, "Alexandre Mongin" <NOSPAM_alexandre.mongin@csgv.fr> writes:-7 >> Did you use my copy or did you find a newer version.tM >> The one problem with my copy is that the reporting by date options are noti >> year 2k compliant.4 >>L >I downloaded your copy from your ftp, but the reporting tool doesn't really >mind to me...E >I'm really surprised with the password exporting tool, which is veryi >impressive... >n >> >> Try inserting >>& >> $ define/user sys$command sys$input >> >already done... :-( >mJ >I've tried many, many things... and I can't make work the exe in a batch. >O& >$ CMD = "$TOOLS:TRANSFER_PWD_TO_NODE"$ >$ DEFINE/USER SYS$COMMAND SYS$INPUT >$ CMD >NODE::t >USER1 >USER2 >$ >u >it doesn't work.3 > - >Unexpected error 136 in TRANSFER_PWD_TO_NODEs$ >After Error Line: OPEN SYS$COMMAND: >Notify your VAX System Managerg
 >Username=
 >FLAGS(0%)= 0p >UAF_SALT= 0 >UAF_ENCRYPT= 0  >?Illegal or illogical accesso >w< >every time I'm facing the same error after OPEN SYS$COMMAND >,
 >any idea ???: >1H Seems to be the way it's opening and closing sys$command in order to get the nodename. 6 I've edited the source and recompiled to remove that. I Try downloading it from ftp://ftp.mdx.ac.uk/vms/transfer_pwd_to_node.exe t4 and ftp://ftp.mdx.ac.uk/vms/transfer_pwd_to_node.bas  " That version should then work with  % $ CMD = "$TOOLS:TRANSFER_PWD_TO_NODE"o $ CMDt NODE:: USER1s USER2s    # Give it a try and see if that works N (no guarantees but I can't see why it's explicitly opening sys$command for the nodename in the first place).a    
 David Webb Security team leader CCSS Middlesex University       >thanks for your help  >e
 >Alexandre M.0 >1 >7 >    ------------------------------  % Date: Tue, 19 Oct 2004 05:34:48 -0700Q, From: James Nykiel <jnykiel@access4less.net>/ Subject: Re: V6 OpenVMS Freeware CD's questions < Message-ID: <pan.2004.10.19.12.34.46.447579@access4less.net>  
 Greetings:    J Thank you all for your assistance, I have obtained the V6 OpenVMS Freeware7 CD's that I was looking for.  Please close this thread.r     Thanks - Jim  7 On Wed, 29 Sep 2004 21:07:49 -0700, James Nykiel wrote:e   > Greetings: >  > I > Does anyone know if you can purchase the V6 OpenVMS Freeware CD's (1-4) B > located at http://h71000.www7.hp.com/openvms/freeware/index.html > H > Unfortunately I am limited to a dial-up connection at this time and itE > would take me forever to download the files! - If you are unable tolI > purchase the CD's would someone with a high speed connection be willingsA > to download the files and burn the CD's for me for a small fee?  >  >  > Thank you... - James NykielA   ------------------------------    Date: 19 Oct 2004 10:30:07 -0700 From: vershinin-vk@mtu-net.rutD Subject: Re: VTfm - OpenVMS file manager for VT-compatible terminals= Message-ID: <b587d953.0410190930.3e9eba78@posting.google.com>n  e David J Dachtera <djesys.nospam@comcast.net> wrote in message news:<4174620A.E47B6086@comcast.net>...   > vershinin-vk@mtu-net.ru wrote: > > F > > I present the new release 2.2-6 of my freeware open source programI > > VTfm which is a Norton Commander-style file manager for OpenVMS. VTfme/ > > works on VAX, Alpha and IA64 under OpenVMS.  > > " > > You can download VTFM.ZIP fromJ > > http://eisner.encompasserve.org/~vershinin/vtfm.zip, unzip archive and > > read README.TXT in [.VTFM].r > E > I pulled the source for this down and uploaded it to HP's testdriveuI > cluster and compiled/linked it on both Alpha and Itanic. The process one > Itanic was MARKEDLY slower!  >  > -- e > David J Dachtera > dba DJE Systemsn > http://www.djesys.com/ > + > Unofficial OpenVMS Hobbyist Support Page:t$ > http://www.djesys.com/vms/support/ > * > Unofficial Affordable OpenVMS Home Page:! > http://www.djesys.com/vms/soho/   
 Hi, David!  > What do you mean: the compile/link process or running program?   Vladimir Vershinin vershinin-vk@mtu-net.ru0   ------------------------------   End of INFO-VAX 2004.581 ************************