1 INFO-VAX	Mon, 01 Dec 2003	Volume 2003 : Issue 665       Contents:5 Announcement OpenVMS backup solution special offer... 9 Re: Announcement OpenVMS backup solution special offer...  CDE Desktop App Mgr crashing  Re: CDE Desktop App Mgr crashing  Re: CDE Desktop App Mgr crashing' Re: defining SYS$OUTPUT for BACKUP/LIST ' Re: defining SYS$OUTPUT for BACKUP/LIST / Re: Display Postscript with MAC as X-terminal ? / Re: Display Postscript with MAC as X-terminal ? + Display Postscript with MAC as X-terminal ?  geld nodig? 6 I wonder if this HP director will resign from HP's BOD& LBR$ library routines and DCX encoding Re: My new Sun e250?  Re: NTP build procedure for VMS?  Re: NTP build procedure for VMS?  Re: NTP build procedure for VMS?. OT: Merger versus Takeover (Chrysler & Compaq)	 Re: Qio ?  Re: Rdb on ODS-5 disk  Re: Rdb on ODS-5 disk  Rdb on ODS-5 disk , Re: recommendations for page/swap/dump files, Re: recommendations for page/swap/dump files! Re: Subprocess termination oddity  Subprocess termination oddity  Re: TELENT in batch + Re: Telnet session with fixed TNAnnn: name? + Re: Telnet session with fixed TNAnnn: name? + Re: Telnet session with fixed TNAnnn: name? = Re: The DCL minute of the day: sh kno links under DECnet Plus  Re: VMS on PDP-10? RE: Wildcard searching3 X-Windows/Motif: Multihost multiwindows navigation.   F ----------------------------------------------------------------------   Date: 1 Dec 2003 08:15:02 -0800 1 From: susan_skonetski@hotmail.com (Sue Skonetski) > Subject: Announcement OpenVMS backup solution special offer...= Message-ID: <857e9e41.0312010815.2df6dc27@posting.google.com>   + Transition seamlessly to a high performance  backup solution for OpenVMS   A LEGATO, in collaboration with the HP OpenVMS Business Unit and HP D Services, is pleased to offer SLS and ABS customers LEGATO's OpenVMSA backup solution and solution training for a special, limited-time E discounted price. Additionally, HP Services has developed a Migration B Services offering to ensure a seamless transition from an existing9 backup solution to the LEGATO NetWorker OpenVMS solution.   0 Imagine a world where you have the ability to:    D        Centrally manage all of your backup and restores on OpenVMS !        Perform LAN-free backups  +        Consolidate your storage resources  E        Backup Oracle RDB and Oracle data without having to bring the  database application down @        Share your off-line storage resources in a heterogeneous environment F        Support disk-to-disk backup, with automatic migration to tape E        Automatically clone savesets or volumes for off-site copies.  A        Perform heterogeneous backups with one centralized backup  server        Support ODS-5         AND reduce your TCO.    B LEGATO and HP can make this vision become a reality. To learn moreF about LEGATO solutions for OpenVMS and a time-limited special offer to4 migrate to the LEGATO/HP solution for OpenVMS, go to www.legato.com/hp/openvms   @ If you would like to download a whitepaper on the LEGATO OpenVMS  solution click here for access. / http://www.legato.com/hp/openvms/whitepaper.cfm   ? We are currently enrolling BETA customers for the most complete F OpenVMS solution for ODS-5 and Oracle Support. To apply for the LEGATO% Oracle/ODS-5 Beta Program click here. ) http://www.legato.com/hp/openvms/beta.cfm   5 You can learn more about our beta program by visiting * www.legato.com/hp/openvms/index.cfm#beta .  9 Please send comments or questions to openvms@legato.com.  F Thank you for your interest in LEGATO software solutions for OpenVMS. @ Together, we look forward to improving your OpenVMS environment.   Sincerely,      F Shaun Ellis                                and                    Andy	 Schneider > Product Manager for OpenVMS Solutions         OpenVMS Business
 Management: LEGATO Software                                            Hewlett-Packard Company B 3210, Porter Drive                                             110 Spit Brook Road  ZKO3-4/T61 F Palo Alto, CA  94304                                        Nashua, NH  03062   ------------------------------  % Date: Mon, 01 Dec 2003 09:34:11 -0700 + From: "Barry Treahy, Jr." <Treahy@MMaz.com> B Subject: Re: Announcement OpenVMS backup solution special offer...' Message-ID: <3FCB6D83.1050409@MMaz.com>    VAXman@TMESIS.COM wrote:  L >>Nice message, but I don't understand why HP does not promote own products.J >>EMC/ Legato is one of HP's toughest competitors. The HP OpenView StorageN >>Data Protector should be positioned as the 1st choice for HP OpenVMS backup. >>Better Together! >>     >> > L >Question:  Can either product allow me to perform an image backup and, moreK >importantly, an image restore of the system disk?  If not, neither product  >is acceptable IMO.  >  >    > H No doubt the question was tailored to be VMS specific, but I would go a H step further and suggest that no backup system is of value if it cannot G backup and restore to the system disk of ANY operating system that the  & system is supposed to be 'protecting.'    H Therefore, if it is supposed to work with VMS, Unix, Linux, Windoze, or I any other OS of choice, it must be able to restore that system's OS disk  D to a fully functional state in a single-pass operation (ie. without  manual preloads, etc)...     Barry    --    > Barry Treahy, Jr                       E-mail: Treahy@MMaz.com> Midwest Microwave                          Phone: 480/314-1320> Vice President & CIO                         FAX: 480/661-7028                            ------------------------------    Date: 30 Nov 2003 12:36:34 -08009 From: ue191@vtn1.victoria.tc.ca (T.L. (Terry) Branscombe) % Subject: CDE Desktop App Mgr crashing * Message-ID: <3fca54d2@news.victoria.tc.ca>   OpenVMS Alpha 7.3, DECW 1.2-6   H I've had my hobbyist system up for over a year now, and thought I'd try F out DECwindows.  I've managed to get Mozilla, Java, and a few plugins G installed and working, and am now trying to register the Mozilla icons  E with HTML file types using the Advanced feature of the Create Action  " dialog in the Application Manager.  I The HP Mozilla site gives some detailed directions on getting this done,  H and I've followed them to the letter.  I click the "Advanced" button in I the "Create Action" dialog, then the "Add..." button.  This opens up the  J "Add Datatype" dialog where I see "Mozilla_FILE_1" already entered in the J "Name of datatype family" field.  The next step is to click the "Edit..." K button, but when I do this both dialogs just close, leaving me back at the  ! main Applications Manager dialog.   E I have security auditing turned on, but that shows nothing unusual.     2 Anyone have any ideas on what else I should check?   Thanks,  Alder    ------------------------------    Date: 30 Nov 2003 17:17:01 -08009 From: ue191@vtn1.victoria.tc.ca (T.L. (Terry) Branscombe) ) Subject: Re: CDE Desktop App Mgr crashing * Message-ID: <3fca968d@news.victoria.tc.ca>  4 Martin P.J. Zinser (zinser@zinser.no-ip.info) wrote: : H Vlems wrote:L : > "T.L. (Terry) Branscombe" <ue191@vtn1.victoria.tc.ca> schreef in bericht : > ! : >>OpenVMS Alpha 7.3, DECW 1.2-6  : >>L : > Well, on a VAX my first thought would be "insufficient memory". I'd lookK : > into SYSGEN at WSMAX and in Authorize for working set values, BYTLM and  : > possibly swapfile usage. : > H : VAX? I have not heard about Mozilla for VAX, so I think it is save to  : assume Alpha in this case.  E True, but I noticed in the DECWindows docs that my system parameters  K PQL_DASTLM and PQL_DBIOLM were set below the recommended values.  I bumped  < them up with SYSGEN and am waiting for the reboot to finish.  4 It might just work, but I'm not holding my breath...   Cheers,  Alder    ------------------------------  % Date: Sun, 30 Nov 2003 20:36:49 -0600 1 From: "David J. Dachtera" <djesys.nospam@fsi.net> ) Subject: Re: CDE Desktop App Mgr crashing ' Message-ID: <3FCAA941.74DCE64F@fsi.net>    "Martin P.J. Zinser" wrote:  >  > H Vlems wrote:L > > "T.L. (Terry) Branscombe" <ue191@vtn1.victoria.tc.ca> schreef in bericht( > > news:3fca54d2@news.victoria.tc.ca... > > ! > >>OpenVMS Alpha 7.3, DECW 1.2-6  > >>K > >>I've had my hobbyist system up for over a year now, and thought I'd try I > >>out DECwindows.  I've managed to get Mozilla, Java, and a few plugins J > >>installed and working, and am now trying to register the Mozilla iconsH > >>with HTML file types using the Advanced feature of the Create Action& > >>dialog in the Application Manager. > >>L > >>The HP Mozilla site gives some detailed directions on getting this done,K > >>and I've followed them to the letter.  I click the "Advanced" button in L > >>the "Create Action" dialog, then the "Add..." button.  This opens up theM > >>"Add Datatype" dialog where I see "Mozilla_FILE_1" already entered in the M > >>"Name of datatype family" field.  The next step is to click the "Edit..." N > >>button, but when I do this both dialogs just close, leaving me back at the% > >>main Applications Manager dialog.  > >>G > >>I have security auditing turned on, but that shows nothing unusual.  > >>6 > >>Anyone have any ideas on what else I should check? > >> > >>Thanks, 	 > >>Alder  > >  > > L > > Well, on a VAX my first thought would be "insufficient memory". I'd lookK > > into SYSGEN at WSMAX and in Authorize for working set values, BYTLM and  > > possibly swapfile usage. > >  > > G > VAX? I have not heard about Mozilla for VAX, so I think it is save to  > assume Alpha in this case.   Also, no CDE on VAX, AFAIK.    --   David J. Dachtera  dba DJE Systems  http://www.djesys.com/  ( Unofficial Affordable OpenVMS Home Page: http://www.djesys.com/vms/soho/    ------------------------------   Date: 1 Dec 2003 07:38:04 -0800 . From: spamsink2001@yahoo.com (Alan E. Feldman)0 Subject: Re: defining SYS$OUTPUT for BACKUP/LIST= Message-ID: <b096a4ee.0312010738.60aac5df@posting.google.com>   | helbig@astro.multiCLOTHESvax.de (Phillip Helbig---remove CLOTHES to reply) wrote in message news:<bq736h$m4b$1@online.de>...I > Why does DEFINE/USER_MODE before BACKUP/LIST not work (i.e. if defined  J > to a file output still goes to the screen) while DEFINE/SUPERVISOR does / > work (with a subsequent DEASSIGN, of course).   C Works for me! (VMS v6.1 and v6.2 on VAX systems) Of course, you can / use BACK/LIST=filename as others have suggeted.      Alan E. Feldman    ------------------------------  % Date: Sun, 30 Nov 2003 17:42:20 -0600 % From: "Mike Naime" <mnaime@kc.rr.com> 0 Subject: Re: defining SYS$OUTPUT for BACKUP/LIST: Message-ID: <gfvyb.102910$M02.16518@twister.rdc-kc.rr.com>  J Define/Usermode is necesary before calling any other COM processes from anK initial script where you do not want to loose the connectivity between your F VT/Keyboard and the process that it is running.  I'm not positive, butJ Define/User mode may be expecting a com proc call to affect.  Not the next line in this com process.   F Backup/List/Output=DEV:[DIR]File.name is related to the backup command# itself and is what you want to use.     J Phillip Helbig---remove CLOTHES to reply <helbig@astro.multiCLOTHESvax.de>/ wrote in message news:bq736h$m4b$1@online.de... H > Why does DEFINE/USER_MODE before BACKUP/LIST not work (i.e. if definedI > to a file output still goes to the screen) while DEFINE/SUPERVISOR does / > work (with a subsequent DEASSIGN, of course).  >    ------------------------------   Date: 1 Dec 2003 05:01:57 -0800 . From: martinkirby12@yahoo.co.uk (Martin Kirby)8 Subject: Re: Display Postscript with MAC as X-terminal ?< Message-ID: <224291b.0312010501.44ad8e81@posting.google.com>  B There were two shared images on the client side that were removed:% XDPS$DPSLIBSHR and XDPS$DPSCLIENTSHR.   I Therefore, you would probably need to stick with a version of DECwindows    that still had those libraries.    Martin Kirby  [ JF Mezei <jfmezei.spamnot@istop.com> wrote in message news:<3FCADF86.5C21A4DF@istop.com>... P > I just read somewhere that current MACs come with Display Postcript, somethingP > they inherited from next. (heard it from someone who heard it from someone who > heard it from someone...)  > P > If that is the case, does this mean that one could still run applications suchO > as DECwrite and the CDA viewer with the output going to a X-terminal on a MAC ) > even though VMS no longer supports it ?  > N > Or must display postscript need to exist on both the client and server (in X
 > parlance) ?    ------------------------------  # Date: Mon, 01 Dec 2003 17:34:42 GMT " From:   VAXman-  @SendSpamHere.ORG8 Subject: Re: Display Postscript with MAC as X-terminal ?0 Message-ID: <00A29B8F.FB9F055A@SendSpamHere.ORG>  b In article <011220031217169852%paul.anderson@hp.com>, Paul Anderson <paul.anderson@hp.com> writes:3 >In article <3FCADF86.5C21A4DF@istop.com>, JF Mezei # ><jfmezei.spamnot@istop.com> wrote:  > Q >> I just read somewhere that current MACs come with Display Postcript, something  >> they inherited from next. > A >I don't think Macs come with Display PostScript.  PDF support is @ >built-in, though.  You can save files as PDF or PostScript from >applications, for example.  > B >If you double-click on a PostScript file, you invoke the TextEditE >application which automatically translates your PostScript file into  >PDF.  >  >Paul  >  >--  > Paul Anderson  >  OpenVMS Engineering >  Hewlett-Packard Company   I just issued:  / $ SET DISPLAY/CREATE/NODE=APBG4/TRANSPORT=TCPIP = $ VIEW/INTERFACE=DECWINDOWS/FORMAT=PS SYS$MANAGER:EYECHART.PS   H The CDA viewer started on the PowerBook but a viewer message window pops> up with the message: Server does not have Postscript Extension  I I ran the VIEW command from a V7.1-2 system that still has the DECWindows  with the PS extensions.    --  L VAXman- A Bored Certified VMS Kernel Mode Hacker    VAXman(at)TMESIS(dot)COM             5   "Well my son, life is like a beanstalk, isn't it?"     ------------------------------  % Date: Mon, 01 Dec 2003 01:29:02 -0500 * From: JF Mezei <jfmezei.spamnot@istop.com>4 Subject: Display Postscript with MAC as X-terminal ?) Message-ID: <3FCADF86.5C21A4DF@istop.com>   N I just read somewhere that current MACs come with Display Postcript, somethingN they inherited from next. (heard it from someone who heard it from someone who heard it from someone...)   N If that is the case, does this mean that one could still run applications suchM as DECwrite and the CDA viewer with the output going to a X-terminal on a MAC ' even though VMS no longer supports it ?   L Or must display postscript need to exist on both the client and server (in X parlance) ?    ------------------------------  # Date: Mon, 01 Dec 2003 05:15:41 GMT  From: "geert" <geert@dds.nl> Subject: geld nodig?2 Message-ID: <18Ayb.36740$LC2.8167@typhoon.bart.nl>   http://www.geld-winkel.nl/   ------------------------------  # Date: Mon, 01 Dec 2003 14:39:43 GMT # From: "John Smith" <a@nonymous.com> ? Subject: I wonder if this HP director will resign from HP's BOD H Message-ID: <PoIyb.49855$ZmO.14621@news01.bloor.is.net.cable.rogers.com>  
 NEWS ALERT from The Wall Street Journal   Dec. 1, 2003  I Boeing Co. announced the resignation of Chairman and Chief Executive Phil G Condit, 62 years old, saying "a new structure for the leadership of the D company is needed." The aerospace manufacturer has been plagued withJ controversy over some of its government contracts and the hiring of an AirJ Force official. Director Lewis Platt, 62, was named nonexecutive chairman.J Retired operating chief and Vice Chairman Harry Stonecipher, 67, was named president and chief executive.   ------------------------------  % Date: Sun, 30 Nov 2003 18:17:40 -0500 * From: JF Mezei <jfmezei.spamnot@istop.com>/ Subject: LBR$ library routines and DCX encoding ) Message-ID: <3FCA7A8A.1E98B581@istop.com>   N Does anyone know where/how I can find out if a library has the DCX compression enabled or not ?  M I have looked at the 20 longword vector supplied with the LBR$OPEN as well as L the 128 longword vector supplied by LBR$GET_HEADER but there doesn't seem toC be any documented bit indicating if a library is compressed or not.    Where it is secretely hidden ?  M Oh, and by the way, the LHIDEF.H module in the DECC software contains a wrong L structure definition. It only defines the documented longwords at the top ofE the structure and doesn't allocate space for the remainder of the 128 L longwords, so supplying such a structure as argument to LBR$GET_HEADER wouldM corrupt memory since the routine would write data well beyond your structure.   C (in the CREDEF module, the proper padding is present in the defined L structure). However, in none of those 2 modules, are the offsets defined, as3 per the LBR$ documentation in the utility routines.     N (It wasn't very obvious for me to find out that LHIDEF and CREDEF were tied toH the LBR$ routines since there are not references by the LBR*.h modules.)   ------------------------------  % Date: Mon, 01 Dec 2003 11:47:24 +0000 O From: Andrew Harrison SUNUK Consultancy <Andrew_No.Harrison_No@nospamn.sun.com>  Subject: Re: My new Sun e250? 0 Message-ID: <bqf9od$a8h$1@new-usenet.uk.sun.com>   Neil Rieck wrote: J > My employer just took delivery of a new Sun e250 for a project where theL > application vendor required it. It came with two CPUs (450 MHz), 2 Gigs ofN > RAM, and six 36 Gig drives. We've been playing with it for the past few daysP > and noticed something strange: CPU utilization is almost totally maxed out forO > quite some time whenever we pull out, then replace, one of the RAID-5 drives. O > (remember, there is no application software installed yet and no one else was  > logged on) > M > After a little head scratching we came to the realization that this must be N > some kind of host-based RAID. We don't see this type of CPU bog down when weD > do something similar on my AlphaServer DS20 which has and externalE > hardware/based RAID controller (a single channel RA-3000 (HSZ-22)).  >   @ The e250 doesn't have hardware RAID support so the system CPU(s)? will be responsible for rebuilding the failed drive in the RAID * 5 volume which is non trivial with RAID 5.  D The e250 has 6 internal drives bays hence the 6 x 36 Gig drives that you have been shipped.  ? One option if you don't need disk capacity would be to only use A RAID 0+1 on the internal drives, this would reduce your available @ space to 106 GB but recreation of a failed drive in a 0+1 set is= much simpler than in a RAID 5 set (its basically just a block  copy).  L > Does Sun make/support a hardware based RAID controller or is this the best > that they've got?  >   < You can add external SCSI based RAID controllers to the E250= The 3310 or the A1000 both support hardware RAID controllers.   > The 3310 is supported on the E250 but you will require Solaris# 8/9 and a Ultra160 LVD PCI adaptor.   = The A1000 I am pretty sure will work with Solaris 2.6 as well  as 8/9 but its more expensive.  . http://www.sun.com/storage/workgroup/index.xml     regards  Andrew Harrison    > Neil Rieck > Kitchener/Waterloo/Cambridge,  > Ontario, Canada.# > http://www3.sympatico.ca/n.rieck/ : > http://www3.sympatico.ca/n.rieck/links/cool_openvms.html >  >    ------------------------------  % Date: Sun, 30 Nov 2003 15:22:44 -0500 3 From: "Richard B. Gilbert" <rgilbert88@comcast.net> ) Subject: Re: NTP build procedure for VMS? 0 Message-ID: <UvOdnbFNPboKzFeiRVn-tw@comcast.com>  D The obvious difference between your system and mine is that you are I running the UNPATCHED version of UCX V4.2.  Mine is patched to ECO 5.  I  F seem to recall that there were good reasons for patching UCX V4.2; to H put it bluntly, there have been good reasons for patching every version E of UCX (and TCPIP Services) that I have ever encountered.(2.0E, 3.0,  5 3.1, 3.2, 3.3, 4.0, 4.1, 4.2, 5.0, 5.0A, and 5.1)!!!!   G The other likely source of difference is your UCX$NTP.CONF file.  Mine   contains, at the moment: peer rfdb08  peer rfdb09  correct-any   I On the rare occasions when it deigns to notice that the two peers exist,  C it announces that it has synchronized with xxx.xxx.xxx.xxx.  A few  H minutes later it annouces that it has lost contact with xxx.xxx.xxx.xxx F and apparently never attempts to regain contact.   xxx.xxx.xxx.xxx is G present, working, peering with another VMS V7.2-1 machine, and serving  I about ten other systems: VMS V7.2-1, Solaris 8 Intel Platform, W2K,  and  D VMWare (hacked Red Hat Linux) all of which synchronize and maintain E synchronization without problems.  The lack of timestamps in the log  " file is another thorn in my flesh.   Bradford J. Hamilton wrote:   T >In article <bqciub$7ke$1@news1.tilbu1.nb.home.nl>, Dirk Munk <munk@home.nl> writes: >!Richard B. Gilbert wrote:  >!snip! ( >!> Do you actually have this working??? >!> R >!I will have to look at some systems at my work how they are configurered. We do N >!have some older system running VMS 6 and UCX 4.2, but I hardly ever look at P >!them. However I'm quite sure they do use NTP for their time source. I hope to ' >!come back to you soon with an answer.  >! > N >I've been watching this thread with some interest.  I have a box at work that+ >seems to have no problem as an NTP client:  >  >XXXXX::BRADH$ ucx sho vers  > 9 >  DIGITAL TCP/IP Services for OpenVMS Alpha Version V4.2 7 >  on a AlphaServer 2000 4/275 running OpenVMS V6.2-1H1  > $ >...and a snippet of the log file... > 3 > 07:30:12.46    -0.000528 seconds, new tick   9764 3 > 07:31:16.47    -0.000120 seconds, new tick   9764 3 > 07:33:24.44    +0.000294 seconds, new tick   9766 3 > 07:34:28.45    -0.000612 seconds, new tick   9764 > > 07:35:32.46    -0.001168 seconds, discarded, trip > 0.010000? > 07:37:40.45    +0.001715 seconds, filtered, aperture 0.001532 3 > 07:39:48.45    -0.000293 seconds, new tick   9764 A > 07:40:52.45    -0.001655 seconds, new tick   9764 for 2 seconds A > 07:41:56.46    -0.000810 seconds, new tick   9764 for 1 seconds  > O >I will admit, the adjustments come frequently (more than I would like to see), = >but I don't think it has an adverse effect upon performance:  >   >XXXXX::BRADH$ sh sys/proc=*ntp*N >OpenVMS V6.2-1H1  on node XXXXX  30-NOV-2003 07:47:49.11  Uptime  16 23:36:17N >  Pid    Process Name    State  Pri      I/O       CPU       Page flts  PagesN >00000118 UCX$NTPD        HIB      7   199658   0 00:04:12.73       333    101 > M >Bottom line:  The system time stays within a second of the NTP server system  >time. > K >__________________________________________________________________________ B >Bradford J. Hamilton                    "All opinions are my own"L >bMradAhamiPltSon-at-coMmcAast.nPeSt     "Lose the MAPS, and replace '-at-' 1 >                                         with @"  >    >    ------------------------------  # Date: Sun, 30 Nov 2003 20:53:08 GMT 4 From: brad@.gateway.2wire.net (Bradford J. Hamilton)) Subject: Re: NTP build procedure for VMS? 0 Message-ID: <UMsyb.260986$mZ5.1922501@attbi_s54>  f In article <UvOdnbFNPboKzFeiRVn-tw@comcast.com>, "Richard B. Gilbert" <rgilbert88@comcast.net> writes:E !The obvious difference between your system and mine is that you are  J !running the UNPATCHED version of UCX V4.2.  Mine is patched to ECO 5.  I G !seem to recall that there were good reasons for patching UCX V4.2; to  I !put it bluntly, there have been good reasons for patching every version  F !of UCX (and TCPIP Services) that I have ever encountered.(2.0E, 3.0, 6 !3.1, 3.2, 3.3, 4.0, 4.1, 4.2, 5.0, 5.0A, and 5.1)!!!! ! H !The other likely source of difference is your UCX$NTP.CONF file.  Mine  !contains, at the moment:  !peer rfdb08 !peer rfdb09 !correct-any !   M Correct on both counts - but I don't remember your mentioning that your .CONF L files were doing "peering" - I have no experience with this behavior, and ofL course cannot give you good advice about what the source of your problem mayO be, at this point.  I always set up my systems to point to servers, rather than O attempting to peer (my systems are inevitably "old" and "slow", and I would not 6 want them attempting to update newer, faster systems).  M That said, there are probably good reasons for "peering" - my systems are not M servers that provide time to other servers, so "peering" is not a good choice = for me, according to Donald Mills' notes for configuring NTP.   L I "assumed" that your configuration was similar to mine - a bad assumption.  Sorry.   !snip!  J __________________________________________________________________________A Bradford J. Hamilton                    "All opinions are my own" K bMradAhamiPltSon-at-coMmcAast.nPeSt     "Lose the MAPS, and replace '-at-'  0                                          with @"   ------------------------------  % Date: Mon, 01 Dec 2003 12:15:19 -0500 & From: David M Smith <dsmit115@csc.com>) Subject: Re: NTP build procedure for VMS? 8 Message-ID: <sitmsvo5qm6mvp434e55h5gung1qu5js2q@4ax.com>  8 On Sun, 30 Nov 2003 15:22:44 -0500, "Richard B. Gilbert" <rgilbert88@comcast.net> wrote:   E >The obvious difference between your system and mine is that you are  J >running the UNPATCHED version of UCX V4.2.  Mine is patched to ECO 5.  I G >seem to recall that there were good reasons for patching UCX V4.2; to  I >put it bluntly, there have been good reasons for patching every version  F >of UCX (and TCPIP Services) that I have ever encountered.(2.0E, 3.0, 6 >3.1, 3.2, 3.3, 4.0, 4.1, 4.2, 5.0, 5.0A, and 5.1)!!!!  N My system is running UCX V4.2 patched to ECO 4 (haven't had a chance to put on ECO 5 and reboot).  H >The other likely source of difference is your UCX$NTP.CONF file.  Mine  >contains, at the moment:  >peer rfdb08 >peer rfdb09 >correct-any >   , My .CONF file has only one uncommented line:   peer 20.1.10.27   : while my .LOG file shows corrections being made routinely:  A  11:09:48.88    -0.002441 seconds, new tick  99999 for 34 seconds A  11:14:04.88    -0.001370 seconds, new tick  99999 for 19 seconds =  11:22:37.04    -0.064702 seconds, discarded, trip > 0.021827 A  11:39:40.92    -0.004558 seconds, new tick  99999 for 65 seconds A  11:56:44.94    -0.001393 seconds, new tick  99999 for 19 seconds A  12:05:20.94    -0.001709 seconds, new tick  99999 for 24 seconds   P According to the V4.2 documentation that I looked at, the only entries supported in the .CONF file are:  & 8.2.1 Creating the Configuration File  	...1 Valid configuration file entries are as follows:    $ peer [internet_address] [version n] & server [internet_address] [version n]  local-master [stratum]   master-clock [stratum]    O I don't see the correct-any that you have in yours -- where is that documented? I ------------------------------------------------------------------------- I David M. Smith 302.391.8533                       dsmit115 at csc dot com I Computer Sciences Corporation     (Opinions are those of the writer only) I -------------------------------------------------------------------------    ------------------------------  % Date: Mon, 01 Dec 2003 03:26:33 -0500 * From: JF Mezei <jfmezei.spamnot@istop.com>7 Subject: OT: Merger versus Takeover (Chrysler & Compaq) ) Message-ID: <3FCAFB0A.E817C9A6@istop.com>   N BBC reports that this week, a lawsuit by Chrysler shareholders will proceed inM US courts. They alledge that the union with Daimler-Benz had been touted as a G MERGER of equals, but was in fact implemented as a straight takeover of  Chrysler by Daimler.  M Not sure about could be done should shareholders win. But I found this tidbit M interesting in the context that HP and Compaq were also touted as a merger of ? equals but in fact, it was a straight takeover by HP of Compaq.   I Perhaps the difference here is that Compaq shareholders are probably very N happy to have been given a chance to dump their Compaq stock and don't care if Compaq ceases to exist.    ------------------------------   Date: 1 Dec 2003 04:31:21 -0800 - From: denis.fayaud@netspace.mc (Denis Fayaud)  Subject: Re: Qio ?= Message-ID: <93820504.0312010431.4fc3e131@posting.google.com>   0 First of all, thanks for all the contributions !* I solved the problem in a very simple way.9 It works fine. The code below is a prototype (in Pascal): A it reads the serial line char by char and writes the message on a  terminal line    by line.     [inherit('SYS$LIBRARY:STARLET')]! PROGRAM CAMDEV_AST(output,input);      TYPE     $WORD=[WORD] -32768..32767;      $BYTE=[BYTE] 0..255;)     nodlin_type =   VARYING [80] OF CHAR; #     io_status       = PACKED RECORD +                          Status_code:$WORD; +                          Byte_count :$WORD; .                          device_dep : INTEGER;                       END;      VAR      stat	: INTEGER:=0;5     ACC_channel : [VOLATILE] $word;   { QIO Channel }      camdev_file : TEXT; 9     INbuffer    : [VOLATILE] PACKED ARRAY [1..1] OF CHAR; +     message 	: [VOLATILE] nodlin_type:='';	 &     first	: [VOLATILE] boolean:=true;	  E {-------------------------------------------------------------------}  [ASYNCHRONOUS] PROCEDURE display_message;    BEGIN    IF (INbuffer <> chr(13)) THEN ;     IF (INbuffer <> chr(10)) THEN message:=message+INbuffer       ELSE BEGIN        writeln(message); #       IF length(message) >= 15 THEN 1        IF SUBSTR(message,12,4)='EXIT' THEN $exit;        message:='';        END;     END; E {-------------------------------------------------------------------}  [ASYNCHRONOUS]  PROCEDURE Rec;      VAR    stat            : INTEGER:=0; 4   IO_Status_Block : io_status;		{ QIO status Block };   OUTbuffer       : [VOLATILE] PACKED ARRAY [1..1] OF CHAR; 7   terminator_mask : [VOLATILE] ARRAY [1..2] OF INTEGER;      BEGIN       terminator_mask[1]:=0; 1    terminator_mask[2]:=8192;           (* <CR> *)       OUTbuffer[1]:=chr(0);         stat:=$QIO  ( 2, 		    acc_Channel ,  		    IO$_READPROMPT,  		    IO_status_block,2                	    ASTADR:= %IMMED IADDRESS(rec), 	            P1:=INbuffer, 	            P2:=1,   . 	            P4:=IADDRESS(terminator_mask[1]),+       	            P5:=IADDRESS(OUTbuffer),  	            P6:=1);       IF NOT first THEN 
      BEGIN       $clref(2);       display_message;      END ELSE first:=false;     END;   E {-------------------------------------------------------------------}  { main }   BEGIN C   WRITELN(' ',chr(27),'[4mCAMDEV$INTERFACE (02/01)',chr(27),'[0m'); *   writeln('  NO timeout - Infinite loop');
   Writeln;=   Writeln(' ATTENTION !!! CAMDEV doit etre en mode PASTHRU'); -   Writeln(' $ set term/perm/pasthru camdev'); 
   Writeln;(   stat := $Assign('CAMDEV',ACC_channel);   rec;
   $HIBER;  END.   @ As one sees, the process is HIB and interrupted by the QIO func.   ------------------------------   Date: 1 Dec 2003 08:45:59 -0600 - From: Kilgallen@SpamCop.net (Larry Kilgallen)  Subject: Re: Rdb on ODS-5 disk3 Message-ID: <PX6FxjQQD9LZ@eisner.encompasserve.org>   i In article <bqf8nl$ur3$1@bozon2.softax.com.pl>, Robert Trawinski <robert.trawinski@softax.com.pl> writes:   6 > Is it possible to create Rdb database on ODS-5 disk?   So long as:   - 	1. You run a sufficiently new version of Rdb    and ? 	2. You restrict the names of files Rdb uses to the ODS-2 rules   7 > We need to change disc structure from ODS-2 to ODS-5.   5 I would be curious to know why you "need" to do that.   G In my own situation, I "need" to have an ODS-5 disk to ensure that code  I write will work with ODS-5.    ------------------------------   Date: 1 Dec 2003 08:56:43 -0800 0 From: keith.cayemberg@conti.de (Keith Cayemberg) Subject: Re: Rdb on ODS-5 disk= Message-ID: <3a65a5c8.0312010856.172a21ab@posting.google.com>   > Some qualified responses to this question kan be found in the / Rdb JCC ListServer. Here a few good examples...   9 http://www.jcc.com/search/rdblistserver/2002/3c43e05e.htm   < http://www.jcc.com/search/rdblistserver/current/3f37f67c.htm  9 http://www.jcc.com/search/rdblistserver/2000/msg.ra93.htm   9 http://www.jcc.com/search/rdblistserver/2002/3d680539.htm   < http://www.jcc.com/search/rdblistserver/current/3f70c8cf.htm  ; You can search for more info in the JCC Search Engine at...   + http://www.jcc.com/searchengine/default.asp    Cheers!    Keith Cayemberg  ICA GmbH, Hannover, Germany     n Robert Trawinski <robert.trawinski@softax.com.pl> wrote in message news:<bqf8nl$ur3$1@bozon2.softax.com.pl>...	 > Hi all,  > 6 > Is it possible to create Rdb database on ODS-5 disk?I > We need to change disc structure from ODS-2 to ODS-5. Most of database   > files resideI > on ODS-2 disc that need to be converted to ODS-5. Some files reside on e > ODS-2 disc that won't beC > converted. Is it posiible to have one database on two discs with g > different structures?o >  > Thanks in advancee >  > Robert   ------------------------------  % Date: Mon, 01 Dec 2003 12:29:49 +0100 7 From: Robert Trawinski <robert.trawinski@softax.com.pl>i Subject: Rdb on ODS-5 disk/ Message-ID: <bqf8nl$ur3$1@bozon2.softax.com.pl>e   Hi all,   4 Is it possible to create Rdb database on ODS-5 disk?G We need to change disc structure from ODS-2 to ODS-5. Most of database l files resideG on ODS-2 disc that need to be converted to ODS-5. Some files reside on ? ODS-2 disc that won't beA converted. Is it posiible to have one database on two discs with   different structures?O   Thanks in advancem   Robert   ------------------------------   Date: 1 Dec 2003 07:57:22 -0600s From: briggs@encompasserve.org5 Subject: Re: recommendations for page/swap/dump filesc3 Message-ID: <d9zFnlrdoVLJ@eisner.encompasserve.org>   V In article <3FC86F35.36C5E673@istop.com>, JF Mezei <jfmezei.spamnot@istop.com> writes:1 > Phillip Helbig---remove CLOTHES to reply wrote:nP >> > Shadow Set - though I would assume at least better if not equal performance >> > than a single disk. >> -H >> I'm sure that WRITEs are slower, since it has to write to both disks. >  > It depends on your setup.m > O > Reads are definitely faster because a host will send read IO to another drives( > if its own drive is already very busy. > M > Writes are slower when one of the drives is not controlled by the same nodeeP > (MSCP). But if they are all physically attached to the same computer, wouldn'tP > a write be more or less about the same speed since both drives would get theirP > write command at about the same time and would complete at about the same time% > ? (no MSCP over ethernet overhead).S  K I agree.  The times should be comparable.  But bear in mind that when doingsB two writes, you're waiting for the slower of the two writes.  HeadD positioning, rotational latency and queue depth may all work to make- the access time for the two drives different.   E If you have write-back caching, your effective write access time goesd: to (close to) zero and you're bound by throughput instead.  G If you're measuring throughput, you should be fair and compare a single D page file on RAID 0+1 to a set of four page files on JBOD.  The JBODE setup should win handily for anything but a single-threaded workload.p   	John Briggs   ------------------------------  % Date: Mon, 01 Dec 2003 08:16:25 -0600 ( From: brandon@dalsemi.com (John Brandon)5 Subject: Re: recommendations for page/swap/dump filese1 Message-ID: <03120108162554@dscis6-0.dalsemi.com>i  1 > Phillip Helbig---remove CLOTHES to reply wrote:t   I like that!  LOLa  Q > > > Shadow Set - though I would assume at least better if not equal performance  > > > than a single disk.  > > I > > I'm sure that WRITEs are slower, since it has to write to both disks.k >  > It depends on your setup.g  K It sure does.  My understanding that with StorageWorks (using HSJ, G, etc.) J dual controllers with RAID 0+1 the writes were split and subsequently much/ faster than a single write.  Again, your setup.c  F > Just a hobbyist system, fast narrow SCSI 2 on the ALPHA, slow narrowH > SCSI 1 on the VAXes, host-based controllers (i.e. the SCSI cards whichJ > came with the machines), locally connected SCSI disks (which can be MSCP > served to the cluster).G  ( Ouch.  However as you say, hobbyist.  :)     J*o*h*n B*r*a*n*d*o*no VMS Systems Administratorn* firstname.lastname.spam.me.not@dalsemi.com   ------------------------------   Date: 1 Dec 2003 08:42:26 -0800c. From: spamsink2001@yahoo.com (Alan E. Feldman)* Subject: Re: Subprocess termination oddity= Message-ID: <b096a4ee.0312010842.2c00b4d0@posting.google.com>   [ JF Mezei <jfmezei.spamnot@istop.com> wrote in message news:<3FCB2086.F627E9B5@istop.com>... 3 > I had the following interactive process hiearchy:  >  > JFMEZEI---+--- A12
 >           |- >           +--- LIBRa > 3 > JFMEZEI was running a TPU session editing FILE1.C, > A1 was running ALLIN1 0 > LIBR was running a TPU session editing FILE2.C > L > This was on a VT terminal on a TX serial line, so I was switching back andF > forth with the ATTACH command available to both TPU and ALLIN1. ThisF > particular set of processes was probably running for a week or more. > O > Then, a batch job to do ALLIN1 house cleaning started, and that one tells anyVL > process running ALLIn1 that it should exit ALLIN1. So the "A1" process notJ > only exited the OA$MAIN.EXE image, but also logged out since it has been? > spanewd only to run ALLIN1. (this is the expected behaviour).I > J > Normally, with just a JFMEZEI (in DCL) and A1 processes, one would see aN > message such as "ALL-IN-1 terminating" folllowed by the "Control returned to. > process JFMEZEI" as the A1 process vanished. > J > However, this time, I was actually in the LIBR process putting finishing" > touches to a set of subroutines. > O > The interesting results were "interesting". Every second key would be sent toeM > JFMEZEI and the other to LIBR.  Not only did it corrupt files because I wasnM > trying commands which would get half their characters typed in to the otherrK > file, and <CTRL-W> would alternately display the TPU screen of one or the M > other process. And there was no way to know if typing <CTRL-z> would end upF > saving one file or the other.  > N > My interpretation, after the fact was that when the A1 image got the messageI > to shutdown, it did so and since the process had nothing else to do, itpI > terminated and returned control to its parent (JFMEZEI).  But since thelN > current subprocess was LIBR, I guess VMS got confused and made both JFMEZEI L > and LIBR foreground. As a result, both had outstanding $QIOs to read input< > from terminal, hence, each process got half my keystrokes. >  > @ > The command procedure to start the A1 process from JFMEZEI is: > $SET NOCONTROL=Y > $A1 :== "ATTACH A1"t > $SPAWN/PROC=A1 ALLIN1/REENTERt > $SET CONTROL=Y > $EXITt >  > M > I assume this is just some odd event caused by weird alignment of the starscP > and that no engineer will get a heart attack when he realises it is a bug deep+ > down in the process management kernel :-)M >   F I tried this without Allin1. I spawned 2 subprocesses, attached to oneD of them, and killed the other from another terminal and got the sameA behavior (VMS v6.2). Every press of the Return key sent me to theaE other process. I tried it again by editing two files with TPU and gotiF behavior to what you got, but not quite the same. It would take me twoE control/W's to switch. Typing would be okay, but navigating the files F was a nightmare! A Next Screen would give an "Attempt to move past the= end of buffer NOTHING.COM" which is the other processes file!c  F So it has nothing to do with Allin1, nor the stars and such. It's just the way it works!   D I vaguely remembering something like this happening to me in the 2nd  half of the 80's (VMS v4 or v5).   > # > And yes, this is with VAX VMS 7.2V   Disclaimer: JMHO Alan E. Feldmanc   ------------------------------  % Date: Mon, 01 Dec 2003 06:06:42 -0500n* From: JF Mezei <jfmezei.spamnot@istop.com>& Subject: Subprocess termination oddity) Message-ID: <3FCB2086.F627E9B5@istop.com>a  1 I had the following interactive process hiearchy:o   JFMEZEI---+--- A1w           |y           +--- LIBR   1 JFMEZEI was running a TPU session editing FILE1.Cd A1 was running ALLIN1 . LIBR was running a TPU session editing FILE2.C  J This was on a VT terminal on a TX serial line, so I was switching back andD forth with the ATTACH command available to both TPU and ALLIN1. ThisD particular set of processes was probably running for a week or more.  M Then, a batch job to do ALLIN1 house cleaning started, and that one tells anytJ process running ALLIn1 that it should exit ALLIN1. So the "A1" process notH only exited the OA$MAIN.EXE image, but also logged out since it has been= spanewd only to run ALLIN1. (this is the expected behaviour)..  H Normally, with just a JFMEZEI (in DCL) and A1 processes, one would see aL message such as "ALL-IN-1 terminating" folllowed by the "Control returned to, process JFMEZEI" as the A1 process vanished.  H However, this time, I was actually in the LIBR process putting finishing  touches to a set of subroutines.  M The interesting results were "interesting". Every second key would be sent to K JFMEZEI and the other to LIBR.  Not only did it corrupt files because I was K trying commands which would get half their characters typed in to the othergI file, and <CTRL-W> would alternately display the TPU screen of one or thehK other process. And there was no way to know if typing <CTRL-z> would end upi saving one file or the other.   L My interpretation, after the fact was that when the A1 image got the messageG to shutdown, it did so and since the process had nothing else to do, it G terminated and returned control to its parent (JFMEZEI).  But since the L current subprocess was LIBR, I guess VMS got confused and made both JFMEZEI J and LIBR foreground. As a result, both had outstanding $QIOs to read input: from terminal, hence, each process got half my keystrokes.    > The command procedure to start the A1 process from JFMEZEI is: $SET NOCONTROL=Y $A1 :== "ATTACH A1"  $SPAWN/PROC=A1 ALLIN1/REENTERn $SET CONTROL=Y $EXITu    K I assume this is just some odd event caused by weird alignment of the stars N and that no engineer will get a heart attack when he realises it is a bug deep) down in the process management kernel :-)     ! And yes, this is with VAX VMS 7.2    ------------------------------  % Date: Mon, 01 Dec 2003 09:57:49 +0100 " From: labadie <labadie_g@decus.fr> Subject: Re: TELENT in batch2 Message-ID: <bqf05e$253$1@news-reader2.wanadoo.fr>   John Brandon wrote:vD > I have a user that is asking me about using TELNET in a batch job.$ > My response is to use RSH instead. > < > Just out of curiousity, can TELNET be used in a batch job? >  > Why or why not?i >  >  > J*o*h*n B*r*a*n*d*o*nN > VMS Systems Administratori, > firstname.lastname.spam.me.not@dalsemi.com Helloe   Have a look at  =   [Compaq C] Telnet Client Accepting Non Interactive Commands   R http://h18000.www1.hp.com/support/asktima/appl_tools/009E60A3-96F8B0E0-1C0186.html   I read, among other things  	 OVERVIEW:M  @ This Telnet client works in both interactive and non interactiveB modes. It can execute from a BATCH file. In both modes, it accepts? commands to remote login and so on. Also, the default IP Telnetp@ port (23) is used unless otherwise specified. This has been used? to automatically reset to a known state DECserver ports. It can C be used for any TCP/IP application, which recognizes ASCII commands  on a given IP port.f         regardst   Gerard   ------------------------------  % Date: Sun, 30 Nov 2003 20:13:35 +0100h2 From: Wilm Boerhout <w.boerhoutOLD@PAINTplanet.nl>4 Subject: Re: Telnet session with fixed TNAnnn: name?* Message-ID: <bqdfo6$c5b$1@reader10.wxs.nl>   Peter Weaver wrote:l   > @ > While I'm sitting here on a sunny Sunday afternoon waiting for= > a 3rd party hardware supplier to bring in a disk to replacee= > the one that died 3:00 this morning because SOMEBODY DIDN'Tf; > REPLACE THE SPARE AFTER THE LAST DISK FAILED! I thought Ie? > would throw in my two cents. You did not mention the IP stacki  ! I know what you're feeling Peter.   > > you are running, but we had this problem using TCPWare until= > we put the PC names into DNS. Once the PC names were in DNSe= > the TT_ACCPORNAM showed the fully qualified PC name withoutR; > the Port. SYSMAN ALF had no problem dealing with it then.i@ > Maybe your IP stack will do the same thing if the PC is listed	 > in DNS?z >   B I run the Compaq/HP TCPIP Services 5.1 stack. I know I can switch H between TELNET reporting the IP address or the host name, but as far as G I can see the port is always included in TT_ACCPORNAM. Do you expect a wI difference between specifying the PC host names in DNS and entering them sG in the local "hosts" database (via TCPIP SET HOST xxxx /ADDR=n.n.n.x) ?s   -- t
 Wilm Boerhouti   w.boerhoutOLD@PAINTplanet.nl(    (remove OLD PAINT from reply address)   ------------------------------  % Date: Sun, 30 Nov 2003 16:23:59 -0500o& From: "Chris Moore" <just@my.twocents>4 Subject: Re: Telnet session with fixed TNAnnn: name?8 Message-ID: <Wctyb.2784$yd.228093@news20.bellglobal.com>  K Putting them in HOSTS. wasn't enough to get this working.  They would stillAK be identified uniquely by the system as "node.domain.com;nn" where 'nn' wassI the port # connected to (and couldn't be counted on to be constant)  OnceDF the source was resolved from DNS, the port # didn't interfere. I don'tH explain 'em, just report 'em.  I did try this successfully with UCX (4.2D ECO10 iirc) a few years back, but we avoid that product (UCX, TCP/IP; Services .... call it what you will) like the plague it is.o  L (btw Peter just needed to look in the right place for the replacement drive,D it was at SOMEBODY's desk --- perfect spares filing system......lol)    ? "Wilm Boerhout" <w.boerhoutOLD@PAINTplanet.nl> wrote in messagei$ news:bqdfo6$c5b$1@reader10.wxs.nl... > Peter Weaver wrote:V >E > >eB > > While I'm sitting here on a sunny Sunday afternoon waiting for? > > a 3rd party hardware supplier to bring in a disk to replaceO? > > the one that died 3:00 this morning because SOMEBODY DIDN'T = > > REPLACE THE SPARE AFTER THE LAST DISK FAILED! I thought I:A > > would throw in my two cents. You did not mention the IP stackd >o# > I know what you're feeling Peter.D >i@ > > you are running, but we had this problem using TCPWare until? > > we put the PC names into DNS. Once the PC names were in DNS ? > > the TT_ACCPORNAM showed the fully qualified PC name without-= > > the Port. SYSMAN ALF had no problem dealing with it then.0B > > Maybe your IP stack will do the same thing if the PC is listed > > in DNS?M > >  >eC > I run the Compaq/HP TCPIP Services 5.1 stack. I know I can switchoI > between TELNET reporting the IP address or the host name, but as far asrH > I can see the port is always included in TT_ACCPORNAM. Do you expect aJ > difference between specifying the PC host names in DNS and entering themI > in the local "hosts" database (via TCPIP SET HOST xxxx /ADDR=n.n.n.x) ?t >i > --   > Wilm BoerhoutA >  > w.boerhoutOLD@PAINTplanet.nl* >    (remove OLD PAINT from reply address) >e   ------------------------------  % Date: Sun, 30 Nov 2003 22:41:45 +0100-2 From: Wilm Boerhout <w.boerhoutOLD@PAINTplanet.nl>4 Subject: Re: Telnet session with fixed TNAnnn: name?* Message-ID: <bqdokp$fr2$1@reader08.wxs.nl>  G This is good news -apart from the "don't touch UCX" part, I don't make a the choices there-.1  E It looks like I can use the ALF facilty after all - will report back t sometime next week.r   Thanks   Chris Moore wrote:M > Putting them in HOSTS. wasn't enough to get this working.  They would stillrM > be identified uniquely by the system as "node.domain.com;nn" where 'nn' wasiK > the port # connected to (and couldn't be counted on to be constant)  Once H > the source was resolved from DNS, the port # didn't interfere. I don'tJ > explain 'em, just report 'em.  I did try this successfully with UCX (4.2F > ECO10 iirc) a few years back, but we avoid that product (UCX, TCP/IP= > Services .... call it what you will) like the plague it is.n   -- C
 Wilm Boerhoutt   w.boerhoutOLD@PAINTplanet.nl(    (remove OLD PAINT from reply address)   ------------------------------  % Date: Mon, 01 Dec 2003 18:37:03 +0100<" From: Didier Morandi <no@spam.com>F Subject: Re: The DCL minute of the day: sh kno links under DECnet Plus4 Message-ID: <3fcb7c6d$0$26787$636a55ce@news.free.fr>   Michael Unger wrote:   > J > Why didn't you use "$ esc[0, 8] = %X1B" to define the "escape character"3 > and insert that symbol later on in the procedure?v  @ because it was (stupidely) used within a type sys$input block... :-(n   D.   ------------------------------    Date: 01 Dec 2003 13:03:00 -05003 From: Rich Alderson <news@alderson.users.panix.com>l Subject: Re: VMS on PDP-10?n. Message-ID: <mddr7zo1kmj.fsf@panix5.panix.com>  $ Didier Morandi <no@spam.com> writes:  + > from http://www.cadigital.com/kildall.htmt   > ../..r  B >> Gary, and his students, wrote a small control program, which heF >> called CP/M (Control Program/Microcomputer). It enabled him to readE >> and write files to and from the disk. Gary copied the commands andnD >> file-naming conventions from the DEC PDP-10 VMS operating system.   > ???a > :-)o   Well, I was going to write:$  F     Ye gods and little fishes!  Where do these children get this shit,     anyway?t  H     CP/M predates VMS by a couple of years, IIRC.  In any case, most DECJ     command interfaces looked alike to the casual observer (Tops-10, OS/8,G     RT-11, RSX-15, etc. etc.), including VMS in its early incarnations.cE     Kildall based the command structure of CP/M on DOS-11, I believe.0  E     There was a hacker-wannabe book in the mid 80s, _Out of the Inner H     Circle_, which talked about not wanting to try hacking the Simtel-20D     archives because it was "some kind of VAX".  Looks like the kids0     haven't gotten any smarter in two decades...  O However, I decided to check out the page in question, and found that (1) it wasiO written by Sol Libes, hardly a kid, and (b) Didier quoted out of context in anyh case.a  M The mistake is still one of knowing that all DEC-originated CLI's look alike,tK and thus assigning the name "VMS" to Tops-10 in 1973, but I think that's anc old-age brain fart.e  = I still thought I knew that CP/M was based on DOS-11, though.d   -- e. Rich Alderson					    | /"\ ASCII ribbon     |< news@alderson.users.panix.com			    | \ / campaign against |L "You get what anybody gets. You get a lifetime."    |  x  HTML mail and    |7 			 --Death, of the Endless    | / \ postings         |(   ------------------------------  * Date: Mon, 1 Dec 2003 18:49:37 +0000 (UTC)- From: lewis@lumina.mitre.org (Keith A. Lewis)s Subject: RE: Wildcard searching . Message-ID: <bqg2g1$bai$1@newslocal.mitre.org>   "Tom Linden" <tom@kednos.com> writes in article <CIEJLCMNHNNDLLOOGNJIOELBIIAA.tom@kednos.com> dated Tue, 25 Nov 2003 18:23:57 -0800: >c >o >>-----Original Message-----* >>From: Tom Linden [mailto:tom@kednos.com]* >>Sent: Tuesday, November 25, 2003 6:06 PM >>To: Info-VAX@Mvb.Saic.Com/! >>Subject: RE: Wildcard searchingo >> >>D >>I have seen in the past implementations for C of the PL/I builtin,) >>Index, whose semantics are described in35 >>http://www.kednos.com/6291pro_037.html#index_x_1662  >>C >>Of course you could always right a one line PL/I subroutine whichn) >                            write        ) >>returns the position to your C program.   H C already has strstr(), which does more-or-less the same thing as INDEX.    4 >>>From: JF Mezei [mailto:jfmezei.spamnot@istop.com] >>> # >>>I have a buffer containing text.n >>>,K >>>I'd like to provide a function to do wildcard searches which returns ther >>>position inside that buffer.  >>> K >>>STR$MATCH_WILD has the widlcard support, but returns essentially true ormE >>>false, and I would have to add * to both the start and end of the d >>>pattern to match. >>>gF >>>(eg, if looking for choco*te , I would have to give STR$MATCH_WILD  >>> *choco*ta*F >>>in which case, it would return the first byte since the whole 2000  >>>line buffer >>>would match.e >>> 1 >>>Is there a magical routine that handles this ?X >>>; >>>(programming language is C).b  K The best way I can think of is to do your STR$MATCH_WILD "*choco*te*" checkIK first, then if it's true, STR$POSITION for "choco" and double-check each of,0 those against "choco*te*" until you get a match.  0 --Keith Lewis              klewis {at} mitre.org> The above may not (yet) represent the opinions of my employer.   ------------------------------  $ Date: Mon, 1 Dec 2003 18:30:47 +0100. From: "Jesper Naur" <jesper.naur@post.tele.dk>< Subject: X-Windows/Motif: Multihost multiwindows navigation.< Message-ID: <3fcb7aab$0$30088$edfadb0f@dtext01.news.tele.dk>   Hello everybody.  & A customer has the following scenario:  > Several separate VMS hosts H1, H2, ... Hn each running its ownH copy of a large Motif client application, consisting of several toplevel windows.  F All hosts H1...Hn have displays on a single X-server (which happens toE be one of them). The assignment is now to implement a function, whichnE allows the operator to select all windows originating from one single@F host Hx and make them visible, and at the same time making the windows1 from all the other hosts invisible (or iconized).e  A It is not an option to modify the large Motif client application.o  4 The systems run a variety of VMS and Motif versions:  ) Some run OpenVMS V7.3-1 with Motif V1.3-0y' Some run OpenVMS V7.3 with Motif V1.2-6 ' Some run OpenVMS V7.2 with Motif V1.2-5e  B Window management is done using 'mwm' and the 'old-style' desktop.  ? By inspecting the WM_CLIENT_MACHINE property of the windows, it H is possible to find out, where a particular window originates from. ThisK leaves the problem of implementing the functions 'make this window visible' ! and 'make this window invisible'.g  % I have explored the following avenue:   E A client window has as its immediate parent a decoration window owned6	 by 'mwm'.:  ' It is possible to iconize the client by   I 1. Sending (with XSendEvent) an MB1 ButtonPress event to the 'mwm' parent0 window7    (this causes the 'mwm' parent window to take focus).sK 2. Sending (with XSendEvent) an <ALT>F9 KeyPress event to the 'mwm' window.p  I This causes 'mwm' to make the client window and its own decoration window4L invisible (xwininfo reports IsUnmapped on the client window and IsUnviewableI on the decoration window). Also, a window containing an icon representinga$ the iconized client becomes visible.  ) It is possible to normalize the client by   I 1. Sending (with XSendEvent) an MB1 ButtonPress event to the icon window.*J 2. Sending (with XSendEvent) an <ALT>F5 KeyPress event to the icon window.    L A major problem with this approach is, that I have found no readily apparentI way of identifying which icon window belongs to which client - presumably  'mwm' I maintains an internal list of 'client window'-'icon window' associations,s but 7 this list seems to be unavailable to the outside world._  G The question is of course now: Does anyone have suggestions about how I"G can solve this problem? Either by pointing out a way of identifying thetE icon window belonging to a given client, or by suggesting an entirely @ different approach (don't be shy, anything will be appreciated).  
  Best regards3  Jesper Naur   ------------------------------   End of INFO-VAX 2003.665 ************************