1 INFO-VAX	Mon, 24 Mar 2003	Volume 2003 : Issue 163       Contents:2 Re: ACCVIO error:  in house COBOL program.  NEWBIE2 Re: ACCVIO error:  in house COBOL program.  NEWBIE2 Re: ACCVIO error:  in house COBOL program.  NEWBIE2 Re: ACCVIO error:  in house COBOL program.  NEWBIE2 Re: ACCVIO error:  in house COBOL program.  NEWBIE- Re: Brian Schenkenberger - Where are you????? - Re: Brian Schenkenberger - Where are you????? - Re: Brian Schenkenberger - Where are you?????  Convert Alpha NT to VMS  Re: Convert Alpha NT to VMS  Re: Convert Alpha NT to VMS  Re: Convert Alpha NT to VMS  Re: Convert Alpha NT to VMS  Re: DECC$UTIME multiply defined  Re: DECC$UTIME multiply defined  Re: DECC$UTIME multiply defined & Determining Serial Port parity via QIO@ Re: How to search for last login, but not on disusered accounts.% Re: Mark Gorham was Re: Rich Marcello $ Re: MicroVAX newbie looking for help5 Re: Msg to OpenVMS Engineering regarding PerfectCache & Re: Sun unix/linux new nasty bug cert!  F ----------------------------------------------------------------------    Date: 23 Mar 2003 14:06:06 -08002 From: rsheehan@ci.bellevue.wa.us (Richard Sheehan); Subject: Re: ACCVIO error:  in house COBOL program.  NEWBIE < Message-ID: <3608772e.0303231406.cfa902f@posting.google.com>  Z p_sture@elias.decus.ch (Paul Sture) wrote in message news:<L+j4FgbYg26s@elias.decus.ch>... > E > Are you sure that the record length seen by the program is the same D > as in the file? The commented out CRC-WORK-ORDER and FILLER fieldsI > point me there, although I would have expected an error along the lines C > of "record too large for user buffer" (I can't remember the exact D > wording) if your program was trying to read a record into a buffer > which is too small.    Thank you for your response.  A Yes, the record lengths match.  The program will actually process A records prior to getting to this point.  This code and the record D layout have not changed in years, and program has been working until just the past week.    ------------------------------    Date: 23 Mar 2003 15:53:35 -0800) From: ben_yates_iii@yahoo.com (Ben Yates) ; Subject: Re: ACCVIO error:  in house COBOL program.  NEWBIE = Message-ID: <154d1183.0303231553.7f01cd56@posting.google.com>   w rsheehan@ci.bellevue.wa.us (Richard Sheehan) wrote in message news:<3608772e.0303230200.1d13936f@posting.google.com>... ? > I'm new to OpenVMS, and even newer to debugging COBOL on this  > platform.  > E > I've been trying to find help through documentation and online.  So 3 > far, this is what I've been able to come up with.  > G > Compiled with debug/nooptimize.  The process will run successfully in + > DEBUG.  *?* Might this be due to quotas??  >  > D Quotas are always a good place to start. On my team, we had programsC that would show a snapshot of a process and its quotas, so we could 8 have it looping and id any quota that might be exceeded.B Access Violation was one of the usual messages we would see when a quota was exceeded.    ------------------------------  % Date: Sun, 23 Mar 2003 20:46:46 -0500 $ From: Hein <hein_cov@eps.zk.dec.com>; Subject: Re: ACCVIO error:  in house COBOL program.  NEWBIE . Message-ID: <3E7E6386.AD37D4E5@eps.zk.dec.com>   Richard Sheehan wrote:  I > I'm new to OpenVMS, and even newer to debugging COBOL on this platform.   4 Well, The process dump suggests an ACCVIO in COBRTL.J It is very unlikely that a regular cobol program can/will pass bad addressB to cobol over which it will stumble (not impossible, just unlikely0 specially if the source lien you give is valid).  @ Therefor one has to suspect the cobol rtl itself, and it becomes& critically important to know the exactE version of the COBRTL in use. You may also want to check with support  whether there is a better  version out there.  G Once you have the exact COBRTL version someone (in support) can map the  relative PC toH a module and an offset for further analyses. Of course it could prove to be in an exception? handler which would go back the the code read soon. No point in  speculating.  H Now if that source line is correct, the program is doing an indexed file read. I There is a very small chance that RMS has send the cobrtl astray after if  got confused with a bad file. E So I'd encourage an ANAL/RMS on the file in question, or better still  CONVERT and try again.    # > Condition signalled to take dump: 5 > %SYSTEM-F-VA_IN_USE, virtual address already in use   G I don't trust that. What VMS version? This may be an ANAL/PROC problem. A When you say there is no problem runing with debug, can you try a  compile/linked debugH program, but with run/nodebug. I seem to recall a problem with anal/proc requiring (wrongly so) an image linked debug.J The particular error mosty happens during image activation and you suggest the code hadH been executing. It can also happen with LIB$FIS calls and SYS$CRMPSC and
 the likes,? but you give no indication of using such memory management call  dynamically.  G I don't expect it to be a quota problem. The process is not using much.  Still, check pagfilquo?   
 Good luck, Hein.        >  > ) > Here is pertinent sections of the code:  >  > *  > *  >     SELECT UB-CRC-FILE >       ASSIGN TO UB-CRC-FILE  >       ORGANIZATION IS INDEXED   >       RECORD KEY IS UB-CRC-KEY >       ACCESS MODE IS DYNAMIC.  >  > *  > *  > *  >  > FD  UB-CRC-FILE. > 01  UB-CRC-RECORD. >   03 UB-CRC-KEY.. >      05  CRC-FIRST-NUMBER          PIC 9(8).. >      05  CRC-UTILITY               PIC X(2).& >      05  CRC-NO            PIC X(8).& >   03  FILLER               PIC X(6).6 > *  03  CRC-WORK-ORDER                      PIC 9(6).# >   03  CRC-STATUS           PIC X. . >   03  CRC-START-DATE               PIC 9(8).. >   03  CRC-END-DATE                 PIC 9(8).' > *  03  FILLER              PIC X(21). * >   03  CRC-UNITS            PIC S9(3)V99.2 >   03  CRC-CHARGE-PER-UNIT          PIC S9(6)V99.2 >   03  CRC-TOTAL-CHARGE             PIC S9(6)V99.3 >   03  CRC-NUM-OF-BILLS-TO-BE-ISSUED     PIC 9(3). . >   03  CRC-NUM-OF-BILLS-ISSUED      PIC 9(3)./ >   03  FILLER                       PIC X(37).  > & > This is a listing from the DEBUGGER: >  >   4243: P-2.. >   4244:         READ UB-CRC-FILE NEXT RECORD >   4245:           AT END  >   4246:             GO TO P-9. >  > TIA. >  > Richard S Sheehan  > Business Systems Analyst > City of Bellevue, WA   ------------------------------  % Date: Sun, 23 Mar 2003 21:12:47 -0400 0 From: JF Mezei <jfmezei.spamnot@vl.videotron.ca>; Subject: Re: ACCVIO error:  in house COBOL program.  NEWBIE / Message-ID: <3E7E5B72.1E6C47F0@vl.videotron.ca>   I > > Compiled with debug/nooptimize.  The process will run successfully in - > > DEBUG.  *?* Might this be due to quotas??     1 I'd suspect the "nooptimize" rather than quotas.    0 What if it is compiled with /NOPTIMIZE/NODEBUG ?  I Optimization per say may not cause your problem, but it may cause another M problem which the compiler was "resistant" to in no-optimize to surface. (for " instance, arguments on the stack).   ------------------------------    Date: 23 Mar 2003 19:43:43 -0800$ From: n8wxs@arrl.net (Jeff Campbell); Subject: Re: ACCVIO error:  in house COBOL program.  NEWBIE = Message-ID: <1a40a0b2.0303231943.471fe9c5@posting.google.com>   w rsheehan@ci.bellevue.wa.us (Richard Sheehan) wrote in message news:?3608772e.0303230200.1d13936f@posting.google.com?... ? ? I'm new to OpenVMS, and even newer to debugging COBOL on this  ? platform.  ? E ? I've been trying to find help through documentation and online.  So 3 ? far, this is what I've been able to come up with.  ? G ? Compiled with debug/nooptimize.  The process will run successfully in + ? DEBUG.  *?* Might this be due to quotas??  ? F ? I also tried just compiling (no debug) with /noopt and process still ? fails. ? 1 ? Here's the info I've been able to put together.  ?  ? = ? %SYSTEM-F-ACCVIO, access violation, reason mask=00, virtual < ? address=0000000000084000, PC=FFFFFFFF80566168, PS=0000001B1 ? %TRACE-F-TRACEBACK, symbolic stack dump follows ; ?   image    module    routine line rel PC           abs PC E ?                                 0 0000000000000000 FFFFFFFF80566168 E ?  DEC$COBRTL                     0 00000000000289B0 000000007C1E09B0 E ?  DEC$COBRTL                     0 00000000000263C4 000000007C1DE3C4 E ?  UB-884  UB_884  UB_884      4244 0000000000010060 0000000000040060 E ?  UB-884                         0 0000000000032410 0000000000042410 E ?                                 0 FFFFFFFF8FB2B3F4 FFFFFFFF8FB2B3F4    [snip]   ?   # ? Condition signalled to take dump:  ?   5 ? %SYSTEM-F-VA_IN_USE, virtual address already in use  ? $  ? ) ? Here is pertinent sections of the code:  ?  ? *  ? *  ?     SELECT UB-CRC-FILE ?       ASSIGN TO UB-CRC-FILE                    ^^^^^^^^^^^  Where is this defined?   ?       ORGANIZATION IS INDEXED   ?       RECORD KEY IS UB-CRC-KEY ?       ACCESS MODE IS DYNAMIC.  ?    [snip]  
 Jeff Campbell  n8wxs@arrl.net   ------------------------------  # Date: Sun, 23 Mar 2003 19:48:39 GMT 4 From: Tim Llewellyn <tim.llewellyn@blueyonder.co.uk>6 Subject: Re: Brian Schenkenberger - Where are you?????0 Message-ID: <3E7E0E28.DC022320@blueyonder.co.uk>  ! VAXman-, @SendSpamHere.ORG wrote:  >     > J > In the states or in the UK?  I have had nothing but positive experiencesI > in London.  Wish I could state the same of Philadelphia and, moreso, of  > NYC. >    J Presumably you were not driving and not subject to a congestion charge:-).   regards  --   tim.llewellyn@blueyonder.co.uk    H * PLEASE NOTE tim.llewellyn@cableinet.co.uk address is NO LONGER VALID *   ------------------------------  # Date: Sun, 23 Mar 2003 21:13:54 GMT " From:   VAXman-  @SendSpamHere.ORG6 Subject: Re: Brian Schenkenberger - Where are you?????0 Message-ID: <00A1D4DF.ABBBF75F@SendSpamHere.ORG>  g In article <3E7E0E28.DC022320@blueyonder.co.uk>, Tim Llewellyn <tim.llewellyn@blueyonder.co.uk> writes:  >  > " >VAXman-, @SendSpamHere.ORG wrote: >>   >  >>  K >> In the states or in the UK?  I have had nothing but positive experiences J >> in London.  Wish I could state the same of Philadelphia and, moreso, of >> NYC.  >>   > K >Presumably you were not driving and not subject to a congestion charge:-).   I I've heard about that.  No, I use the Underground (damn that Central Line J being out) and the shoe-leather express when in London.  The rail has alsoJ simplified traversing the Southern parts of England.  Save for a taxi fromJ Salisbury to Stonehenge for 35 and a 40 taxi from Taunton's rail stationI to the Butlins in Minehead, there was nothing we couldn't see for want of I public transportation to get us there.  Even a bus ride from Brighton out J to Rottingdean's chalk cliffs precluded any need for a taxi and it was far cheaper, I'm sure.   --O VAXman- OpenVMS APE certification number: AAA-0001     VAXman(at)TMESIS(dot)COM              5   "Well my son, life is like a beanstalk, isn't it?"     ------------------------------  % Date: Sun, 23 Mar 2003 19:14:21 -0400 0 From: JF Mezei <jfmezei.spamnot@vl.videotron.ca>6 Subject: Re: Brian Schenkenberger - Where are you?????/ Message-ID: <3E7E3FB9.C4052838@vl.videotron.ca>   ! VAXman-, @SendSpamHere.ORG wrote: K > public transportation to get us there.  Even a bus ride from Brighton out L > to Rottingdean's chalk cliffs precluded any need for a taxi and it was far > cheaper, I'm sure.    " Been there, done that, by bike :-)   ------------------------------    Date: 23 Mar 2003 16:50:26 -0800' From: nrt@nxrt.cncdsl.com (Norm Taylor)   Subject: Convert Alpha NT to VMS= Message-ID: <97e2760a.0303231650.7bc8332c@posting.google.com>   A I have an Alpha 500 workstation, purchased off E-Bay. Seems to be = functional, but has 2 new completely empty hard drives in it.   F When I boot it, the ROM looks for Windows NT to boot. I presume that'sA what it was originally configured for. I now have my VMS hobbyist F license, and a complete media set for VMS 7.3 borrowed from my work. ID just can't figure out how to get it to use the VMS CD, since all theA instructions that come with that give instructions to type at the C ">>>" prompt.  When I try to use the ROM-based utility for changing A the OS setup, it won't run because it says it can't find any hard  drives. So I'm stuck.   5 Does anyone have suggestions for how I might proceed?    thanks, 
   Norm Taylor    ------------------------------  % Date: Sun, 23 Mar 2003 21:09:42 -0500 3 From: "Homer J. Simpson" <hsimpson@burnsenergy.com> $ Subject: Re: Convert Alpha NT to VMS: Message-ID: <mGtfa.32690$uO.23954@fe06.atl2.webusenet.com>  L IIRC, for Unix or VMS, you can't use the Adaptec 2940UW that comes with a NT@ 500.  You need to change out the SCSI controller with a KZPBA-CA& (FR-KZPBA-CA), also known as KZPBA-CY.  4 "Norm Taylor" <nrt@nxrt.cncdsl.com> wrote in message7 news:97e2760a.0303231650.7bc8332c@posting.google.com... C > I have an Alpha 500 workstation, purchased off E-Bay. Seems to be ? > functional, but has 2 new completely empty hard drives in it.  > H > When I boot it, the ROM looks for Windows NT to boot. I presume that'sC > what it was originally configured for. I now have my VMS hobbyist H > license, and a complete media set for VMS 7.3 borrowed from my work. IF > just can't figure out how to get it to use the VMS CD, since all theC > instructions that come with that give instructions to type at the E > ">>>" prompt.  When I try to use the ROM-based utility for changing C > the OS setup, it won't run because it says it can't find any hard  > drives. So I'm stuck.  > 7 > Does anyone have suggestions for how I might proceed?  > 	 > thanks,  >   Norm Taylor    ------------------------------  + Date: Mon, 24 Mar 2003 02:11:36 +0000 (UTC) ) From: Dan Foster <dsf@globalcrossing.net> $ Subject: Re: Convert Alpha NT to VMS5 Message-ID: <slrnb7sqan.532.dsf@gaia.tf.roc.gblx.net>   f In article <97e2760a.0303231650.7bc8332c@posting.google.com>, Norm Taylor <nrt@nxrt.cncdsl.com> wrote:C > I have an Alpha 500 workstation, purchased off E-Bay. Seems to be ? > functional, but has 2 new completely empty hard drives in it.  > H > When I boot it, the ROM looks for Windows NT to boot. I presume that'sC > what it was originally configured for. I now have my VMS hobbyist H > license, and a complete media set for VMS 7.3 borrowed from my work. IF > just can't figure out how to get it to use the VMS CD, since all theC > instructions that come with that give instructions to type at the E > ">>>" prompt.  When I try to use the ROM-based utility for changing C > the OS setup, it won't run because it says it can't find any hard  > drives. So I'm stuck.  > 7 > Does anyone have suggestions for how I might proceed?   K I've never touched that machine and don't have an in-depth familiarity with I SRM or AlphaBIOS, so I can't suggest anything. However, you might want to 4 post the exact model -- is it a PWS500a or PWS500au?  L And what is the exact SRM firmware version? And the OpenVMS PALcode version?9 (I think 'show config' at the >>> prompt will show this.)   @ What do you get when you enter 'show devices' at the >>> prompt?  E Exactly what make/model are the hard drives, CD drive, *and* the SCSI  adapter?  G Knowing each and every of the above details may be important in solving 
 your problem.    -Dan   ------------------------------  + Date: Sun, 23 Mar 2003 22:17:14 -0600 (CST)  From: sms@antinode.org$ Subject: Re: Convert Alpha NT to VMS) Message-ID: <03032322171373@antinode.org>   3 From: "Homer J. Simpson" <hsimpson@burnsenergy.com> B > [...] You need to change out the SCSI controller with a KZPBA-CA( > (FR-KZPBA-CA), also known as KZPBA-CY.  E    I believe that the -CA is also known as a -CX.  The -CB/-CY is the 5 differential card, while the -CA/-CX is single-ended.   G    I have a pristine -CY, for which I have no use, so I speak from some 1 knowledge here.  (Cross my palm with silver, ...)   H ------------------------------------------------------------------------  4    Steven M. Schweda               (+1) 651-699-98183    382 South Warwick Street        sms@antinode.org     Saint Paul  MN  55105-2547    ------------------------------  % Date: Sun, 23 Mar 2003 23:27:04 -0500 3 From: "Homer J. Simpson" <hsimpson@burnsenergy.com> $ Subject: Re: Convert Alpha NT to VMS; Message-ID: <LEvfa.49404$gS4.24777@fe02.atl2.webusenet.com>   G Yes, you are right.  My mistake.  -CA = -CX.  Thanks for catching that.   G <sms@antinode.org> wrote in message news:03032322171373@antinode.org... 5 > From: "Homer J. Simpson" <hsimpson@burnsenergy.com> D > > [...] You need to change out the SCSI controller with a KZPBA-CA* > > (FR-KZPBA-CA), also known as KZPBA-CY. > G >    I believe that the -CA is also known as a -CX.  The -CB/-CY is the 7 > differential card, while the -CA/-CX is single-ended.  > I >    I have a pristine -CY, for which I have no use, so I speak from some 3 > knowledge here.  (Cross my palm with silver, ...)  > J > ------------------------------------------------------------------------ > 6 >    Steven M. Schweda               (+1) 651-699-98185 >    382 South Warwick Street        sms@antinode.org  >    Saint Paul  MN  55105-2547    ------------------------------  # Date: Sun, 23 Mar 2003 19:46:01 GMT - From: "John E. Malmberg" <wb8tyw@qsl.network> ( Subject: Re: DECC$UTIME multiply defined< Message-ID: <Z9ofa.2673$TW2.1279002@news1.news.adelphia.net>   sms@antinode.org wrote: > > From:	SMTP%"Theo.Jakobus@iaf.fhg.de" 13-DEC-2001 01:58:24.67 > J >    Well, it took only a little over a year, but now that I have gotten aI > VMS V7.3 hobbyist kit (VAX), I finally loaded it up and duplicated this 
 > problem. >   5 >>I started to build GNU WGET and got a LINK warning:  >>; >>LINK /TRACE/NOMAP/EXEC=[.ALPHA]WGET.EXE wget.opt /options 4 >>%LINK-W-MULDEF, symbol DECC$UTIME multiply definedD >>         in module DECC$SHR file SYS$COMMON:[SYSLIB]DECC$SHR.EXE;1S >>%MMS-F-ABORT, For target [.ALPHA]WGET.EXE, CLI returned abort status: %X10648268.  >>$  >>4 >>I'm using: Compaq C V6.4-008 on OpenVMS Alpha V7.3 >  > J >    After some experimentation, I decided that I liked the behavior of myJ > "utime()" more than that of the new VMS C RTL "utime()".  Mine sets bothH > the created and modified dates of a file, while the new C RTL one sets5 > only the modified date.  (Good UNIX compatibility.)   F This was just covered under a thread about the build for perl failing  for the same reason.    $ And I will quote what I posted then:   [quote] @ This is actually a cross-platform problem with many open-source G programs.  They provide replacement routines that have the same symbol  D names as either missing or routines that they determine do not work 
 correctly.  E This is bad practice as many C compilers will when optimizations are  G enabled (default HP/COMPAQ/DEC C) optimize inline common routines that  ? the compiler knows about.  This also occurs on other platforms.     C If the open-source product or the porter had named the replacement  D routine for foo to be replace_foo, or my_foo, and then used a macro 5 replacement as below, then there would be no problem.   
 #ifdef foo
 #undef foo #endif* #define foo(x, y, z) replace_foo(x, y, z).    D Maintainers or porters of open-source programs need to take note of F this, because as the C RTL shipped with OpenVMS improves in it's UNIX 3 compatability, more of these problems will show up.   @ And these problems have been reported on other platforms on the B SAMBA-TECHNICAL mailing list.  Usually the better the compiler or D optimization settings, the more likely you are to have this problem.  D  From a programmer's perspective, the names of the standard library I routines in UNIX should be treated as reserved words.  No public routine  6 should have the same symbol name as a library routine.  D Doing so can cause undesired behavior that other platforms will not I diagnose, because the compiler will normally use the inline optimization  D   instead of the routine that the programmer thought that they were 	 invoking.   H To make the compiler and linker behave the classic library search order G where a local routine covers up a system routine requires dumbing down  H both the optimizer and the linking procedure.  This of course will have A performance impacts.  And this problem is not limited to OpenVMS.   H It is much simpler to code the replacement routine with it's own public 1 symbol, and use a macro to invoke it when needed.   B A complete online listing of the official UNIX routines and their C functions can be found at http://www.opengroup.org/.  They request  I registration to access their material.  I have never received any e-mail   from them after registering. [end quote]     F If a c routine is not behaving according to the UNIX standard, please A first check to see if there is a feature logical or setting that  E controls this.  In some cases this is needed because there is either  E existing code that depends on the broken behavior, or there is not a  G direct 1 to 1 match between the UNIX behavior and the possible OpenVMS  H behaviors.  File modification times are one of the areas where the UNIX ' standard is not clear on what it means.   G OpenVMS updates the file modification time anytime that the protection  F is changed or a file is open for write/modify/append.  UNIX just says H that the modification time will be updated for a significant event.  So ? far I have not found a definition for a significant event, but  H apparently just opening a file for write access and closing it does not  count.  H I do not know why the utime() call does not update the creation time, I I would have to look that up in the C manual and "The Open Group" which is  5 the official standard on how a routine should behave.   F If it is not to the specification, and there is not a feature logical + controlling this, please file a bug report.   H Again, any code that is naming local routines the same as existing C or B future C library routine is going to have problems as the CRTL in  OpenVMS improves.      -John  wb8tyw@qsl.network Personal Opinion Only    ------------------------------  # Date: Sun, 23 Mar 2003 21:06:27 GMT ; From: "John Gemignani, Jr." <jon-nope@thiswontworkossc.net> ( Subject: Re: DECC$UTIME multiply defined< Message-ID: <nlpfa.4310$TW2.1304161@news1.news.adelphia.net>  G <sms@antinode.org> wrote in message news:03032221173155@antinode.org... > > From: SMTP%"Theo.Jakobus@iaf.fhg.de" 13-DEC-2001 01:58:24.67 > J >    Well, it took only a little over a year, but now that I have gotten aI > VMS V7.3 hobbyist kit (VAX), I finally loaded it up and duplicated this 
 > problem. > 7 > > I started to build GNU WGET and got a LINK warning:  > > = > > LINK /TRACE/NOMAP/EXEC=[.ALPHA]WGET.EXE wget.opt /options 6 > > %LINK-W-MULDEF, symbol DECC$UTIME multiply definedF > >          in module DECC$SHR file SYS$COMMON:[SYSLIB]DECC$SHR.EXE;1I > > %MMS-F-ABORT, For target [.ALPHA]WGET.EXE, CLI returned abort status:  %X10648268.  > > $  > > 6 > > I'm using: Compaq C V6.4-008 on OpenVMS Alpha V7.3 > J >    After some experimentation, I decided that I liked the behavior of myJ > "utime()" more than that of the new VMS C RTL "utime()".  Mine sets bothH > the created and modified dates of a file, while the new C RTL one sets5 > only the modified date.  (Good UNIX compatibility.)  > F >    Interestingly, a UNIX "ls -l" shows the modified date for a file,I > while a VMS "DIRECTORY /DATE" shows the created date.  Thus, when doing B > a recursive FTP fetch, a VMS format FTP server shows a file dateG > (created) which will not be set at the receiving end if the new VMS C H > RTL "utime()" is used.  When both the dates are set, the file dates at= > the receiving (VMS client) end look better, at least to me.  > ? >    So, if you get the latest (1.5.3h) Wget kit from my server H > (<http://www.antinode.org/dec/sw/wget.html>), only the DESCRIP.MMS hasJ > been changed to suppress the DECC$ prefix on "utime", and thus Wget willG > work the same on VMS V7.3 as it did on earlier VMS versions.  Earlier H > suggestions to use "#if __CRTL_VER < 70300000" to excise my substituteB > "utime()" code on VMS V7.3 (and up) will also eliminate the LINK? > warnings, but will also result in different file time stamps.  >  >    Everything's complicated. > J > ------------------------------------------------------------------------ > 6 >    Steven M. Schweda               (+1) 651-699-98185 >    382 South Warwick Street        sms@antinode.org  >    Saint Paul  MN  55105-2547   J At compile time the DECC compiler looks into the DECC$SHR.EXE and replaces all K names of routines in the library to have the decc$ prefix.  For example, in  the library the H procedure utimes is listed and replaced by decc$utimes.  Therefore, YOUR procedure's E name is being replaced by decc$utimes, causing a multiple definition.   D What John Malmberg is saying is correct, in how such names should be
 considered by  coders.   L I have had to do what you are trying to do many times within TCP/IP Services
 either due@ to the need to emulate or replace the feature due to a different environment.  The DECCL compiler provides a workaround via the /[NO]PREFIX qualifier.  John suggests	 replacing I the definition with the preprocessor.  If you have a special routine just  for VMS, you could:    #ifdef __VMS #define UTIMES my$utimes
 UTIMES(...) {      }  #else  /* __VMS */ #define UTIMES utimes  #endif /* __VMS */         .      .        i = UTIMES( ... );   ------------------------------  + Date: Sun, 23 Mar 2003 22:23:34 -0600 (CST)  From: sms@antinode.org( Subject: Re: DECC$UTIME multiply defined) Message-ID: <03032322233464@antinode.org>   H    Thanks to all, but I _really_ don't need any more explanations of theH origin of and ways around this problem, as I understood it all just fineB before.  I posted my recent findings merely as a general notice toD potential users of Wget on VMS V7.3 (and up) who might have used theF __CRTL_VER test to suppress the MULDEF warning, as the behavior of the/ V7.3 C RTL "utime()" differs from that of mine.   E    As I explained, I like mine better, so if you get a fresh (1.5.3h) E Wget kit from me, it should now compile and link cleanly on VMS V7.3, - _and_ work as it did on earlier VMS versions.   H ------------------------------------------------------------------------  4    Steven M. Schweda               (+1) 651-699-98183    382 South Warwick Street        sms@antinode.org     Saint Paul  MN  55105-2547    ------------------------------    Date: 23 Mar 2003 21:21:27 -0800/ From: stuie_norris@yahoo.com.au (Stuart Norris) / Subject: Determining Serial Port parity via QIO = Message-ID: <51262235.0303232121.414c5372@posting.google.com>    Hello Readers,  M I am trying to determine the partiy setting on a serial port.  I have looked  J at the I/O reference guide and coded the following.  When I get the partiyM returned the value is 00011000.  I how do I equate that to the port parity of L none in this case?  Also how do I determine the number of stop bits on these ports?   Thanks   Stuart   #include <stdio.h> #include <ints.h>  #include <descrip.h> #include <iodef.h> #include <ttdef.h> #include <ssdef.h> #include <starlet.h> #include <lib$routines.h>   (    struct sensemode_iosb    //ZK-0698-GE {     uint16 status;     uint8  transmit_speed;     uint8  receive_speed;    uint8  crfill_count;     uint8  lffill_count;     uint8  parity_flags;     uint8  unused;  };   int main(int argc, char **argv)  {     int status = SS$_NORMAL;     unsigned short int channel;    long chrbuf[3];    long parity;     unsigned char x;     long baud_rate;      struct sensemode_iosb iosb;      $DESCRIPTOR(devnam,"tta0:");   .    status = sys$assign(&devnam, &channel,0,0);7    if ((status & SS$_NORMAL) != 1) lib$signal (status);   P    status = sys$qiow(0,channel,IO$_SENSEMODE,&iosb,0,0,&chrbuf,12,0,0,&parity,0) ; 7    if ((status & SS$_NORMAL) != 1) lib$signal (status); A    if ((iosb.status & SS$_NORMAL) != 1) lib$signal (iosb.status);       switch (iosb.transmit_speed)     {       case TT$C_BAUD_50:            baud_rate = 50;            break;        case TT$C_BAUD_75:            baud_rate = 75;            break;        case TT$C_BAUD_110:             baud_rate = 110;             break;        case TT$C_BAUD_134:             baud_rate = 134;             break;        case TT$C_BAUD_150:             baud_rate = 150;             break;        case TT$C_BAUD_300:             baud_rate = 300;             break;        case TT$C_BAUD_600:             baud_rate = 600;             break;        case TT$C_BAUD_1200:            baud_rate = 1200;            break;        case TT$C_BAUD_1800:            baud_rate = 1800;            break;        case TT$C_BAUD_2000:            baud_rate = 2000;            break;        case TT$C_BAUD_2400:            baud_rate = 2400;            break;        case TT$C_BAUD_3600:            baud_rate = 3600;            break;        case TT$C_BAUD_4800:            baud_rate = 4800;            break;        case TT$C_BAUD_7200:            baud_rate = 50;            break;        case TT$C_BAUD_9600:            baud_rate = 9600;            break;        case TT$C_BAUD_19200:             baud_rate = 19200;             break;        case TT$C_BAUD_38400:             baud_rate = 38400;             break;        case TT$C_BAUD_57600:             baud_rate = 57600;             break;        case TT$C_BAUD_76800:             baud_rate = 76800;             break;        case TT$C_BAUD_115200:            baud_rate = 11520;             break;        default:            baud_rate = 0;     }       .    printf("Baudrate is %d baud\n", baud_rate);         status = sys$dassgn(channel);7    if ((status & SS$_NORMAL) != 1) lib$signal (status);       return SS$_NORMAL;  }      show term tta0@ Terminal: _TTA0:      Device_Type: Unknown       Owner: No Owner  B    Input:    2400     LFfill:  0      Width:  80      Parity: None0    Output:   2400     CRfill:  0      Page:   24   ------------------------------    Date: 23 Mar 2003 21:17:00 -0800: From: craig.berry@SignalTreeSolutions.com (Craig A. Berry)I Subject: Re: How to search for last login, but not on disusered accounts. ; Message-ID: <7f15589f.0303232117.5d9944@posting.google.com>   Z Hein <hein_cov@eps.zk.dec.com> wrote in message news:<3E7C9AAD.D0F6671E@eps.zk.dec.com>...  P > Anyway... here is a much similar example to the above, this time done in PERL.  F And here's another example in Perl, which doesn't really save much but; illustrates the fact that the decision to use Perl does not C necessarily mean just parsing through the output of other commands.    use VMS::User;  @users = VMS::User::user_list();   for my $user (@users) { (   $uairef = VMS::User::user_info($user);2   next if $uairef->{FLAGS} & 0x10; # UAI$M_DISACNT/   print "$user last login (interactive)    : "  $     . $uairef->{LASTLOGIN_I} . "\n";/   print "$user last login (non-interactive): "  $     . $uairef->{LASTLOGIN_N} . "\n"; }   ? The VMS::User extension (like just about any Perl extension) is E available from http://search.cpan.org.  It's still officially in beta C but seems to work ok for simple lookups like this.  Eventually it's 5 supposed to be able to do $setuai as well as $getuai.   R > The 'date' manipulation is tricky again, this time because the ascii date in theL > listing can not directly be used for comparisons. I decided to use a quickQ > subroutine to convert to an encoded intege in the form <YYYY><MM><DD>. I'm sure @ > there are neat date package available for other manipulations.  C The package is VMS::Misc and includes routines for dealing with VMS D quadword dates and assorted conversions.  Using that you could add aC check for disusered in the last 30 days (or whatever) to my example  above.  ? Yet another Perl option would be to use VMS::IndexedFile, which D creates a tied hash to any indexed file, and of course SYSUAF.DAT isB one of those.  Conveniently, it's the one used in the example that' comes in the extension's documentation.    ------------------------------  % Date: Sun, 23 Mar 2003 14:08:49 -0400 0 From: JF Mezei <jfmezei.spamnot@vl.videotron.ca>. Subject: Re: Mark Gorham was Re: Rich Marcello/ Message-ID: <3E7DF830.25297DF9@vl.videotron.ca>    Sue Skonetski wrote:6 > Mark Gorham is the best VP I have ever worked with.  ... G > thing a VP can tell his people.  He also NEVER EVER says anything bad  > about anybody. n     Not even about me ????    ) :-) :-) :-) Talk about self restraint :-)r   ------------------------------  % Date: Sun, 23 Mar 2003 20:22:06 -0600a1 From: "David J. Dachtera" <djesys.nospam@fsi.net>a- Subject: Re: MicroVAX newbie looking for helpr' Message-ID: <3E7E6BCE.333A814F@fsi.net>-   Immo Birnbaum wrote: >  > Hi everybody,r > H > this afternoon my uncle gave me an old computer he said he didn't needF > anymore, and which was once the server in a company where he worked.G > Now I'm sitting in front of this thing and don't know what to do withEH > it. (I don't even know if this is the right newsgroup...). I hope thatD > someone around here can supply me with some informations about the > machine and how to use it. >  > It's labeled:u > digital MicroVAX 3100 . > on the front cover, and on the back it says: > Model DV31-ETA-AH > I just opened the cover to take a look inside, there's a harddisk withG > nothing but a digital label on it and a tape drive for tapes that are- > labeled "CompacTape".rH > He also gave me lots of tapes, including the system installation tapesB > (VMS 5.5), and lots of licenses. These date back to 1986 for theD > machine he used before the MicroVAX 3100, the licenses for VMS 5.5C > that have the 3100's serial number printed on them were issued inuE > 1992. The most recent backup tape is labeled "8/25/93", that's when / > the machine was replaced and he took it home.  > H > Can someone please tell me what the specs of this system are and, mostC > important, how I can get it up and running? As I also have an SGIrA > machine and a PC with NetBSD, I have some basic knowledge aboutn= > UNIX-like systems, but I have never worked with VMS before.nA > Also, I can't find any connectors for a keyboard or a monitor /gE > terminal. It has BNC/AUI ethernet connectors, a parallel connector,iE > several connectors that look similar to RJ45 which are probably fora? > connecting terminals, a SCSI connector, an unknown Centronics-H > connector labeled "<-- --> 4-11" and another unkown connector which isH > about as wide as a parallel port but is higher and has 3 rows of pins.( > How can I at least connect a terminal?  9 You'll want to inquire over on the comp.os.vms newsgroup.h  G A MicroVAX does not have a graphics card or naything like that. It usesoA a serial terminal for its "console". Likewise the "BIOS", or moreO: correctly, the console program, is CLI oriented instead of0 screen-oriented with a character-cell interface.  F In place of a serial terminal, you can use a terminal program on a PC.E You will need to read up on what we call "MMJ" connections ("ModifiedhD Modular Jack", so called because it is modified from the more commonG modular connectors in that the locking tab is offset to one side rathernG than centered). The pinouts can be found on the web via the OpenVMS FAQeA at http://www.openvms.compaq.com/wizard/openvms_faq.html The data C rate/format will likely be 9600-N-1 (9600 baud, no parity, one stop  bit).u  B "CompacTape" is the predecessor to the more common "DLT" ("DigitalE Linear Tape"). The drive model is probably TZ30, a SCSI-I device. TheDF cartridges, called TK50-K, have a native capacity of 95MB. These early( drives provided no hardware compression.  B VMS has very few similarities to UNIX beyond the basics of being aG multi-processing operating system. VMS can be thought of as a super-setaC of UNIX. However, discussing this in detail tens to open "operatingjF system wars" in the newsgroups. So, best to just read up on it and ask questions where appropriate.  G If the machine is in any condition to boot (possible, but don't take itoG for granted) here are some possible scenarios, assuming you can console 3 terminal connected to the lowest-numbered MMJ jack:v  F 1. The machine will power up and boot VMS. As you will likely not knowE the password for the SYSTEM accound or any other, this will not be ofvG much use to you. See the OpenVMS FAQ (link above) for information abouto@ how to get into a mahcine when you don't know the SYSTEM account password or any other.  G 2. The machine will power up and come to the console prompt which lookse
 like this:   >>>   G At that point, typing HELP should provide some entertainment. (BTW, youiD press <Return> after entering commands at the console. Sorry if thatF sounds condescending, but you'd be surprised how many folks don't know that.)  @ 3. The machine will fail to power up, and may or may not displayH anything on the console terminal. There are some diagnostic LEDs visibleF in a small window at the rear of the unit. Note the LEDs which are litD and whether they are on solid, flashing, etc.; then, ask about it in$ comp.os.vms or comp.sys.dec or both.  H A "MicroVAX" is a smaller, usually desktop-size version of the mid-rangeD VAX computer system line sold in the 1980s and 1990s by then DigitalC Equipment Corporation. Digital was bought by Compaq circa. 1998 andyD Compaq was just acquired by HP. A "VAXstation" is likewise usually aG desktop machine, but does include graphics (usually RGB, sync on green) - and does need a keyboard and usually a mouse.   F The VAX CPU is a CISC machine ("Complex Instruction Set Computing"), aD 32-bit processor. It was replaced in the Digital product line circa.C 1992 by the 64-bit RISC ("Reduced Instruction Set Computing") AlphaC
 processor.  D VMS ("Virtual Memory System") was written based on the VAX processorH architecture and uses many elements of earlier Digital operating systemsG such as RMS ("Record Management Services") which helps user programs do H record-level I/O by eliminating the need to provide such services in theF program itself, and ODS (On-Disk Structure) which is the "filesystem".= ODS is an early predecessor to ISO-9660. It provides for file < "versioning", in that a version number is tacked onto a fileH specification by the system so that you can have more than one "version"F of a file. For example, "MYFILE.DAT;1" and "MYFILE.DAT;2". The commandG language is called DCL ("Digital Command Language") which is extensibleRD and can be very consistent from one command to the next, unlike UN*XE where a command option such as "-r" might mean one thing to "dd", butEG something very different to "tar", for example. DCL has an API allowing D that command line processing can be done by system-provided services= rather than having to include such code in each user program.   H I could go on, of course, but that's the basics. Hope you enjoy learningG about VMS. It's been a great environment for me to work in for the last 	 20 years.E  . (Cross posted to comp.os.vms and comp.sys.dec)   -- i David J. Dachteraw dba DJE Systemsl http://www.djesys.com/  ( Unofficial Affordable OpenVMS Home Page: http://www.djesys.com/vms/soho/q   ------------------------------  # Date: Sun, 23 Mar 2003 22:42:38 GMTi/ From: brooks@cuebid.zko.dec.nospam (Rob Brooks) > Subject: Re: Msg to OpenVMS Engineering regarding PerfectCache- Message-ID: <uNPyqic5gP0R@cuebid.zko.dec.com>e  & svieth@wi.rr.com (Scott Vieth) writes:F > From the messages from Jim, it sounds like he is working hard on theG > problem but OpenVMS Engineering is not being as helpful or responsive  > as they could be.b  K It's my group that is in the process of making the needed changes, althoughr( I'm not directly involved with the work.  3 I'll forward this note along to the relevant folks.    -- V  M Rob Brooks    VMS Engineering -- I/O Exec Group     brooks!cuebid.zko.dec.comi   ------------------------------    Date: 23 Mar 2003 17:53:30 -0800( From: bob@instantwhip.com (Bob Ceculski)/ Subject: Re: Sun unix/linux new nasty bug cert!w= Message-ID: <d7791aa1.0303231753.48bb3046@posting.google.com>o  ^ david20@alpha2.mdx.ac.uk (David Webb) wrote in message news:<b5kmid$1pt$1@aquila.mdx.ac.uk>...j > In article <d7791aa1.0303211729.730c8be3@posting.google.com>, bob@instantwhip.com (Bob Ceculski) writes: > >Andrew Harrison SUNUK Consultancy <Andrew_No.Harrison_No@nospamn.sun.com> wrote in message news:<3E7B2646.70809@nospamn.sun.com>... > >> Bob Ceculski wrote:2 > >> > the inquirer has the details, and Andrew, I1 > >> > guess certs aren't important, esp. if they  > >> > point to your os ...  > >> > 0/ > >> > http://www.theinquirer.net/?article=8436  > >>   > >> Well to be more clear.0 > >> E/ > >> Solaris, HP-UX, Tru64, OpenVMS, AIX, Linux0% > >> IRIX are all probably vunerable.  > >>   > >> Regards > >> Andrew Harrison > >0. > >what's vms got to do with a unix/linux bug? >  > P > Probably referring to the fact that the DEC TCPIP services stack uses the sameP > codebase as the TRU64 TCPIP stack and the fact that the HP VMS Kerberos server > is based on MIT's code.oL > However the substantial differences between VMS and Unix doesn't mean thatL > these facts automatically make VMS vulnerable even if TRU64 is vulnerable. >  > David Webb > VMS and Unix team leader  ; and TCPware is the only vms ip stack that is still based onb7 the vms kernel and not unix based ... also vms security : would most likely squelch any unix bug, not always, but on% many instances it has in the past ...:   ------------------------------   End of INFO-VAX 2003.163 ************************