1 INFO-VAX	Mon, 22 Aug 2005	Volume 2005 : Issue 468       Contents:P "...turning to India for answers is not a good thing..." - Interesting article.  Re: 7.3 supported ethernetD Re: AlphaServer 4100 CPU speed upgrade and/or backplane speed ranges@ Re: Cerner hopefully is using Zotob to show why hospitals shouldI Re: Cerner hopefully is using Zotob to show why hospitals should run VMS! I RE: Cerner hopefully is using Zotob to show why hospitals should run VMS! I Re: Cerner hopefully is using Zotob to show why hospitals should run VMS! I Re: Cerner hopefully is using Zotob to show why hospitals should run VMS! P Re: Cerner hopefully is using Zotob to show why hospitals should run VMS! VMS! VP Re: Cerner hopefully is using Zotob to show why hospitals should run VMS! VMS! VP Re: Cerner hopefully is using Zotob to show why hospitals should run VMS! VMS! VP Re: Cerner hopefully is using Zotob to show why hospitals should run VMS! VMS! V* Re: Chaining a .COM  or .EXE from an image* Re: Chaining a .COM  or .EXE from an image* Re: Chaining a .COM  or .EXE from an image* Re: Chaining a .COM  or .EXE from an image* Re: Chaining a .COM  or .EXE from an image* Re: Chaining a .COM  or .EXE from an image* Re: Chaining a .COM  or .EXE from an image* Re: Chaining a .COM  or .EXE from an image* Re: Chaining a .COM  or .EXE from an image* Re: Chaining a .COM  or .EXE from an image* Re: Chaining a .COM  or .EXE from an image* Re: Chaining a .COM  or .EXE from an image) Re: Chaining a .COM or .EXE from an image ) Re: Chaining a .COM or .EXE from an image  Re: EFI is out to lunch  Re: EFI is out to lunch   Re: Highwater marking experience& HP to dump itanium - bring back alpha?* Re: HP to dump itanium - bring back alpha?* Re: HP to dump itanium - bring back alpha?* Re: HP to dump itanium - bring back alpha?( Re: Next project, C programming problem.( Re: Next project, C programming problem.( Re: Next project, C programming problem.( Re: Next project, C programming problem.( RE: Next project, C programming problem.( Re: Next project, C programming problem.( Re: Next project, C programming problem.( Re: Next project, C programming problem.( Re: Next project, C programming problem.( Re: Next project, C programming problem.( Re: Next project, C programming problem.( Re: Next project, C programming problem.( Re: Next project, C programming problem.( Re: Next project, C programming problem.( Re: Next project, C programming problem.( Re: Next project, C programming problem.( Re: Next project, C programming problem. Re: RAM / Rumor Mill:  OpenVMS moving from Itanic to x86?  Re: SET FILE/ATTRIB LRL vs MRS5 Re: [TCPware, VMS V8.2] What is the official status ?   F ----------------------------------------------------------------------  % Date: Mon, 22 Aug 2005 10:30:14 -0500 + From: Chuck Aaron <caaron@ceris.purdue.edu> Y Subject: "...turning to India for answers is not a good thing..." - Interesting article.  4 Message-ID: <decr25$sq$1@mailhub227.itcs.purdue.edu>  , This is a multi-part message in MIME format.& --------------0104030701030102090906019 Content-Type: text/plain; charset=us-ascii; format=flowed  Content-Transfer-Encoding: 7bit   Y http://www.fortune.com/fortune/technology/articles/0,15114,1096414,00.html?promoid=yahoo      & --------------010403070103010209090601) Content-Type: text/html; charset=us-ascii  Content-Transfer-Encoding: 7bit   ? <!DOCTYPE html PUBLIC "-//W3C//DTD HTML 4.01 Transitional//EN">  <html> <head>I   <meta content="text/html;charset=ISO-8859-1" http-equiv="Content-Type">    <title></title>  </head> ' <body bgcolor="#339999" text="#ffffff">  <a class="moz-txt-link-freetext" href="http://www.fortune.com/fortune/technology/articles/0,15114,1096414,00.html?promoid=yahoo">http://www.fortune.com/fortune/technology/articles/0,15114,1096414,00.html?promoid=yahoo</a>  </body>  </html>   ( --------------010403070103010209090601--   ------------------------------  # Date: Mon, 22 Aug 2005 17:23:41 GMT / From: Adam Stouffer <adam_stouffer@hotmail.com> # Subject: Re: 7.3 supported ethernet * Message-ID: <xSnOe.5874$Fq2.1544@trndny03>   johnhreinhardt@yahoo.com wrote:   I > Support for Compaq's Fast Ethernet Network Interface Cards (NICs) based F > on Intel's i8255x Ethernet controllers has been added to the OpenVMSG > Alpha Version 7.2-1 release. These NICs replace the end-of-life DE500 @ > family of adapters. These PCI based NICs support 10/100BaseTX,< > 100BaseFX, Full Duplex, Half Duplex, and Auto Negotiation.  . Many thanks. I have some of those Intel cards.     Adam   ------------------------------    Date: 22 Aug 2005 11:14:10 -0500/ From: brooks@cuebid.zko.dec.nospam (Rob Brooks) M Subject: Re: AlphaServer 4100 CPU speed upgrade and/or backplane speed ranges - Message-ID: <sxZAZU6Amv9a@cuebid.zko.dec.com>   0 Duncan Brown <brown_du@eisner.decus.org> writes: > Norman Lastovica wrote:  > J > Hmm, good point.  I was so convinced I had done a hardware no-no that I H > didn't think about having tripped up on some latent VMS problem.  I'd H > still like to hear some voice (or PDF file...) of authority bless the G > basic concept of putting a 533MHz CPU board in a box that originally    > came with 400MHz ones, though.  G I recently upgraded a couple of 4x00's from 400Mhz CPUs to 600Mhz CPUs, I running both V7.3-2 and V8.2.  I only have two CPUs in each box, however.   * The extra 4MB of CPU cache really is nice!  M I've had no problems since the upgrade, but the heritage of some of the stuff H we've got in the lab can be a bit suspect; it's quite possible I'm usingD prototype hardware (that would be about 10 years old at this point).   --    M Rob Brooks    VMS Engineering -- I/O Exec Group     brooks!cuebid.zko.dec.com    ------------------------------    Date: 22 Aug 2005 09:16:05 -0500; From: koehler@eisner.nospam.encompasserve.org (Bob Koehler) I Subject: Re: Cerner hopefully is using Zotob to show why hospitals should 3 Message-ID: <WR6SmwjhvDPG@eisner.encompasserve.org>   Z In article <11gjl3im3n6j3d6@corp.supernews.com>, Dave Froble <davef@tsoft-inc.com> writes: > bob@instantwhip.com wrote:J >> you can open anything I give you access to ... that is called security! >>   > M > Open them with what tool?  Opening a Word document with EDT isn't much use.   5    Opening them with catdoc works well enough for me.   F    And I'm quite happy not to be able to open them with an application4    that will implement the application virus inside.   ------------------------------   Date: 22 Aug 2005 14:00:09 GMT( From: bill@cs.uofs.edu (Bill Gunshannon)R Subject: Re: Cerner hopefully is using Zotob to show why hospitals should run VMS!, Message-ID: <3mu438F18h7uaU1@individual.net>  0 In article <11gifqe14a4oha4@corp.supernews.com>,* 	Dave Froble <davef@tsoft-inc.com> writes: > Bill Gunshannon wrote:3 >> In article <11gfhh5ahbroqe4@corp.supernews.com>, - >> 	Dave Froble <davef@tsoft-inc.com> writes:  >>   >>>Bill Gunshannon wrote:  >>> G >>>>In article <1124565228.925988.212330@g43g2000cwa.googlegroups.com>,   >>>>	bob@instantwhip.com writes: >>>> >>>>I >>>>>"This is a good example of why you can't protect a weak OS with some K >>>>>router only because if something happens from tyhe inside, you are hit  >>>>>just as badly." >>>>> 7 >>>>>so why haven't all these windoze cios figured this ! >>>>>out yet and switched to vms?  >>>> >>>>  I >>>>Because the security provided by VMS carries with it the same penalty ; >>>>as a secured windows box.  No real world functionality.  >>>> >>>>bill >>>> >>>   >>>Now that I'll not agree with. >>   >>  3 >> Wether you agree or not does not change reality.  > F > If I'm right it defines reality.  You write as if reality is as you  > define it.  F No, I just think there are many people here (in c.o.v) who's vision ofA reality somehow has VMS in a much loftier place than it currently D occupies.  Don't get me wrong, I find my job moving more and more toC Windows, just like everyone else and like most of the people here I D don't like it either.  It's more work and a lot more headaches.  But4 the position MS holds in the IT industry is reality!   > K >>>If you define real world functionality being able to run MS Office, and  J >>>some do, then you have a point.  But MS Office isn't the only thing in  >>>the world worth doing.  >>   >>  F >> What is the current function of the largest percentage of computers >> in use today? > H > First, we could get into a lengthly and meaningless argument over the J > definition of 'a computer'.  That deviates from the topic, so let's not. > H > It could be argued that 'upgraded typewriters' are one of the largest A > segments of computing today.  I'll agree with that.  Ditto for  $ > calculators, telephones, and such.  E Well, the origin of this discussion was "Why not replace Windows with F VMS?" so I think we can eliminate typewriters, phones and calculators.   > K > Large segments of the whole are not everything.  Majority, yes.  All, no.  > 9 >>  Server apps?  Or Desktop apps?  Where are the largest H >> number of Windows boxes that get infected with the virus-du-jour?  In' >> the server room?  Or on the desktop?  > 5 > I'd expect the desktop, gameroom, gossip tool, etc.  > I >> So, where does VMS have to take over to have any impact on the current E >> situation?  And, what possibility is there of that ever happening?  > 2 > Whoops!  Hold on there.  Don't change the topic.  E Not changing the subject.  That was the original subject.  Brought up C by Boob. He, once again, suggested that VMS could replace all these 6 infected Windows boxes and reality is that it can not.   > G > The topic, and you named it, is whether a secured VMS system has any   > real world functionality.   F No, it wasn't. The original subject was VMS ability to replace all theF virus infected Windows boxes.  And my argument was that if you look atH what the majority of those machines are doing, VMS lacks that real world functionality.   > I > I design, write, install, and support real world applications that run  C > on VMS systems.  Unfortunately, I do the same for windoz systems.   G Sure, but what percentage of either class of machines are running those C particular applications?  Like it or not, the majority of the virus B infected machines are running things like Web Browsers and Office.G Can VMS replace them to the level that the user community would demand?    > F > Not one of my customers using VMS has ever had any type of security H > issue, not even a minor issue.  Why?  Because security is an integral E > part of the design and implementation and operations.  An employee  F > leaves?  A pre-determined set of procedures takes place.  Users are I > formally trained to avoid sharing security information.  User tracking  E > is built into the application, not tacked on afterward.  Users are  G > instructed that if a transaction shows up with their identification,  G > they are responsible for the transaction, regardless of who actually  
 > entered it.  > H > Ok, you may ask what role VMS plays in such a scenario.  The security K > design and implementation is really OS independent.  The reason for such  J > designed in security comes from a long background of being aware of the E > importance of security.  A VMS environment tends to encourage such  J > considerations.  Your typical windoz user and shop do not.  They could,  > but normally do not. >  >> Sorry, reality check. > D > Yes, a reality check.  Using even one customer, I can refute your I > statement about a secured VMS system.  From my experience, starting in  H > 1973, I can state that with VMS it's easier to implement security, in F > the application, and in the environment.  My experience with windoz 1 > tends to make me appreciate VMS more than ever.   E Once again, the point is missed.  I am not and have never argued that D VMS isn't secure.  I said it is irrelevant because the users can notD do what they want to do on a VSM system.  They could, but the ownersE of VMS choose not to make that possibility a reality.  At this point, D even if they could it is unlikely that VMS could regain that market.C Heck, Linux is free and it still hasn't driven MS from the desktop.    > G > I'm a VMS bigot and advocate.  I'm not a glutton for punishment.  My  D > e-mail, newsreader, and office stuff in done on a PC.  A decently C > protected network environment, prompted by a VMS background, and  1 > vindicated by today's insanity on the Internet.   F And I'm not any kind of a bigot.  I have used more OSes than I care toH remember all had their strengths and their weaknesses (Ok, some may onlyH have weaknesses) and just like in the language debate I think you shouldF use the OS that best fits the job.  VMS "could" be the right fit for aJ lot more than it currently is, but that is up to HP and they do not appearF to agree.  All the bigots int he world can't fix that.  Unless you you= consider them all monkeys and give them all keyboards........    bill      --  J Bill Gunshannon          |  de-moc-ra-cy (di mok' ra see) n.  Three wolvesD bill@cs.scranton.edu     |  and a sheep voting on what's for dinner. University of Scranton   |A Scranton, Pennsylvania   |         #include <std.disclaimer.h>       ------------------------------  % Date: Mon, 22 Aug 2005 11:39:17 -0400 ' From: "Main, Kerry" <Kerry.Main@hp.com> R Subject: RE: Cerner hopefully is using Zotob to show why hospitals should run VMS!R Message-ID: <FD827B33AB0D9C4E92EACEEFEE2BA2FB6B2078@tayexc19.americas.cpqcorp.net>   > -----Original Message-----$ > From: bill@triangle.cs.uofs.edu=20A > [mailto:bill@triangle.cs.uofs.edu] On Behalf Of Bill Gunshannon   > Sent: August 22, 2005 10:00 AM > To: Info-VAX@Mvb.Saic.Com = > Subject: Re: Cerner hopefully is using Zotob to show why=20  > hospitals should run VMS!  >=202 > In article <11gifqe14a4oha4@corp.supernews.com>,, > 	Dave Froble <davef@tsoft-inc.com> writes: > > Bill Gunshannon wrote:5 > >> In article <11gfhh5ahbroqe4@corp.supernews.com>, / > >> 	Dave Froble <davef@tsoft-inc.com> writes:  > >>=20  > >>>Bill Gunshannon wrote:  > >>>   D Bill - I may be mistaken, but I believe the original statement aboutC replacing Windows with OpenVMS was wrt to the server side - not the  desktop. Two different worlds.  G A great deal of the recent virus activity has been wrt to issues caused 6 on Windows 2000 Servers as well as W2K based desktops.   Regards   
 Kerry Main Senior Consultant  HP Services Canada Voice: 613-592-4660  Fax: 613-591-4477  kerryDOTmainAThpDOTcom (remove the DOT's and AT)=20  4 OpenVMS - the secure, multi-site OS that just works.   ------------------------------  + Date: Mon, 22 Aug 2005 15:56:01 +0000 (UTC)  From: david20@alpha2.mdx.ac.ukR Subject: Re: Cerner hopefully is using Zotob to show why hospitals should run VMS!) Message-ID: <decsih$4hc$1@news.mdx.ac.uk>   \ In article <4307714E.F15400D6@teksavvy.com>, JF Mezei <jfmezei.spamnot@teksavvy.com> writes: >Bill Gunshannon wrote: E >> Or like most modern viruses it just means the infected machine has C >> a .mil address in it's addressbook.  Most of the emailed viruses / >> are .exe, .com, .pif, etc. DOS Executables.   >  > E >This current virus isn't transmitted by email, it accesses an opened I >port on Windows machines directly, and then FTPs itself to that machine.  >   L Later variants of the Zotob worm added other secondary transmission methods./ W32/Zotob-c added in transimission by email see   7 http://www.sophos.com/virusinfo/analyses/w32zotobc.html     
 David Webb Security team leader CCSS Middlesex University   ------------------------------   Date: 22 Aug 2005 16:38:00 GMT( From: bill@cs.uofs.edu (Bill Gunshannon)R Subject: Re: Cerner hopefully is using Zotob to show why hospitals should run VMS!, Message-ID: <3mudb8F15ivdcU1@individual.net>  R In article <FD827B33AB0D9C4E92EACEEFEE2BA2FB6B2078@tayexc19.americas.cpqcorp.net>,* 	"Main, Kerry" <Kerry.Main@hp.com> writes: >  >> -----Original Message----- % >> From: bill@triangle.cs.uofs.edu=20 B >> [mailto:bill@triangle.cs.uofs.edu] On Behalf Of Bill Gunshannon! >> Sent: August 22, 2005 10:00 AM  >> To: Info-VAX@Mvb.Saic.Com> >> Subject: Re: Cerner hopefully is using Zotob to show why=20 >> hospitals should run VMS! >>=20 3 >> In article <11gifqe14a4oha4@corp.supernews.com>, - >> 	Dave Froble <davef@tsoft-inc.com> writes:  >> > Bill Gunshannon wrote: 6 >> >> In article <11gfhh5ahbroqe4@corp.supernews.com>,0 >> >> 	Dave Froble <davef@tsoft-inc.com> writes: >> >>=20 >> >>>Bill Gunshannon wrote: >> >>> > F > Bill - I may be mistaken, but I believe the original statement aboutE > replacing Windows with OpenVMS was wrt to the server side - not the   > desktop. Two different worlds.  F Go read Boob's rantings again.  He has always claimed that all WindowsG could be replaced by VMS.  This recent virus is from desktops boxes and B he once again asked why everybody isn't running VMS to prevent it.  H I am not saying he has a grip on reality or is even coherent most of theD time, but he does constantly claim that VMS could replace Microsoft.   > I > A great deal of the recent virus activity has been wrt to issues caused 8 > on Windows 2000 Servers as well as W2K based desktops.   E All I have seen is more of the usual.  Desktops running with services E they don't need leaving them wide open to infection because the users A don't even know what they are running.  I have never even seen an E attempted attack agains my 2K or 2K3 servers.  On the other hand, the C lab PC's seem to be prime targets (unsuccessfully as you can't even E access any of the random ports from outside our own network). And all E it takes is about 5 minutes before a new Unix server becomes a target B (also unsuccessful!)  I haven't bothered looking to see if the VAXC has been targeted since I brought it up but then, why would I care?    bill   --  J Bill Gunshannon          |  de-moc-ra-cy (di mok' ra see) n.  Three wolvesD bill@cs.scranton.edu     |  and a sheep voting on what's for dinner. University of Scranton   |A Scranton, Pennsylvania   |         #include <std.disclaimer.h>       ------------------------------    Date: 22 Aug 2005 06:03:52 -0700 From: bob@instantwhip.com Y Subject: Re: Cerner hopefully is using Zotob to show why hospitals should run VMS! VMS! V B Message-ID: <1124715832.819741.96410@g14g2000cwa.googlegroups.com>  > a vms mail server w/pmdf, precisemail, and sophos tco is cheapB when even though it costs a little more up front, after 7 years of< zero patches or problems, that makes those extra few dollars@ well worth it ... and I better add web access as Process will be releasing it soon for vms ...    ------------------------------    Date: 22 Aug 2005 06:08:44 -0700 From: bob@instantwhip.com Y Subject: Re: Cerner hopefully is using Zotob to show why hospitals should run VMS! VMS! V C Message-ID: <1124716124.898173.124890@g14g2000cwa.googlegroups.com>   G you can open anything I give you access to ... that is called security!    ------------------------------    Date: 22 Aug 2005 06:07:26 -0700 From: bob@instantwhip.com Y Subject: Re: Cerner hopefully is using Zotob to show why hospitals should run VMS! VMS! V B Message-ID: <1124716046.085472.56130@z14g2000cwz.googlegroups.com>  : I haven't had to patch either 7.1-1H2 or 7.1-2 VMS since I8 initially built it, and patches to TCPware have been few9 mainly to eliminate those nasty "access violation errors" ( that stop hacks dead on their tracks ...   ------------------------------  % Date: Mon, 22 Aug 2005 09:38:45 -0400 ' From: Dave Froble <davef@tsoft-inc.com> Y Subject: Re: Cerner hopefully is using Zotob to show why hospitals should run VMS! VMS! V 0 Message-ID: <11gjl3im3n6j3d6@corp.supernews.com>   bob@instantwhip.com wrote:I > you can open anything I give you access to ... that is called security!  >   K Open them with what tool?  Opening a Word document with EDT isn't much use.    --  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 Road  Vanderbilt, PA  15486    ------------------------------  % Date: Mon, 22 Aug 2005 05:01:01 -0700 ( From: Jeff Cameron <roktsci@comcast.net>3 Subject: Re: Chaining a .COM  or .EXE from an image 0 Message-ID: <BF2F0C8D.12E21%roktsci@comcast.net>  L On 8/21/05 10:26 PM, in article bmdOe.3466$WO2.357@fe06.lga, "Z" <Z@no.spam> wrote:   > JF Mezei wrote: I >> NETSERVER.EXE seems to be able to chain the execution of a .COM and/or   >> .EXE within the same process. >>  E >> How does it do it ? There is no code in netserver.com (which calls I >> netserver.exe) to look at some symbol set by netserver.exe and execute D >> it. So Netserver.EXE seems to have the ability to cause a command7 >> procedure to execute immediatly after it has exited.  >>   >> How is that done ?  > G > Are you sure NETSERVER.EXE has exited when the next command executes?  > G > If NETSERVER is exit_ING_, then the next .EXE/.COM could be called as # > part of NETSERVER's exit handler.   L  I don't Think so. Don't exit Handlers run in executive mode? There is a lot* in a .COM that wont Run in executive mode.   ------------------------------  % Date: Mon, 22 Aug 2005 05:09:12 -0700 ( From: Jeff Cameron <roktsci@comcast.net>3 Subject: Re: Chaining a .COM  or .EXE from an image 0 Message-ID: <BF2F0E78.12E23%roktsci@comcast.net>  J On 8/21/05 10:28 PM, in article 4309625F.C3ADB0B5@teksavvy.com, "JF Mezei"% <jfmezei.spamnot@teksavvy.com> wrote:    > Jeff Cameron wrote: M >> LIB$DO_COMMAND merely spawns a sub process and executes the command passed F >> to LIB$DO_COMMAND, all in the CONTEXT of the Executable that calledN >> LIB$DO_COMMAND. Once the sub-process completes, then the calling executable& >> continues on it's execution thread. >  > I > The documentation (HELP RTL LIB LIB$DO_COMMAND) says otherwise. It says G > that if LIB$DO_COMMAND succeeds, it is the last thing your image does  > (eg: your image exits).  > G > Looking at what netserver.exe does, if you show process/cont, you see J > that one process execute the commands which the network connection wantsG > to execute (for instance run FAL.EXE). So they do execute in the same 5 > process as NETSERVER.EXE, not a spawned subprocess.   I I stand Corrected. I was under the mistaken assumption LIB$DO_Command was G kind of a "Poor-Man's" $SPAWN system service, which I have always used.   K Damn, I love this OS! Been working with it for over 22 years, and I'm still  learning about it!   ------------------------------    Date: 22 Aug 2005 07:13:26 -0500- From: Kilgallen@SpamCop.net (Larry Kilgallen) 3 Subject: Re: Chaining a .COM  or .EXE from an image 3 Message-ID: <mVLkyF6THf6P@eisner.encompasserve.org>   [ In article <BF2F0C8D.12E21%roktsci@comcast.net>, Jeff Cameron <roktsci@comcast.net> writes:   , > Don't exit Handlers run in executive mode?  6 Exit handlers can also run in Supervisor or User mode.  C http://h71000.www7.hp.com/doc/82FINAL/4527/4527pro_035.html#jun_182    ------------------------------  # Date: Mon, 22 Aug 2005 12:15:13 GMT " From:   VAXman-  @SendSpamHere.ORG3 Subject: Re: Chaining a .COM  or .EXE from an image 0 Message-ID: <00A48A7A.15F22978@SendSpamHere.ORG>  \ In article <430958DC.5E247103@teksavvy.com>, JF Mezei <jfmezei.spamnot@teksavvy.com> writes: >Martin Vorlaender wrote:  >> > How is that done ?  >>   >> LIB$DO_COMMAND perhaps? >  > 	 >Thanks.   >  >So if you have: >  >CHOCOLATE.COM:  >  >$LOOP:  >$RUN PASTRY.EXE >$SHOW TIME  >$GOTO LOOP: > H >and PASTRY.EXE does a LIB$DO_COMMAND,  when that command completes, the >SHOW TIME will execute ?   K If the argument to LIB$DO_COMMAND is a descriptor which contains the string ! SHOW TIME then the answer is YES.  --  K 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, 22 Aug 2005 12:21:49 GMT " From:   VAXman-  @SendSpamHere.ORG3 Subject: Re: Chaining a .COM  or .EXE from an image 0 Message-ID: <00A48A7B.019D8A87@SendSpamHere.ORG>  [ In article <BF2EAE95.12DFA%roktsci@comcast.net>, Jeff Cameron <roktsci@comcast.net> writes: J >On 8/21/05 9:47 PM, in article 430958DC.5E247103@teksavvy.com, "JF Mezei"& ><jfmezei.spamnot@teksavvy.com> wrote: >  >> Martin Vorlaender wrote:  >>>> How is that done ?  >>>  >>> LIB$DO_COMMAND perhaps?  >>   >>   >> Thanks.   >>   >> So if you have: >>   >> CHOCOLATE.COM:  >>  	 >> $LOOP:  >> $RUN PASTRY.EXE
 >> $SHOW TIME  >> $GOTO LOOP: >>  J >> and PASTRY.EXE does a LIB$DO_COMMAND,  when that command completes, the >> SHOW TIME will execute ?  >  >What??? > L >In the procedure given above, assuming there was no error in PASTRY.EXE the? >SHOW TIME command will execute regardless of if PASTRY.EXE did  >LIB$DO_COMMAND or not.  > K >LIB$DO_COMMAND merely spawns a sub process and executes the command passed D >to LIB$DO_COMMAND, all in the CONTEXT of the Executable that calledL >LIB$DO_COMMAND. Once the sub-process completes, then the calling executable$ >continues on it's execution thread.  , That's the LIB$SPAWN RTL not LIB$DO_COMMAND.  I LIB$DO_COMMAND writes the command to a "buffer" for DCL and signals it to % execute the command in this "buffer".   J There are mechanics in DCL to "CHAIN" programs as well but there is no RTLJ to access this.  One must built build a dispatch block and inform DCL thatK a chain is desired.  Simple enough to do but it will require elevated/priv-  ileged access.    --  K 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: 22 Aug 2005 07:42:24 -0500 From: briggs@encompasserve.org3 Subject: Re: Chaining a .COM  or .EXE from an image 3 Message-ID: <4B1FdSHq015T@eisner.encompasserve.org>   [ In article <BF2F0C8D.12E21%roktsci@comcast.net>, Jeff Cameron <roktsci@comcast.net> writes: N > On 8/21/05 10:26 PM, in article bmdOe.3466$WO2.357@fe06.lga, "Z" <Z@no.spam> > wrote: >  >> JF Mezei wrote:J >>> NETSERVER.EXE seems to be able to chain the execution of a .COM and/or! >>> .EXE within the same process.  >>> F >>> How does it do it ? There is no code in netserver.com (which callsJ >>> netserver.exe) to look at some symbol set by netserver.exe and executeE >>> it. So Netserver.EXE seems to have the ability to cause a command 8 >>> procedure to execute immediatly after it has exited. >>>  >>> How is that done ? >>  H >> Are you sure NETSERVER.EXE has exited when the next command executes? >>  H >> If NETSERVER is exit_ING_, then the next .EXE/.COM could be called as$ >> part of NETSERVER's exit handler. > N >  I don't Think so. Don't exit Handlers run in executive mode? There is a lot, > in a .COM that wont Run in executive mode.  G No.  Exit handlers do not run in executive mode.  Think about it.  That D would be a pretty nasty security hole -- user-chosen code running in executive mode.   F You may be thinking of rundown handlers.  But I believe that user mode rundown is done in kernel mode.   C My understanding of LIB$DO_COMMAND is that it is a service provided B by the supervisor.  The current image requests the invocation of aA chained command and then runs down normally.  When the supervisor B regains control as a result of image rundown, it then services the* request and activates the chained command.  0 Yes.  Read the help -- "directs the supervisor".   	John Briggs   ------------------------------  % Date: Mon, 22 Aug 2005 13:52:02 +0100 0 From: Chris Sharman <chris.sharman@sorry.nospam>3 Subject: Re: Chaining a .COM  or .EXE from an image 4 Message-ID: <dechpj$8lm$1$8300dec7@news.demon.co.uk>    VAXman- @SendSpamHere.ORG wrote:L > There are mechanics in DCL to "CHAIN" programs as well but there is no RTLL > to access this.  One must built build a dispatch block and inform DCL thatM > a chain is desired.  Simple enough to do but it will require elevated/priv-  > ileged access.  H There's also lib$put_common & lib$get_common - no need for privilege to  chain.   Chris    ------------------------------  % Date: Mon, 22 Aug 2005 13:48:39 +0100 0 From: Chris Sharman <chris.sharman@sorry.nospam>3 Subject: Re: Chaining a .COM  or .EXE from an image 4 Message-ID: <dechj8$gpt$1$8302bc10@news.demon.co.uk>    VAXman- @SendSpamHere.ORG wrote:K > LIB$DO_COMMAND writes the command to a "buffer" for DCL and signals it to ' > execute the command in this "buffer".  > L > There are mechanics in DCL to "CHAIN" programs as well but there is no RTLL > to access this.  One must built build a dispatch block and inform DCL thatM > a chain is desired.  Simple enough to do but it will require elevated/priv-  > ileged access.   There's also lib$run_programI I'd guess the basic rtl has some kind of chain, too, since chain is part   of the basic language.2 It would all use the undocumented sys$cli routine.   Chris    ------------------------------  # Date: Mon, 22 Aug 2005 14:18:28 GMT " From:   VAXman-  @SendSpamHere.ORG3 Subject: Re: Chaining a .COM  or .EXE from an image 0 Message-ID: <00A48A8B.4D58C1AB@SendSpamHere.ORG>  g In article <dechj8$gpt$1$8302bc10@news.demon.co.uk>, Chris Sharman <chris.sharman@sorry.nospam> writes: ! >VAXman- @SendSpamHere.ORG wrote: L >> LIB$DO_COMMAND writes the command to a "buffer" for DCL and signals it to( >> execute the command in this "buffer". >>  M >> There are mechanics in DCL to "CHAIN" programs as well but there is no RTL M >> to access this.  One must built build a dispatch block and inform DCL that N >> a chain is desired.  Simple enough to do but it will require elevated/priv- >> ileged access.  >  >There's also lib$run_program   H This requires the image specification.  I don't believe this will accept
 block chains.   J >I'd guess the basic rtl has some kind of chain, too, since chain is part  >of the basic language.    Basic?    3 >It would all use the undocumented sys$cli routine.    --  K 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: 22 Aug 2005 09:19:51 -0500- From: Kilgallen@SpamCop.net (Larry Kilgallen) 3 Subject: Re: Chaining a .COM  or .EXE from an image 3 Message-ID: <Tf6fDH$llyht@eisner.encompasserve.org>   T In article <4B1FdSHq015T@eisner.encompasserve.org>, briggs@encompasserve.org writes:] > In article <BF2F0C8D.12E21%roktsci@comcast.net>, Jeff Cameron <roktsci@comcast.net> writes:   g >>  I don't Think so. Don't exit Handlers run in executive>> in a .COM that wont Run in executive mode.  > I > No.  Exit handlers do not run in executive mode.  Think about it.  That F > would be a pretty nasty security hole -- user-chosen code running in > executive mode.  > H > You may be thinking of rundown handlers.  But I believe that user mode! > rundown is done in kernel mode.   D I do not believe there is any difference between "Exit handlers" andF "Rundown handlers" for an executable image.  Each executes in the mode from which it was declared.   F There are separate entrypoints for protected images, at least for exec and kernel modes as I recall.    ------------------------------    Date: 22 Aug 2005 09:39:13 -0500; From: koehler@eisner.nospam.encompasserve.org (Bob Koehler) 3 Subject: Re: Chaining a .COM  or .EXE from an image 3 Message-ID: <E2TY8vXR5nou@eisner.encompasserve.org>   T In article <4B1FdSHq015T@eisner.encompasserve.org>, briggs@encompasserve.org writes: > I > No.  Exit handlers do not run in executive mode.  Think about it.  That F > would be a pretty nasty security hole -- user-chosen code running in > executive mode.   D    Yes, exit handlers do run in executive mode.  It's not a securityA    hole because only a privileged user can add an exit hanlder in     executive mode.   ------------------------------    Date: 22 Aug 2005 11:47:52 -0500 From: briggs@encompasserve.org3 Subject: Re: Chaining a .COM  or .EXE from an image 3 Message-ID: <CMrZmljRzRiP@eisner.encompasserve.org>   q In article <E2TY8vXR5nou@eisner.encompasserve.org>, koehler@eisner.nospam.encompasserve.org (Bob Koehler) writes: V > In article <4B1FdSHq015T@eisner.encompasserve.org>, briggs@encompasserve.org writes: >>  J >> No.  Exit handlers do not run in executive mode.  Think about it.  ThatG >> would be a pretty nasty security hole -- user-chosen code running in  >> executive mode. > F >    Yes, exit handlers do run in executive mode.  It's not a securityC >    hole because only a privileged user can add an exit hanlder in  >    executive mode.  D Executive mode exit handlers run in executive mode.  Sure.  But user$ mode exit handlers run in user mode.  F If you get to the point where executive mode exit handlers are kickingC in, you're headed for process termination and LIB$DO_COMMAND is not  going to prevent it.   	John Briggs   ------------------------------  + Date: Mon, 22 Aug 2005 12:13:50 +0000 (UTC) 7 From: moroney@world.std.spaamtrap.com (Michael Moroney) 2 Subject: Re: Chaining a .COM or .EXE from an image( Message-ID: <decfhu$dnp$1@pcls4.std.com>  * Jeff Cameron <roktsci@comcast.net> writes:  K >LIB$DO_COMMAND merely spawns a sub process and executes the command passed D >to LIB$DO_COMMAND, all in the CONTEXT of the Executable that calledL >LIB$DO_COMMAND. Once the sub-process completes, then the calling executable$ >continues on it's execution thread.  C No, LIB$DO_COMMAND runs down the current image and runs the command E specified.  It does not return to the caller.  It operates _as if_ it I inserts the specified command immediately after the one being executed in A the .com file and simply exits. (this is not really what happens)    ------------------------------  + Date: Mon, 22 Aug 2005 13:13:38 +0000 (UTC) 7 From: moroney@world.std.spaamtrap.com (Michael Moroney) 2 Subject: Re: Chaining a .COM or .EXE from an image( Message-ID: <decj22$k7o$1@pcls4.std.com>    briggs@encompasserve.org writes:  O >>  I don't Think so. Don't exit Handlers run in executive mode? There is a lot - >> in a .COM that wont Run in executive mode.   H >No.  Exit handlers do not run in executive mode.  Think about it.  ThatE >would be a pretty nasty security hole -- user-chosen code running in  >executive mode.  G >You may be thinking of rundown handlers.  But I believe that user mode   >rundown is done in kernel mode.  G Not 100% sure what you mean by either, but what VMS calls exit handlers F can run in user, supervisor or executive modes.  The exit handlers runD when an image would otherwise terminate (SYS$EXIT or a fatal error).J An example might be an editor, which puts the terminal into a funky state,F may declare a user mode exit handler to reset the terminal to a normal mode upon exit.    ------------------------------  + Date: Mon, 22 Aug 2005 10:32:57 +0000 (UTC) % From: Dan Foster <usenet@evilphb.org>   Subject: Re: EFI is out to lunch5 Message-ID: <slrndgjaft.pfq.usenet@zappy.catbert.org>   m In article <00A47D4E.709018DB@SendSpamHere.ORG>, VAXman-  @SendSpamHere.ORG <VAXman-@SendSpamHere.ORG> wrote:  > My EFI console is horked.   G Just out of curiosity, manage to get any further, successfully, without ! divine (HP service) intervention?    -Dan   ------------------------------  # Date: Mon, 22 Aug 2005 12:24:26 GMT " From:   VAXman-  @SendSpamHere.ORG  Subject: Re: EFI is out to lunch0 Message-ID: <00A48A7B.5F7207F7@SendSpamHere.ORG>  ] In article <slrndgjaft.pfq.usenet@zappy.catbert.org>, Dan Foster <usenet@evilphb.org> writes: n >In article <00A47D4E.709018DB@SendSpamHere.ORG>, VAXman-  @SendSpamHere.ORG <VAXman-@SendSpamHere.ORG> wrote: >> My EFI console is horked. > H >Just out of curiosity, manage to get any further, successfully, without" >divine (HP service) intervention?  J I've been busy working on a kernel (driver) hack port to Itanium.  I have,I fortunately, access to 2 rx2600s.  I'm still in loop mode with the EFI on  the one rx2600. :(  I I hope to get back to it this week.  I hate to have to pull my rack apart A again for this but it look like that's just what I will be doing.    --  K 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, 22 Aug 2005 10:01:02 -0400 $ From: "Hein" <hein.nomail@hp.nomail>) Subject: Re: Highwater marking experience , Message-ID: <4309da9e$1@usenet01.boi.hp.com>  ; "JF Mezei" <jfmezei.spamnot@teksavvy.com> wrote in message  & news:43077EAC.D45F2042@teksavvy.com...4 > VAX VMS 7.2 on a volume without highwater marking. > + > $COPY SYS$SYSTEM:NETSERVER.EXE []temp.exe " > $set file temp.exe/attr=(lrl=80)   The simple/real test is: $COPY/ALLO=NL: tmp.tmp $SET FILE /END tmp.tmp  D > However, DUMP/RECORD still stops at record 70 and doesn't spit out2 > record 71 which contains 30 bytes of valid data.  L You lied to the system about the file. Don't complain if you don't like the  result.   E > Since I removed highwater marking some time ago to test performance L > difference (and forgot to put it back on), I guess I got what I asked for.   Ayup.   J > I just never realised that high-water marking would be such an importantE > basic security measure, I had figured it was just some neat add on.   I It is critical, if there are security concerns with the data on the disk.      > Question:  > Between : 1 > SET VOLUME/HIGHWATER_MARKING /NOERASE_ON_DELETE  > and 2 > SET VOLUME /NOHIGHWATER_MARKING /ERASE_ON_DELETE= > Is there a functional difference in the level of security ?   K Should end up the same. If the disk stays under the current system manager   control.E With erase_on_delete, you can not change your mind. The data will go. C With highwater marking the data is there to be grabbed but will be  & overwritten just before it is grabbed.  @ Highwatermarking should turn out to be cheaper: more performing.J If the application writes data before reading that data (like a 99.99% of J the applications do), then the new business data will perform the cleanup L and the actuall writing of zeroes rarely happens (only for set file/end and 4 random access writes beyonf eof to sequential files)   hth, Hein.    ------------------------------    Date: 22 Aug 2005 06:30:21 -0700 From: bob@instantwhip.com / Subject: HP to dump itanium - bring back alpha? C Message-ID: <1124717421.487511.188260@z14g2000cwz.googlegroups.com>   ) http://www.theinquirer.net/?article=25573    ------------------------------    Date: 22 Aug 2005 08:38:19 -0700! From: susan_skonetski@hotmail.com 3 Subject: Re: HP to dump itanium - bring back alpha? C Message-ID: <1124725099.461469.184730@g43g2000cwa.googlegroups.com>   F The article on the Inquirer is pure rumor please ignore it, it is justD wasting everyone's time.  I promise if I hear anything I will let myB distribution lists know.  We are working as hard and as well as we" possibly can on Integrity servers.   Thanks,    Sue    ------------------------------    Date: 22 Aug 2005 09:11:41 -0700 From: bob@instantwhip.com 3 Subject: Re: HP to dump itanium - bring back alpha? B Message-ID: <1124727101.955899.54860@f14g2000cwb.googlegroups.com>  - where there is smoke there is fire ... we the * customers cannot ignore this because we do- not want to have to port back to alpha or x86  just after porting to itanium!   ------------------------------  % Date: Mon, 22 Aug 2005 13:25:54 -0400 - From: JF Mezei <jfmezei.spamnot@teksavvy.com> 3 Subject: Re: HP to dump itanium - bring back alpha? , Message-ID: <430A0A9A.5DC2F4FD@teksavvy.com>   bob@instantwhip.com wrote: > + > http://www.theinquirer.net/?article=25573   H I could have written that article based on what I *think* is happening. D This article doesn't bring any facts to the table to support the badH IA64 prognostic. It is just as speculative as my posts in this newsgroupE because it doesn't point to any real HP statement that strengtens the  view that IA64 has no future.   E I may agree with the writer's opinion, but not sure it deserves to be C printed as "fact" in the inquirer at this point in time. When facts $ emerge, then that will be important.  H So, when Hoff and/or FredK admits he has been working on the port of VMS. to the 64 bit 8086, then we'll know for sure !G (notice how Hoff isn't posting much anymore... he probably doesn't have E access to comp.os.vms from the basement of ZKO where he is working on & the secret port of VMS to the 8086 :-)   ------------------------------  % Date: Mon, 22 Aug 2005 07:55:26 -0400 - From: "John E. Malmberg" <wb8tyw@qsl.network> 1 Subject: Re: Next project, C programming problem. 1 Message-ID: <yIKdncOytcCtIJTeRVn-qw@adelphia.com>    Bill Gunshannon wrote:3 > In article <aKmdnUwyHfpLn5TeRVn-gA@adelphia.com>, 2 > 	"John E. Malmberg" <wb8tyw@qsl.network> writes: >  >>Bill Gunshannon wrote: >>E >>>OK, now that I am ready to turn the students  loose on the VAX for E >>>programming projects I have another problem to solve.  While SAMBA H >>>now makes it trivial to move files between the VAX and out fileserverE >>>the files will not be in a proper VMS text format and the compiler  >>>won't deal with them. >>K >>The compiler should accept files that are in the STREAM-LF format, which  7 >>is how SAMBA should be creating them on the VMS side.  >  > J > SAMBA does.  But all the files are missing all the CR's and the compiler > does not accept that.   H I compile Stream-LF files on the C compilers all the time.  That is how G they come off of a tar-ball, which is the only way that I can get some   source modules.   I The compiler should only care if the CR's are missing if the file format  
 is STREAM-CR.   I The handling of CR's and LF's is not in the compiler it is in RMS, which  G does the record handling based on what the attributes are for the file.    > B >>Microsoft Windows will be either creating the files in what VMS * >>considers STREAM-LF or STREAM-CR format. >>E >>If they are not showing up properly on the VMS side, then there is  E >>probably a SAMBA option that should be set in the smb.conf file to   >>enable this. >   < > The problem is my program which read the file character by? > character writting it to a new file adding the missing CR's.  = > The problem is that contrary to any documetation I can find < > the fopen() call seems to be creating Fixed Length 512byte: > per record files instead of StreamLF.  All I need is for< > someone to tell me how to force fopen() to behave properly  D SAMBA on VMS has stuff in it to convert the VFS format to Stream-LF 9 format when presenting text files from VMS to PC clients.   D Now I did not do any work on that section of the current version of G SAMBA, nor have I intensively worked on it, but what you are trying to  E do is probably one of the main reasons that people use SAMBA, and if  I there was not a way to make it work, then I would expect to see a lot of  * complaints about it on the SAMBA-VMS list.  E If you are new to SAMBA on VMS, due to the very non-existence of VMS  I specific documentation, the SAMBA-VMS mailing list is the fastest way to  @ get help.  Most of the people that apparently have the hands on G experience and the person that is maintaining the 2.2.8 release do not  5 seem to be monitoring comp.os.vms on a regular basis.   H For text files created on a SAMBA share, there are two factors that may  be involved.  C If a version of the  file already exists on VMS, SAMBA will try to  G maintain the VMS attributes.  For files that are non-stream LF, it may  G get the conversion wrong.  There also may be an setting in smb.conf to   override this.  C If the version of the file does not exist, it should be created in  J STREAM-LF format, but there may also be a setting in smb.conf to fix this.   > = >>The option should also be visible through the SWAT program.  >>J >>The folks on the SAMBA-VMS mailing list can probably give more detailed  >>help.  >   N > This is not a SAMBA problem it is a problem with C on the VAX, specifically: > ' > Compaq C V6.4-005 on OpenVMS VAX V7.3      PYTHON> cc/ver% Compaq C V6.4-005 on OpenVMS VAX V7.3  PYTHON> dir vms.*;  - Directory PROJECT2_ROOT:[PERL.PERL-5_8_4-RC1]   5 VMS.C;6                  475  22-MAR-2004 14:13:04.00 5 VMS.DIR;1                  2  27-APR-2004 08:18:31.57 5 VMS.LIS;4              11273  29-APR-2004 01:30:25.12 5 VMS.OBJ;8                262  29-APR-2004 06:41:48.33    Total of 4 files, 12012 blocks.  PYTHON> dir/full vms.c;   - Directory PROJECT2_ROOT:[PERL.PERL-5_8_4-RC1]   2 VMS.C;6                       File ID:  (5533,1,0)8 Size:          475/477        Owner:    [SYSOP,MALMBERG]" Created:   22-MAR-2004 14:13:04.00& Revised:   29-APR-2004 06:41:47.16 (3) Expires:   <None specified>  Backup:    <No backup recorded>  Effective: <None specified>  Recording: <None specified>  File organization:  Sequential Shelved state:      Online  Caching attribute:  WritethroughF File attributes:    Allocation: 477, Extend: 0, Global buffer count: 0%                      No version limit C Record format:      Stream_LF, maximum 0 bytes, longest 32767 bytes 4 Record attributes:  Carriage return carriage control RMS attributes:     None Journaling enabled: NoneA File protection:    System:RWED, Owner:RWED, Group:RWED, World:RE  Access Cntrl List:  None Client attributes:  None   PYTHON> dump vms.cE Dump of file PROJECT2_ROOT:[PERL.PERL-5_8_4-RC1]VMS.C on 22-AUG-2005   06:52:01.32                ;6: File ID (5533,1,0)   End of file block 475 / Allocated 477  3 Virtual block number 1 (00000001), 512 (0200) bytes   =   56202A20 0A2A200A 632E736D 76202A2F /* vms.c. *. * V 000000 =   74756F72 20636966 69636570 732D534D MS-specific rout 000010 =   200A356C 72657020 726F6620 73656E69 ines for perl5.  000020 =   302E372E 35203A6E 6F697372 6556202A * Version: 5.7.0 000030 =   30322074 73756775 41202A20 0A2A200A . *. * August 20 000040   I That specific version of the compiler seems quite happy with a file that   does not have any 'CR's in it.   -John  wb8tyw@qsl.network Personal Opinion Only    ------------------------------   Date: 22 Aug 2005 12:18:30 GMT( From: bill@cs.uofs.edu (Bill Gunshannon)1 Subject: Re: Next project, C programming problem. , Message-ID: <3mtu4mF18favaU1@individual.net>  , In article <43093012.30EAC3DE@teksavvy.com>,0 	JF Mezei <jfmezei.spamnot@teksavvy.com> writes: > Bill Gunshannon wrote:E >> The files are binary copied as far as I can tell.  They seem to be D >> StreamLF but missing the "0x0D" paired with the "0x0A" that marks4 >> the end of the record.  This upsets the compiler. >  > K >> SAMBA isn't creating anything.  SAMBA is just a file server.  It can not E >> (and should not) make any assumptions about the contents of files.  > $ > Actually it must. Similar to FTP,   J Actually, FTP does not.  FTP relies on the user telling it how to transferI files and if you tell it wrong it happily destroys the file.  It makes no D assumptions on it's own beyond having a default mode it comes up in.  F >                                   it should know what types of filesC > should be considered binary and which should be considered text.    D No, it should not and it can't possibly.  What criteria?  Extension?F How can it assume that the extension means anything? For example, from my VMS 7.3 Install CD:  7   DECW073.C;1   DECW073.D;1   DECW073.E;1   DECW073.F;1   H Guess what, the first one isn't C Source and the last one isn't ForTran.    H >                                                                  (sameH > with web servers BTW). If it knows that a file is a text file, it willA > then copy it to the target platform's native text file format.    H To the best of my knowledge web servers transfer all files as binary andI rely on the browser to interpret it.  This is done with associations that F are made by some human and not by trying to automagically identify the; filetype.  Even the Unix "file" command can't do that!  :-)    > Q > STREAMLF without RAT=CR should require only LF characters as record delimiters.   L StreamLF would be fine.  The offending file is being created as "rfm=fixed".   bill   --  J Bill Gunshannon          |  de-moc-ra-cy (di mok' ra see) n.  Three wolvesD bill@cs.scranton.edu     |  and a sheep voting on what's for dinner. University of Scranton   |A Scranton, Pennsylvania   |         #include <std.disclaimer.h>       ------------------------------   Date: 22 Aug 2005 12:38:53 GMT( From: bill@cs.uofs.edu (Bill Gunshannon)1 Subject: Re: Next project, C programming problem. , Message-ID: <3mtvatF18favaU3@individual.net>  0 In article <11giiq1a0vh2o85@corp.supernews.com>,* 	Dave Froble <davef@tsoft-inc.com> writes: > Bill Gunshannon wrote:4 >> In article <aKmdnUwyHfpLn5TeRVn-gA@adelphia.com>,3 >> 	"John E. Malmberg" <wb8tyw@qsl.network> writes:  >>   >>>Bill Gunshannon wrote:  >>> F >>>>OK, now that I am ready to turn the students  loose on the VAX forF >>>>programming projects I have another problem to solve.  While SAMBAI >>>>now makes it trivial to move files between the VAX and out fileserver F >>>>the files will not be in a proper VMS text format and the compiler >>>>won't deal with them.  >>> L >>>The compiler should accept files that are in the STREAM-LF format, which 8 >>>is how SAMBA should be creating them on the VMS side. >>   >>  K >> SAMBA does.  But all the files are missing all the CR's and the compiler  >> does not accept that. >>   >>  C >>>Microsoft Windows will be either creating the files in what VMS  + >>>considers STREAM-LF or STREAM-CR format.  >>> F >>>If they are not showing up properly on the VMS side, then there is F >>>probably a SAMBA option that should be set in the smb.conf file to  >>>enable this.  >>   >>  = >> The problem is my program which read the file character by @ >> character writting it to a new file adding the missing CR's. > >> The problem is that contrary to any documetation I can find= >> the fopen() call seems to be creating Fixed Length 512byte ; >> per record files instead of StreamLF.  All I need is for = >> someone to tell me how to force fopen() to behave properly  >>   >>  > >>>The option should also be visible through the SWAT program. >>> K >>>The folks on the SAMBA-VMS mailing list can probably give more detailed   >>>help. >>   >>  O >> This is not a SAMBA problem it is a problem with C on the VAX, specifically:  >>  ( >> Compaq C V6.4-005 on OpenVMS VAX V7.3 >>   >> bill  >>   > F > I've read through the entire thread.  I figured that this cannot be C > something new, and one or more good solutions would be suggested.  > I > Until you get things working better, if you consider this a C compiler  E > problem, then how about making the issue transparent to the users.    D Why does everyone assume the students will be using C?  If they wereF doing C they wouldn't need VMS and there would be no chance of gettingF them on it at all.  Right now, it is COBOL.  But I suspect the problemG may exist in all the compilers (except maybe C) because it isn't really 2 a compiler problem but a text file format problem.  I > Redirect the invocation of the C compiler to a procedure that will run  J > the source file through a pre-processing program similar to what you're = > writing, and then submit the source file to the C compiler.   D Yeah, but if I can't write the preprocessing program what is it thatH I would redirect it through?  And, I would then hace to write proceduresG for every compiler.  The likelihood that these would fail at some point K in their use (based on my inexperience with writting VMS commands as should I have been obvious from the problem I had making UNZIP work) would be very  high.    > J > If you cannot get a C program to work for you, try BASIC, it's great at  > handling strings.   M And I know even less about VMS BASIC.  The problem is in the default filetype 8 being created.  I don't know what it is in BASIC either!   bill      --  J Bill Gunshannon          |  de-moc-ra-cy (di mok' ra see) n.  Three wolvesD bill@cs.scranton.edu     |  and a sheep voting on what's for dinner. University of Scranton   |A Scranton, Pennsylvania   |         #include <std.disclaimer.h>       ------------------------------   Date: 22 Aug 2005 12:33:22 GMT( From: bill@cs.uofs.edu (Bill Gunshannon)1 Subject: Re: Next project, C programming problem. , Message-ID: <3mtv0iF18favaU2@individual.net>  , In article <43092E19.E708A84C@teksavvy.com>,0 	JF Mezei <jfmezei.spamnot@teksavvy.com> writes: > Bill Gunshannon wrote:K >> SAMBA does.  But all the files are missing all the CR's and the compiler  >> does not accept that. > K > Could you post the output of a DIR/FULL of one of the offending C files ?   L First, why assume the offending files are "C"?  I am actually more concernedH about COBOL.  I don;t know that the C Compiler would have a problem with files with justr "lf" in them.  ! > Could you post the output from:  > ! > DUMP/BLOCK=COUNT=1 filename.C ?   H I could, but it won't clarify any more than what I have already said.  IF can read a FileDump too.  It was the first thing I did to see what the problem was.     > > >> The problem is that contrary to any documetation I can find= >> the fopen() call seems to be creating Fixed Length 512byte ) >> per record files instead of StreamLF.   > G > fopen on VMS has extensions which allow it to create any file format.  > K > fopen("chocolate.C","w","rfm=fix","mrs=512") would do what you say above.  > 7 > fopen("chocolate.C","w")  would open a streamlf file.   M And that is what the documentation says.  But reality is, as usual, something  quite different.   >  > H > Also, if you EDIT/TPU the C file, do you see any control characters at" > end or at beginning of records ?  D I don't know if it's using TPU or EVE, but when I edit the files see, the offending "lf" at the end of the record.  E I found how to make my files StreamLF so now I have something else to 
 play with.   bill   --  J Bill Gunshannon          |  de-moc-ra-cy (di mok' ra see) n.  Three wolvesD bill@cs.scranton.edu     |  and a sheep voting on what's for dinner. University of Scranton   |A Scranton, Pennsylvania   |         #include <std.disclaimer.h>       ------------------------------  % Date: Mon, 22 Aug 2005 09:31:42 -0400 ' From: "Main, Kerry" <Kerry.Main@hp.com> 1 Subject: RE: Next project, C programming problem. R Message-ID: <FD827B33AB0D9C4E92EACEEFEE2BA2FB6B2060@tayexc19.americas.cpqcorp.net>   > -----Original Message-----$ > From: bill@triangle.cs.uofs.edu=20A > [mailto:bill@triangle.cs.uofs.edu] On Behalf Of Bill Gunshannon  > Sent: August 22, 2005 8:33 AM  > To: Info-VAX@Mvb.Saic.Com 3 > Subject: Re: Next project, C programming problem.  >=20. > In article <43092E19.E708A84C@teksavvy.com>,2 > 	JF Mezei <jfmezei.spamnot@teksavvy.com> writes: > > Bill Gunshannon wrote:? > >> SAMBA does.  But all the files are missing all the CR's=20  > and the compiler > >> does not accept that. > >=20   Bill,   D I can not believe that you are the first to ever encounter this textE format issue with Samba. This seems like a way to basic type issue to / not have some parameter setup to resolve it.=20   @ As previously suggested, have you tried posting to the VMS SambaF newsgroup? How did they setup their systems as they would have had the same issues ..   Regards   
 Kerry Main Senior Consultant  HP Services Canada Voice: 613-592-4660  Fax: 613-591-4477  kerryDOTmainAThpDOTcom (remove the DOT's and AT)=20  4 OpenVMS - the secure, multi-site OS that just works.   ------------------------------   Date: 22 Aug 2005 13:38:40 GMT( From: bill@cs.uofs.edu (Bill Gunshannon)1 Subject: Re: Next project, C programming problem. , Message-ID: <3mu2r0F18s5taU1@individual.net>  1 In article <yIKdncOytcCtIJTeRVn-qw@adelphia.com>, 0 	"John E. Malmberg" <wb8tyw@qsl.network> writes: > Bill Gunshannon wrote:4 >> In article <aKmdnUwyHfpLn5TeRVn-gA@adelphia.com>,3 >> 	"John E. Malmberg" <wb8tyw@qsl.network> writes:  >>   >>>Bill Gunshannon wrote:  >>> F >>>>OK, now that I am ready to turn the students  loose on the VAX forF >>>>programming projects I have another problem to solve.  While SAMBAI >>>>now makes it trivial to move files between the VAX and out fileserver F >>>>the files will not be in a proper VMS text format and the compiler >>>>won't deal with them.  >>> L >>>The compiler should accept files that are in the STREAM-LF format, which 8 >>>is how SAMBA should be creating them on the VMS side. >>   >>  K >> SAMBA does.  But all the files are missing all the CR's and the compiler  >> does not accept that. > J > I compile Stream-LF files on the C compilers all the time.  That is how I > they come off of a tar-ball, which is the only way that I can get some   > source modules.   A Once again, it's not the C Compiler I care about.  It's the COBOL 	 compiler.    > K > The compiler should only care if the CR's are missing if the file format   > is STREAM-CR.   E Yes, but the problem was that the default format wasn't coming out as D Stream anything.  I think I have fixed it.  In any case, it seems toE be working and I am not going to try to figure out why it didn't work = int he first place.  I have no time to troubleshoot the CRTL.    > K > The handling of CR's and LF's is not in the compiler it is in RMS, which  I > does the record handling based on what the attributes are for the file.   C True, I knew that.  That was why I was asking how to get fopen() to - create the files as Stream rather than Fixed.    >  >>  C >>>Microsoft Windows will be either creating the files in what VMS  + >>>considers STREAM-LF or STREAM-CR format.  >>> F >>>If they are not showing up properly on the VMS side, then there is F >>>probably a SAMBA option that should be set in the smb.conf file to  >>>enable this.  >>  = >> The problem is my program which read the file character by @ >> character writting it to a new file adding the missing CR's. > >> The problem is that contrary to any documetation I can find= >> the fopen() call seems to be creating Fixed Length 512byte ; >> per record files instead of StreamLF.  All I need is for = >> someone to tell me how to force fopen() to behave properly  > F > SAMBA on VMS has stuff in it to convert the VFS format to Stream-LF ; > format when presenting text files from VMS to PC clients.   B But while it may be able to tell what a windows file is (actually,B I doubt even that as DOS has always had two totally different text? file formats) there is no way it can know what a file on a Unix @ system is as all files are merely streams of bytes.  That is why? no program should ever make assumptions about the contents of a  file,    > F > Now I did not do any work on that section of the current version of I > SAMBA, nor have I intensively worked on it, but what you are trying to  G > do is probably one of the main reasons that people use SAMBA, and if  K > there was not a way to make it work, then I would expect to see a lot of  , > complaints about it on the SAMBA-VMS list.  D Somehow I doubt that there are many people using a SAMBA served UnixE file system, a VMS served file system and a Windows box in the middle  to actually move the files.    > G > If you are new to SAMBA on VMS, due to the very non-existence of VMS  K > specific documentation, the SAMBA-VMS mailing list is the fastest way to  
 > get help.     E I certainly have no time to read yet another mailing list.  I need to / stick to the ones that actually mean something.   A >           Most of the people that apparently have the hands on  I > experience and the person that is maintaining the 2.2.8 release do not  7 > seem to be monitoring comp.os.vms on a regular basis.  > J > For text files created on a SAMBA share, there are two factors that may  > be involved. > E > If a version of the  file already exists on VMS, SAMBA will try to  I > maintain the VMS attributes.  For files that are non-stream LF, it may  I > get the conversion wrong.  There also may be an setting in smb.conf to   > override this.  G I repeat, SAMBA should not be converting files.  There is no way it can H know what the data contained in a Unix file is so the only safe thing toI do is transfer it as binary keeping everything intact. If you can tell me H how anyone expects SAMBA to correctly identify the payload, I would love to hear it.    > E > If the version of the file does not exist, it should be created in  L > STREAM-LF format, but there may also be a setting in smb.conf to fix this.  B And it was, which makes one wonder why a unix style file with LF'sC doesn't work, unless it something about the COBOL Compiler.  In any D event, the real problem was getting the output file of my conversionB program to be a Stream file and I have figured out how to do that.' And the COBOL Compiler now seems happy.    >  >>  > >>>The option should also be visible through the SWAT program. >>> K >>>The folks on the SAMBA-VMS mailing list can probably give more detailed   >>>help. >>  O >> This is not a SAMBA problem it is a problem with C on the VAX, specifically:  >>  ( >> Compaq C V6.4-005 on OpenVMS VAX V7.3 >  >  > PYTHON> cc/ver' > Compaq C V6.4-005 on OpenVMS VAX V7.3  > PYTHON> dir vms.*; > / > Directory PROJECT2_ROOT:[PERL.PERL-5_8_4-RC1]  > 7 > VMS.C;6                  475  22-MAR-2004 14:13:04.00 7 > VMS.DIR;1                  2  27-APR-2004 08:18:31.57 7 > VMS.LIS;4              11273  29-APR-2004 01:30:25.12 7 > VMS.OBJ;8                262  29-APR-2004 06:41:48.33  > ! > Total of 4 files, 12012 blocks.  > PYTHON> dir/full vms.c;  > / > Directory PROJECT2_ROOT:[PERL.PERL-5_8_4-RC1]  > 4 > VMS.C;6                       File ID:  (5533,1,0): > Size:          475/477        Owner:    [SYSOP,MALMBERG]$ > Created:   22-MAR-2004 14:13:04.00( > Revised:   29-APR-2004 06:41:47.16 (3) > Expires:   <None specified> ! > Backup:    <No backup recorded>  > Effective: <None specified>  > Recording: <None specified>   > File organization:  Sequential > Shelved state:      Online" > Caching attribute:  WritethroughH > File attributes:    Allocation: 477, Extend: 0, Global buffer count: 0' >                      No version limit E > Record format:      Stream_LF, maximum 0 bytes, longest 32767 bytes 6 > Record attributes:  Carriage return carriage control > RMS attributes:     None > Journaling enabled: NoneC > File protection:    System:RWED, Owner:RWED, Group:RWED, World:RE  > Access Cntrl List:  None > Client attributes:  None >  > PYTHON> dump vms.cG > Dump of file PROJECT2_ROOT:[PERL.PERL-5_8_4-RC1]VMS.C on 22-AUG-2005  
 > 06:52:01.32  >               ;6< > File ID (5533,1,0)   End of file block 475 / Allocated 477 > 5 > Virtual block number 1 (00000001), 512 (0200) bytes  > ? >   56202A20 0A2A200A 632E736D 76202A2F /* vms.c. *. * V 000000 ? >   74756F72 20636966 69636570 732D534D MS-specific rout 000010 ? >   200A356C 72657020 726F6620 73656E69 ines for perl5.  000020 ? >   302E372E 35203A6E 6F697372 6556202A * Version: 5.7.0 000030 ? >   30322074 73756775 41202A20 0A2A200A . *. * August 20 000040  > K > That specific version of the compiler seems quite happy with a file that    > does not have any 'CR's in it.  M But, as I have said, it isn't the C compiler that the students will be using.    But, it's fixed!!    bill   --  J Bill Gunshannon          |  de-moc-ra-cy (di mok' ra see) n.  Three wolvesD bill@cs.scranton.edu     |  and a sheep voting on what's for dinner. University of Scranton   |A Scranton, Pennsylvania   |         #include <std.disclaimer.h>       ------------------------------   Date: 22 Aug 2005 13:39:47 GMT( From: bill@cs.uofs.edu (Bill Gunshannon)1 Subject: Re: Next project, C programming problem. , Message-ID: <3mu2t3F18s5taU2@individual.net>  * In article <FubOe.8563$1g2.5396@fe05.lga>, 	Z <Z@no.spam> writes: > Bill Gunshannon wrote:E >> OK, now that I am ready to turn the students  loose on the VAX for E >> programming projects I have another problem to solve.  While SAMBA H >> now makes it trivial to move files between the VAX and out fileserverE >> the files will not be in a proper VMS text format and the compiler D >> won't deal with them.  No big thing.  I merely wrote 2 utilities,G >> "unix2vms" and "vms2unix".  Simple on the unix side, but the idea is I >> to have the students work on the VMS side.  If they  have to be logged G >> into the Unix server to convert every file before moving it, nothing F >> is gained in trying to get them to work in the VMS environment. So,: >> I need these utilities to work on the VMS side as well. >>  D >> Now the problem.  I convert the file by writing it to a temp fileG >> while playing with the "lf"/"lf cr" combinations and then rename the K >> tempfile to the original files name.  This (according to the description J >> of "rename") guarantees that the file can not be lost even in the eventH >> of a system crash.  But the problem is the tempfile is not created asH >> StreamLF.  According to the CC docs I have found on the web it shouldG >> have been as that is the default for "fopen(filename, "w")".  Anyone G >> here tell me what I need to do to make sure the new file I create is G >> StreamLF so it doesn't end out with garbage padding at the end which 8 >> upsets the compiler even more  thanthe lack of "cr"s. > K > Perhaps one of these optional keywords will allow you force the creation  : > of the right type of temporary file on the fopen() call: > G > http://h71000.www7.hp.com/commercial/c/docs/5763p021.html#index_x_600  >  > ... 
 > rfm = <val> # > 	fix 	Fixed-length record format.   > 	stm	RMS stream record format.1 > 	stmlf	Stream format with line-feed terminator. 7 > 	stmcr	Stream format with carriage-return terminator. % > 	var	Variable-length record format. 1 > 	vfc	Variable-length record with fixed control.  > 	udf	Undefined.  > ...   ? Yeah, that was what I needed and eventually found.  The problem B seems to be fixed at this point. Just have to install my programs.   bill   --  J Bill Gunshannon          |  de-moc-ra-cy (di mok' ra see) n.  Three wolvesD bill@cs.scranton.edu     |  and a sheep voting on what's for dinner. University of Scranton   |A Scranton, Pennsylvania   |         #include <std.disclaimer.h>       ------------------------------  + Date: Mon, 22 Aug 2005 14:03:56 +0000 (UTC) ( From: m.kraemer@gsi.de (Michael Kraemer)1 Subject: Re: Next project, C programming problem. 5 Message-ID: <decm0c$991$1@lnx107.hrz.tu-darmstadt.de>   % sorry to interrupt your conversation, H but I encounter a similar problem with ftp (hobbyist VAX/VMS) transfers  of plain ASCII files. , Even with ASCII mode they arrive mangled. A     "set file/attr=rfm:stmlf"/ helps for a particular file. but how can I make / that the default for all ASCII file transfers ?    ------------------------------  % Date: Mon, 22 Aug 2005 16:18:08 +0200 3 From: "Gorazd Kikelj" <gorazd.kikelj@nospam.hp.com> 1 Subject: Re: Next project, C programming problem. , Message-ID: <4309dea1$1@usenet01.boi.hp.com>  6 "Bill Gunshannon" <bill@cs.uofs.edu> wrote in message & news:3mtvatF18favaU3@individual.net...   > F > Why does everyone assume the students will be using C?  If they wereH > doing C they wouldn't need VMS and there would be no chance of gettingH > them on it at all.  Right now, it is COBOL.  But I suspect the problemI > may exist in all the compilers (except maybe C) because it isn't really 4 > a compiler problem but a text file format problem.  L Did you try to set SYSGEN parameter RMS_HEURISTIC to 1? This will force RMS C to look into the contents of the file when determining a file type.   9 I have some success with this. Not perfect, but can work.   
 Best, Gorazd     ------------------------------   Date: 22 Aug 2005 14:44:32 GMT( From: bill@cs.uofs.edu (Bill Gunshannon)1 Subject: Re: Next project, C programming problem. , Message-ID: <3mu6mfF18qtovU1@individual.net>  3 In article <r71M57+LpTA3@eisner.encompasserve.org>, > 	koehler@eisner.nospam.encompasserve.org (Bob Koehler) writes:Y > In article <3msipiF18kfipU6@individual.net>, bill@cs.uofs.edu (Bill Gunshannon) writes: 4 >> In article <aKmdnUwyHfpLn5TeRVn-gA@adelphia.com>,3 >> 	"John E. Malmberg" <wb8tyw@qsl.network> writes:  >>  K >> SAMBA does.  But all the files are missing all the CR's and the compiler  >> does not accept that. >>   > G >    But there is no CR separator in a StreamLF file, so something else C >    is goin on here.  Can you post the compiler message, a dump of A >    a couple of blocks (not records) of the source file, and the  >    dir/full output?  >   B I appreciate all the help, but at this point it works and I really@ don't have the time (all the students will be back in full force@ next week and I still have a lot to do!) to figure out why COBOLB didn't like a StreamLF file with "lf"s as the record separators orA why the default for fopen() was fixed while all the documentation 8 (and at least one person here) said it should be Stream.  + Maybe in a few weeks if anyone still cares.    bill   --  J Bill Gunshannon          |  de-moc-ra-cy (di mok' ra see) n.  Three wolvesD bill@cs.scranton.edu     |  and a sheep voting on what's for dinner. University of Scranton   |A Scranton, Pennsylvania   |         #include <std.disclaimer.h>       ------------------------------    Date: 22 Aug 2005 09:43:00 -0500; From: koehler@eisner.nospam.encompasserve.org (Bob Koehler) 1 Subject: Re: Next project, C programming problem. 3 Message-ID: <eV$I1zvHAXCt@eisner.encompasserve.org>   W In article <3mtvatF18favaU3@individual.net>, bill@cs.uofs.edu (Bill Gunshannon) writes:   F > Why does everyone assume the students will be using C?  If they wereH > doing C they wouldn't need VMS and there would be no chance of gettingH > them on it at all.  Right now, it is COBOL.  But I suspect the problemI > may exist in all the compilers (except maybe C) because it isn't really 4 > a compiler problem but a text file format problem.  E    The only compiler on VMS that is source record format sentitive is F    Java, which somehow bypasses RMS.  And even it will compile code inF    other formats, it's just the debug and traceback utilities that get    confused.  G    For all other compilers,you just need to get the attributes to match G    the meta-data.  And CR is not meta-data in StreamLF, it is meta-data     in Stream or StreamCR.    ------------------------------    Date: 22 Aug 2005 09:34:36 -0500; From: koehler@eisner.nospam.encompasserve.org (Bob Koehler) 1 Subject: Re: Next project, C programming problem. 3 Message-ID: <u5wtsTkGuc1c@eisner.encompasserve.org>   W In article <3msi2tF18kfipU4@individual.net>, bill@cs.uofs.edu (Bill Gunshannon) writes: . > In article <4308B81E.1D6B8158@teksavvy.com>,2 > 	JF Mezei <jfmezei.spamnot@teksavvy.com> writes: >> Bill Gunshannon wrote: G >>>> programming projects I have another problem to solve.  While SAMBA I >>> now makes it trivial to move files between the VAX and out fileserver F >>> the files will not be in a proper VMS text format and the compiler >>> won't deal with them.  >>  F >> When you move a file from unix to VMS with SAMBA, how are the files. >> created on VMS ? what does a DIR/FULL say ? > D > The files are binary copied as far as I can tell.  They seem to beC > StreamLF but missing the "0x0D" paired with the "0x0A" that marks 3 > the end of the record.  This upsets the compiler.   *    OD with OA is not SteamLF, it's Stream.   ------------------------------    Date: 22 Aug 2005 09:36:59 -0500; From: koehler@eisner.nospam.encompasserve.org (Bob Koehler) 1 Subject: Re: Next project, C programming problem. 3 Message-ID: <r71M57+LpTA3@eisner.encompasserve.org>   W In article <3msipiF18kfipU6@individual.net>, bill@cs.uofs.edu (Bill Gunshannon) writes: 3 > In article <aKmdnUwyHfpLn5TeRVn-gA@adelphia.com>, 2 > 	"John E. Malmberg" <wb8tyw@qsl.network> writes: > J > SAMBA does.  But all the files are missing all the CR's and the compiler > does not accept that.  >   E    But there is no CR separator in a StreamLF file, so something else A    is goin on here.  Can you post the compiler message, a dump of ?    a couple of blocks (not records) of the source file, and the     dir/full output?    ------------------------------    Date: 22 Aug 2005 09:50:10 -0500; From: koehler@eisner.nospam.encompasserve.org (Bob Koehler) 1 Subject: Re: Next project, C programming problem. 3 Message-ID: <W1KqmGDMn59e@eisner.encompasserve.org>   W In article <3mu2r0F18s5taU1@individual.net>, bill@cs.uofs.edu (Bill Gunshannon) writes:   D > But while it may be able to tell what a windows file is (actually,D > I doubt even that as DOS has always had two totally different textA > file formats) there is no way it can know what a file on a Unix B > system is as all files are merely streams of bytes.  That is whyA > no program should ever make assumptions about the contents of a  > file,   B    At the OS level, both UNIX and Windows consider all files to beF    a stream of bytes.  At the text processing level they use different1    conventions for line separation (LF vs. CRLF).   I > I repeat, SAMBA should not be converting files.  There is no way it can J > know what the data contained in a Unix file is so the only safe thing toK > do is transfer it as binary keeping everything intact. If you can tell me J > how anyone expects SAMBA to correctly identify the payload, I would love
 > to hear it.   =    SAMBA, like some FTP clients, can probably be told to make F    assumptions about files based on file type.  That's not as reliableD    as having the user specify the information every time, but it's a    good time saver.   F    If SAMBA always transfers all data as unmodified bytes, then it canG    never deal with Windows vs. UNIX vs. VMS vs. MacOS ... text files.       PCNFS has that problem.  D > And it was, which makes one wonder why a unix style file with LF'sE > doesn't work, unless it something about the COBOL Compiler.  In any F > event, the real problem was getting the output file of my conversionD > program to be a Stream file and I have figured out how to do that.) > And the COBOL Compiler now seems happy.    ------------------------------   Date: 22 Aug 2005 15:04:55 GMT( From: bill@cs.uofs.edu (Bill Gunshannon)1 Subject: Re: Next project, C programming problem. , Message-ID: <3mu7snF18qtovU2@individual.net>  3 In article <W1KqmGDMn59e@eisner.encompasserve.org>, > 	koehler@eisner.nospam.encompasserve.org (Bob Koehler) writes:Y > In article <3mu2r0F18s5taU1@individual.net>, bill@cs.uofs.edu (Bill Gunshannon) writes:  > E >> But while it may be able to tell what a windows file is (actually, E >> I doubt even that as DOS has always had two totally different text B >> file formats) there is no way it can know what a file on a UnixC >> system is as all files are merely streams of bytes.  That is why B >> no program should ever make assumptions about the contents of a >> file, > D >    At the OS level, both UNIX and Windows consider all files to beH >    a stream of bytes.  At the text processing level they use different3 >    conventions for line separation (LF vs. CRLF).   G Actually, Windows has two different types of Text Files (at least under ) FAT, maybe with NTFS they only have one)     > J >> I repeat, SAMBA should not be converting files.  There is no way it canK >> know what the data contained in a Unix file is so the only safe thing to L >> do is transfer it as binary keeping everything intact. If you can tell meK >> how anyone expects SAMBA to correctly identify the payload, I would love  >> to hear it. > ? >    SAMBA, like some FTP clients, can probably be told to make 2 >    assumptions about files based on file type.    H In the absence of any form of Meta-data how can SAMBA assume the type of a file?   G >                                                That's not as reliable F >    as having the user specify the information every time, but it's a >    good time saver.   & And pretty much a recipe for disaster.   > H >    If SAMBA always transfers all data as unmodified bytes, then it canI >    never deal with Windows vs. UNIX vs. VMS vs. MacOS ... text files.    >    PCNFS has that problem.  D It probably shouldn't.  What precisely is the determining factor forG telling the difference between a binary data file, an executable binary ' file or a text file on a Unix system?      bill     --  J Bill Gunshannon          |  de-moc-ra-cy (di mok' ra see) n.  Three wolvesD bill@cs.scranton.edu     |  and a sheep voting on what's for dinner. University of Scranton   |A Scranton, Pennsylvania   |         #include <std.disclaimer.h>       ------------------------------  % Date: Mon, 22 Aug 2005 12:54:43 -0400 - From: JF Mezei <jfmezei.spamnot@teksavvy.com> 1 Subject: Re: Next project, C programming problem. , Message-ID: <430A034E.F77875E2@teksavvy.com>   Bill Gunshannon wrote:9 > Why does everyone assume the students will be using C?    * Perhaps the subject of this thread ???????    H If it is COBOL, then you really do need properly delimited records sinceA the COBOL language really treats "records" in the source code, as H opposed to the C compiler which treats a file as a stream of bytes (withF exception of the // comment tag which ignores everything until the end of current line).   G COBOL also may not like control characters in the source. (although the H VAX Cobol was more flexible than COBOLs on some other platfor in that it4 accepted tab character, lowercase or uppercase etc).  C If your files are coming from the foreign host as "rfm=fix" but the E contents are really unix files, then it would explain why Cobol would  complain a lot.   G If you are unable to get SAMBA to create the files properly on VMS, one B option would be to define the symbol "COBOL" to point to a commandG procedure which would then fix the files and invoke the cobol compiler. H The trick would be to skip the conversion if the source file has alreadyM been converted. (F$FILE_ATTRIBUTES can be used to pickup the RFM of the file)    ------------------------------   Date: 22 Aug 2005 17:17:19 GMT( From: bill@cs.uofs.edu (Bill Gunshannon)1 Subject: Re: Next project, C programming problem. , Message-ID: <3mufkvF18jjjqU1@individual.net>  , In article <430A034E.F77875E2@teksavvy.com>,0 	JF Mezei <jfmezei.spamnot@teksavvy.com> writes: > Bill Gunshannon wrote:: >> Why does everyone assume the students will be using C?  > , > Perhaps the subject of this thread ???????  F The subject is correct.  I was writting a C Program to handle cleaningF up the incoming COBOL source files and something the documentation forF C says was wrong. That is, all the docs say that if you oen a file forG writting ("w") with fopen() it will be Stream.  This was not what I was F getting.  I was opening a file for writting with fopen() and getting a! Fixed file with 512 byte records.    >  > J > If it is COBOL, then you really do need properly delimited records sinceC > the COBOL language really treats "records" in the source code, as J > opposed to the C compiler which treats a file as a stream of bytes (withH > exception of the // comment tag which ignores everything until the end > of current line).   B The COBOL source files are the data my program needs to work with.   > I > COBOL also may not like control characters in the source. (although the J > VAX Cobol was more flexible than COBOLs on some other platfor in that it6 > accepted tab character, lowercase or uppercase etc). > E > If your files are coming from the foreign host as "rfm=fix" but the G > contents are really unix files, then it would explain why Cobol would  > complain a lot.   D No, they are coming over as Stream but for whatever reason the COBOLE Compiler didn't like recordds terminated with an LF.  I was trying to C write a program to get rid of the Unix LF's and put something there D that the COBOL Compiler would be happy with.  But the file had to be@ Stream regardless, otherwise the padding blew the COBOL compilerA away.  I have succeeded.  I now convert the COBOL source files to - StreamCR and the compiler is happy with that.    > I > If you are unable to get SAMBA to create the files properly on VMS, one D > option would be to define the symbol "COBOL" to point to a commandI > procedure which would then fix the files and invoke the cobol compiler. J > The trick would be to skip the conversion if the source file has alreadyO > been converted. (F$FILE_ATTRIBUTES can be used to pickup the RFM of the file)   E And people here still think VMS is easier than Unix.  I have had more F than a half-dozen different "solutions" to the original problem all ofG which are more complicated than mine and all of which have the tendency E to drop the problem in the users lap.  In Unix it took me less than 5 D minutes to write a utility that can change the end of record symbolsE to pretty much anything I might want.  I think I will need to come up C with an exit survey before the end of the semester and see what the F students actually think of working on VMS.  Anybody want to offer some& sample questions for this survey?  :-)   bill   --  J Bill Gunshannon          |  de-moc-ra-cy (di mok' ra see) n.  Three wolvesD bill@cs.scranton.edu     |  and a sheep voting on what's for dinner. University of Scranton   |A Scranton, Pennsylvania   |         #include <std.disclaimer.h>       ------------------------------  % Date: Mon, 22 Aug 2005 19:29:37 +0200 6 From: =?ISO-8859-1?Q?Arne_Vajh=F8j?= <arne@vajhoej.dk> Subject: Re: RAM< Message-ID: <430a0b78$0$85749$edfadb0f@dread16.news.tele.dk>   Reinhard Eigner wrote:2 > "Jim" <none@example.com> schrieb im Newsbeitrag A > news:ra-dnZ2dnZ3PhrrhnZ2dnTmpmN6dnZ2dRVn-0J2dnZ0@comcast.com... 1 >>"Reinhard Eigner" <antispam@garnix.de> wrote in ) >>news:de2p8o$4qo$01$1@news.t-online.com: B >>>The AS1000A needs the good old 72pin PS/2 SIMMs (of course with >>>Parity).    >>SIMMs, eh?  P > Yes! The memory modules which were in your 80486 or Pentium 1/Pro computer...   , Will 486/Pentium SIMM's work in an AS1000A ?   Arne   ------------------------------  % Date: Mon, 22 Aug 2005 14:04:10 -0000 1 From: wspencer@ap.dontspamme.org (Warren Spencer) 8 Subject: Rumor Mill:  OpenVMS moving from Itanic to x86?1 Message-ID: <96BA6B167wspenceraporg@216.168.3.30>   G A fertile source of OpenVMS *rumors* - The Inq' - speculates HP may be  : dumping Itanic:  http://www.theinquirer.net/?article=25573   ws   --   Warren Spencer Senior Software Engineer The Associated Press   ------------------------------  % Date: Mon, 22 Aug 2005 09:51:03 -0400 $ From: "Hein" <hein.nomail@hp.nomail>' Subject: Re: SET FILE/ATTRIB LRL vs MRS , Message-ID: <4309d847$1@usenet01.boi.hp.com>  + Yeah there is some confusion in this space. 8 The best you can do is to just make LRL=MRS=some-number.L I see little purpose in pondering what setting one different fron the other " for fixed length files might mean.  C LRL is the one that counts for fixed length files but might not be  * maintained for some variable length files.I MRS is the hard upper limit for variable length files and is ignored for  $ more mixed length file applications.J MRS may be 0 for variable length file in which case applications which do = dynamic buffer allocation have to hope the LRL is up to date. M ANAL/RMS seem to have a bug in this space as it would not re-create the same  $ bogus file. Don't do it if it hurts!   Verification is trivial:  
 $ cre tmp.tmp % aap noot mies teun vuur boom kees zus . $ set file/att=(rfm=fix,lrl=10,mrs=20) tmp.tmp $ dump/rec tmp.tmp : @ Record number 1 (00000001), 10 (000A) bytes, RFA(0001,0000,0000)  <               746F 6F6E2070 61610025 %.aap noot...... 000000  @ Record number 2 (00000002), 10 (000A) bytes, RFA(0001,0000,000A)  <               6E75 65742073 65696D20  mies teun...... 000000 :             ; "JF Mezei" <jfmezei.spamnot@teksavvy.com> wrote in message  & news:43078077.8A77724C@teksavvy.com... > Odd behaviour: >  > + > $COPY SYS$SYSTEM:NETSERVER.EXE []TEMP.EXE " > $SET FILE TEMP.EXE ATTR=(LRL=80) > 9 > DIR/FULL TEMP.EXE shows a fixed format 80 byte records.  >  > ANA/RMS/FDL TEMP.EXE shows > RECORD$ >        BLOCK_SPAN              yes% >        CARRIAGE_CONTROL        none & >        FORMAT                  fixed$ >        SIZE                    512 >  >  > G > If I set MRS to 80, then both DIR/FULL and ANA/RMS/FDL  show a record  > size of 80 bytes.  >  >  > QUESTION:  > I > What is the difference between LRL and MRS in the context of fixed size  > record files ? > J > I was under the impression that MRS was just a convenience attribute forI > variable record length files to indicate the size of the longest record  > (and is not always there). > H > But it seems that ANA/RMS/FDL looks at MRS and disregards LRL totally. >  > Is there an explanation ?    ------------------------------  % Date: Mon, 22 Aug 2005 17:42:18 +0100 : From: "Tom Garcia" <tgarcia@hivemind.REMOVE-SPAM-TRAP.org>> Subject: Re: [TCPware, VMS V8.2] What is the official status ?3 Message-ID: <ftidnQn1JKYdnZfeRVnygg@eclipse.net.uk>   D "Peter 'EPLAN' LANGSTOEGER" <peter@langstoeger.at> wrote in message & news:430664ad$1@news.langstoeger.at...F > Since TCPware V5.6-2 is now over 3 years old, what is the status of  > TCPware ?   5 Having considered the same question, I found hints at J http://www.process.com/techsupport/tcpware%5Fcompat%5Fvms%5Fversions.htm ,' but I could not find any release dates.   M I guess it depends on demand, of course. TCPware has been my stack of choice  D as being the most VMSy and technically interesting, and for the VAX J certainly it's unbeatable (I miss ipv6, tho can understand that not being K justifiable with the current abomination that is nat being so popular) --   K the enterprise customer is understandably more conservative than hobbyists  I like myself when it comes to upgrades and bells/whistles. The recent big  K SSH_V562P060 certainly suggests fair maintenance; here's hoping TCPware is    in full continued development...   --  & Tom | tgarcia@hivemind.no-spam-thx.org   ------------------------------   End of INFO-VAX 2005.468 ************************