1 INFO-VAX	Fri, 05 May 2000	Volume 2000 : Issue 250       Contents:( Re: Compaq's business practices (update) dec/vax help needed  Re: dec/vax help needed $ External authentication in a cluster
 faster gunzip  Re: faster gunzip  Re: faster gunzip  Re: faster gunzip  Re: faster gunzip C Re: HP Laserjet Escape Sequences for VMS printing documented where? 1 Re: Humble request for a modification to VMS Mail 1 Re: Humble request for a modification to VMS Mail 1 Re: Humble request for a modification to VMS Mail 1 Re: Humble request for a modification to VMS Mail & Mailbox and synchronization in OpenVMS* Re: Mailbox and synchronization in OpenVMS Marketing opportunity  Re: Marketing opportunity  Re: OpenVMS disk space?  Re: OpenVMS marketing  Re: OpenVMS marketing ) Re: Printing in Landscape - (Please help)  Re: Reading VMS disks  Re: RMS indexed file cleanup Re: RMS indexed file cleanup Re: RMS indexed file cleanup Re: TPU for unix" Re: Web-enabling VMS reporting ..." Re: Web-enabling VMS reporting ..." Re: Web-enabling VMS reporting ...  F ----------------------------------------------------------------------  % Date: Thu, 04 May 2000 23:04:04 -0400 2 From: "Richard B. Gilbert" <DRAGON@compuserve.com>1 Subject: Re: Compaq's business practices (update) 7 Message-ID: <200005042304_MC2-A3D7-3D12@compuserve.com>   J         A couple of people contacted me about the original message I post= edJ with this topic.  They assured me that Bryan Marler was "full of it".   I=  H already believed that he was "full of it"; I have dealt with CPU OptionsF before, with complete satisfaction.   I ordered DEC/Compaq parts and I= expect to receive them, Mr. Marler's opinions to the contrary G notwithstanding!  I have been promised licenses that will both work and > stand up in court if Compaq is so unwise as to challenge them.  H         I refused to be bullied and my boss, Mr. Lloyd Curtis, backed meJ up.  There's not a pointy hair on his head!!!!   We have not canceled our=   order and will not.   J         The other addressees of the original message were Michael Capella= s J and his VPs plus the Antitrust Division of the U.S. Department of Justice= . =   G  I have not had a response from Compaq to date.  Neither have I had any G response from the DOJ.  I'm not certain that Mr. Marler's tactics would H form the basis for an Antitrust case against Compaq but I thought it was worth raising the point.  J         I hope, that if *you* are the victim of such sleazy tactics on th= e D part of Compaq that you will also stand your ground and refuse to beH bullied.  I'm certain that Compaq will ultimately sell more computers by@ being the low bidder than by threatening its would-be customers!   ------------------------------  # Date: Fri, 05 May 2000 02:43:56 GMT  From: chrislebeck@usa.net  Subject: dec/vax help needed) Message-ID: <8etch9$hr3$1@nnrp1.deja.com>   F i just aquired several pieces of dec equipment and i don't really knowC what they are all are or what i can do with them. all seem to be in  working order. here's the list:   G 2-vs49k-ab (vaxstation 4000 90). each with tape drives and dual seagate 
 st3610 drives   $ 3-vx20a-ea (not sure what these are)   2-gdm1961 monitors   3-vr320 monitors   1-vrc16 monitor   ! plus keyboards, mice, cords, etc.   / ----------------------------------------------- 8 the vaxstations boot up (with vax 5.5-2hw) but require a8 login name and password. is there a backdoor or default?  G can i do something else with these? like re-format the drives? possibly  put linux on one to play with?  % what type of processors are in these?   # are they worth anything to anybody?   H i know it's alot of questions....but for those of you who reply, thanks!    & Sent via Deja.com http://www.deja.com/ Before you buy.    ------------------------------  % Date: Thu, 04 May 2000 21:11:49 -0700 # From: Mark Tarka <markZERO@mcn.net>   Subject: Re: dec/vax help needed# Message-ID: <39124A05.6434@mcn.net>    chrislebeck@usa.net wrote: > K > i just a[c]quired several pieces of dec equipment and i don't really know E > what they are all are or what i can do with them. all seem to be in ! > working order. here's the list: 	 [snip...] : > the vaxstations boot up (with vax 5.5-2hw) but require a: > login name and password. is there a backdoor or default?  @ You need to do some homework.  Start with the FAQ for this group9 (somewhere around here....), then move to a decent search 7 engine (at the least, to get some feel for the "playing , field" -- vocabulary, reality checks, etc.).  I > can i do something else with these? like re-format the drives? possibly   > put linux on one to play with?   The sky's the limit!   ' > what type of processors are in these?   ; ERK!  Propriatary VAX cpus...running software known as VMS. 7 (Check this assertion...I know squat about the system.)      % > are they worth anything to anybody?   7 At the risk of being labled a predator...$100, and I'll  come and pick up the junk.   J > i know it's alot of questions....but for those of you who reply, thanks!  < Thanks isn't enough!  How much abuse can you take, or, OTOH, how deep are your pockets :-)         Mark   ------------------------------  # Date: Fri, 28 Apr 2000 22:28:11 GMT " From: falk@arc.ab.ca (Alfred Falk)- Subject: External authentication in a cluster 2 Message-ID: <8F24A6BF3falkarcabca@205.233.108.180>  D I've been testing out external authentication, and am having trouble with non-pathworks nodes.    Cluster setup is as follows 2   Node1: AlphaServer 400, OVMS 7.2, pathworks 6.0B?   Node2: VAXserver 3100, OVMS 7.2, external authentication only   1 External authentication works perfectly on node1. ; If I try to signon to Node2, I get following opcom message: D     %LOGIN-F-NOEXTAUTH, external authentication service disabled or  unavailable	  B If I turn on the debug trace (SYS$SINGLE_SIGNON = 80000001), I get    / 16:21:00.81: SYS$LOGONW - ACME state = 80000005 - 16:21:00.85: Function = ACME$_MAP_TO_USERNAME 0 16:21:00.85: Attempting to map Userid "Z_TESTER"# 16:21:00.86: Userid found in SYSUAF $ 16:21:00.86: ExtAuth is set for user9 16:21:30.70: Return status from ACME map routine follows: : 16:21:30.83: %SYSTEM-F-NOSUCHID, unknown rights identifier! 16:21:30.83: ACME reason code = 0 ! 16:21:30.88: IOSB status follows: G 16:21:30.89: %SYSTEM-F-ERRMAPAUTH, error mapping or authenticating user $ 16:21:30.89: IOSB devdepend follows:: 16:21:30.89: %SYSTEM-F-NOSUCHID, unknown rights identifier  2 Anyone know what this might mean or how to fix it?  @ ----------------------------------------------------------------A   A L B E R T A         Alfred Falk               falk@arc.ab.ca  @ R E S E A R C H         Information Systems Dept   (780)450-5185+   C O U N C I L         250 Karl Clark Road 1                         Edmonton, Alberta, Canada  http://www.arc.ab.ca/   T6N 1E4   http://www.arc.ab.ca/staff/falk/   ------------------------------   Date: 4 May 2000 18:34:25 GMT 2 From: mathog@seqaxp.bio.caltech.edu (David Mathog) Subject: faster gunzip, Message-ID: <8esfrh$rc9@gap.cco.caltech.edu>  J gunzip 1.2-4 on our OpenVMS 7.2-1 DS10 is very much slower than gunzip on G a DS10 running Linux/Alpha.  A big part of that is because on Linux you K don't really get the data onto disk unless you force it there - instead it  L is cached into the free memory on the system.  We've hashed over the pluses J and minuses of that method before and I don't want to start another thread on  it here.  J Just considering the direct write to disk though, gunzip, and sadly, just K about any simple C program, go about it on OpenVMS by doing about a zillion < small extends. One can speed things up a lot by first doing    $ set rms/extend=16000 $ gunzip bigfile.gz   K This is all a long winded way of asking if anybody has a version of gunzip  4 which has been optimized for write speed on OpenVMS.  H And an even longer winded way of asking why DEC C, by default, produces 5 programs which seem to be so poorly optimized for IO.   L I just went to see what MGFTP had done to speed up its IO - then found that L it's written in BLISS.  Not much help for speeding up C programs unless you " want to rewrite every one of them.  J Since we don't have any way to do disk caching as on Linux or WNT it wouldK be really nice if there was some way instead to control the behavior of the H standard C writes in DEC C better.  In particular I'd like to be able to@ configure it - without changing the C code - in these two ways:   $ 1.  cache up to N bytes before write* 2.  hold up to M milliseconds before writeG 3.  write the cache to disk in the fastest possible manner supported by      OpenVMS.  H To preserve current behavior, the default values for N and M would be 0.J When the file is closed the cache would flush to disk immediately without % waiting for more bytes or a time out.   G This could either go in as compiler switches or be run time adjustable  , (perhaps through some change to the C RTL).   E For gunzip a recompile would probably use parameters about like this:   7 $ cc/diskcache=(bytes=1000000,milliseconds=1000)/etc...   V Small files would unpack directly into cache and then be written out on the file closeL in one fell swoop.  Big files would fill the cache and write out a Megabyte 
 at a time.  $ If it was run time adjustable, then   0 $ define decc$rtl_diskcache_bytes        1000000- $ define decc$rtl_diskcache_milliseconds 1000  $ gunzip bigfile.gz   H A little tweak like this would go a long way towards speeding up many of& the C tools that come over from Unix.    Regards,   David Mathog mathog@seqaxp.bio.caltech.edu ? Manager, sequence analysis facility, biology division, Caltech     ------------------------------  " Date: Thu, 4 May 2000 21:33:22 GMT- From: young_r@eisner.decus.org (Robert Young)  Subject: Re: faster gunzip+ Message-ID: <N9OeQJt$hhSf@eisner.decus.org>   a In article <8esfrh$rc9@gap.cco.caltech.edu>, mathog@seqaxp.bio.caltech.edu (David Mathog) writes:   L > gunzip 1.2-4 on our OpenVMS 7.2-1 DS10 is very much slower than gunzip on I > a DS10 running Linux/Alpha.  A big part of that is because on Linux you M > don't really get the data onto disk unless you force it there - instead it  N > is cached into the free memory on the system.  We've hashed over the pluses L > and minuses of that method before and I don't want to start another thread > on  it here. > L > Just considering the direct write to disk though, gunzip, and sadly, just M > about any simple C program, go about it on OpenVMS by doing about a zillion > > small extends. One can speed things up a lot by first doing  >  > $ set rms/extend=16000 > $ gunzip bigfile.gz  > M > This is all a long winded way of asking if anybody has a version of gunzip  6 > which has been optimized for write speed on OpenVMS. > J > And an even longer winded way of asking why DEC C, by default, produces 7 > programs which seem to be so poorly optimized for IO.  > N > I just went to see what MGFTP had done to speed up its IO - then found that N > it's written in BLISS.  Not much help for speeding up C programs unless you $ > want to rewrite every one of them. > L > Since we don't have any way to do disk caching as on Linux or WNT it wouldM > be really nice if there was some way instead to control the behavior of the J > standard C writes in DEC C better.  In particular I'd like to be able toB > configure it - without changing the C code - in these two ways:  > & > 1.  cache up to N bytes before write, > 2.  hold up to M milliseconds before writeI > 3.  write the cache to disk in the fastest possible manner supported by  >     OpenVMS. >   D 	I suppose it has lot to do with environment too... disks I am usingB 	are hanging off controllers with write-back caching enabled (fast/ 	writes)... similar to what you are describing.   ? 	Hein sent this to me or pointed me to it a while ago, I posted 2 	it last November.  Realize this URL may go stale:  Q http://x25.deja.com/[ST_rn=ps]/getdoc.xp?AN=544491857&CONTEXT=957466989.896532492   G 	But can be easily located via www.deja.com Power search of comp.os.vms    	for:	 heuvel sampson     A 	If nothing else, one could grab the output file size from stored ? 	gunzip info (zip stores this info, suppose gunzip does to) and A 	allocate the exact needed size (the C code referenced in the URL ? 	above does that among other things).  Also, using this you can @ 	see it launches multiple I/Os using setup and io_perform.  This 	baby flies.  
 	Bottom line:   A 	Using this C code as a starting point , you can make I/O scream.   > 	I'll race Unix any old day.  Unix may win.  But then maybe we@ 	can drop similar code into a database program and run it acrossE 	multiple cluster members and see who can hammer disks faster.  Owww! D 	How do I write to memory in a Linux cluster and maintain filesystem 	integrity?  Tricky .. tricky.  C 	Then maybe we move the cluster members inside the same cabinet and 1 	call them "instances" and then we really race...    				Rob    ------------------------------  % Date: Thu, 04 May 2000 15:05:28 -0700 5 From: "Larry D Bohan, Jr" <LBohan@dbc.spam_less..com>  Subject: Re: faster gunzip2 Message-ID: <lPERORUnx4EysVkXpICGTSIuV3TI@4ax.com>  @ On 4 May 2000 18:34:25 GMT, mathog@seqaxp.bio.caltech.edu (David Mathog) wrote:  K >gunzip 1.2-4 on our OpenVMS 7.2-1 DS10 is very much slower than gunzip on  H >a DS10 running Linux/Alpha.  A big part of that is because on Linux youL >don't really get the data onto disk unless you force it there - instead it M >is cached into the free memory on the system.  We've hashed over the pluses  K >and minuses of that method before and I don't want to start another thread 
 >on  it here.   1 this was posted on C.O.V. a few weeks/months ago.   ? if the speed of gnuzip is really irritating, you might perhaps  B try to apply the techniques seen here.  I was saving this example 8 for the day I ran into something similarly irritating.    3 the perceived difference in speed btw this example  7 and a regular $COPY is impressive.  Even while already  8 using other techniques to speed up $COPY (large extents,- disls w/ lotsa of contiguous free space, etc)   7 It would be interesting if this  BUFOBJ implementation  > could be made optionally available for  $COPY, $CONVERT, etc. 1 (i imagine (hyper)SORT  and  BACKUP already have    similar optimizations)   + or for that matter, the C run-time library.      /*4  * FAST_IO_COPY.C - disk file copy Fast I/O example.5  * This program requires OpenVMS V7.0 (VAX or Alpha). 3  * The DEC C V5.2 (or newer) compiler is preferred.   */  #include atrdef  #include descrip #include fibdef  #include iodef   #include iosadef   #include lib$routines 
 #include rms   #include ssdef   #include starlet   #include stdio   #include stdlib  #include string    #include "fatdef.h"      #define NBYTES_PER_BLOCK 512 #define MAX_BUFFERS 16 #define MAX_BLOCKS_PER_XFER 127 A #define MAX_BYTES_PER_XFER (NBYTES_PER_BLOCK*MAX_BLOCKS_PER_XFER) B #define MAX_BYTES_PER_BUFFER (NBYTES_PER_BLOCK+MAX_BYTES_PER_XFER)   #pragma extern_model save % #pragma extern_model common_block shr  #pragma nostandard" extern volatile _align (page) char* buffer[MAX_BUFFERS][MAX_BYTES_PER_BUFFER]; #pragma standard) extern volatile fandle ifan[MAX_BUFFERS]; ) extern volatile fandle ofan[MAX_BUFFERS]; * extern volatile int iohan[MAX_BUFFERS][2]; extern struct FAB ifab;  extern struct NAM inam; $ extern char iexpand[NAM$C_MAXRSS+1];$ extern char iresult[NAM$C_MAXRSS+1]; extern struct FAB ofab;  extern struct NAM onam; $ extern char oexpand[NAM$C_MAXRSS+1];$ extern char oresult[NAM$C_MAXRSS+1]; extern struct XABFHC xabfhc; extern struct XABSUM xabsum;! extern struct XABALL xaball[255]; ! extern struct XABKEY xabkey[255];  extern char knm[255][32];  extern volatile int ivbn;  extern volatile int iefn; ' extern volatile int nreads_outstanding; ( extern volatile int nwrites_outstanding;* extern volatile int max_reads_outstanding;+ extern volatile int max_writes_outstanding;  extern volatile int nbuffers; % extern volatile int nblocks_per_xfer; $ extern volatile int nbytes_per_xfer;& extern volatile int nbytes_per_buffer; #pragma extern_model restore  % void read_completion_ast(iosa *piosa)  { -   register int status = piosa->iosa$l_status;0-   register int nbytes = piosa->iosa$ih_count;s3   register int ibuf   = piosa->iosa$q_context_q[0];e  1   if (nreads_outstanding > max_reads_outstanding)t/     max_reads_outstanding = nreads_outstanding;r     nreads_outstanding--;   1   if ((status != SS$_ENDOFFILE) && !(status & 1))s   {V     int msgvec[2] = {1,0};     msgvec[1] = status;rD     printf("Could not perform input from VBN %0d to buffer #%0d.\n",*         piosa->iosa$q_context_q[1], ibuf);     sys$putmsg(msgvec,0,0,0);r     goto read_return;o   }V  
   if (nbytes)    {y     ifab.fab$l_ctx++;S  7     status = sys$io_perform(ofan[ibuf], ofab.fab$l_stv,o:         &buffer[ibuf][0], &buffer[ibuf][NBYTES_PER_BLOCK],,         nbytes, piosa->iosa$q_context_q[1]);     if (!(status & 1))     {        int msgvec[2] = {1,0};       msgvec[1] = status; 
       printf(eB         "Could not start output (%0d bytes) to VBN %0d from buffer	 #%0d.\n", 2         nbytes, piosa->iosa$q_context_q[1], ibuf);       sys$putmsg(msgvec,0,0,0);V       goto read_return;-     }-     nwrites_outstanding++;   }-   read_return:>   if ((nreads_outstanding == 0) && (nwrites_outstanding == 0))     sys$setef(iefn);	   return;o } /* read_completion_ast */i  & void write_completion_ast(iosa *piosa) {0-   register int status = piosa->iosa$l_status; -   register int nbytes = piosa->iosa$ih_count;n3   register int ibuf   = piosa->iosa$q_context_q[0];d  3   if (nwrites_outstanding > max_writes_outstanding) 1     max_writes_outstanding = nwrites_outstanding;h     nwrites_outstanding--;     if (!(status & 1))   {e     int msgvec[2] = {1,0};     msgvec[1] = status;r     printf(iD         "Could not perform output (%0d bytes) to VBN %0d from buffer	 #%0d.\n", 2         nbytes, piosa->iosa$q_context_q[1], ibuf);     sys$putmsg(msgvec,0,0,0);      goto write_return;   }    ofab.fab$l_ctx++;d     if (ivbn <= ifab.fab$l_alq)n   {p+     piosa->iosa$ih_count = nbytes_per_xfer;e&     piosa->iosa$q_context_q[1] = ivbn;7     status = sys$io_perform(ifan[ibuf], ifab.fab$l_stv,h:         &buffer[ibuf][0], &buffer[ibuf][NBYTES_PER_BLOCK],         nbytes_per_xfer, ivbn);      if (!(status & 1))     {n       int msgvec[2] = {1,0};       msgvec[1] = status;e
       printf( ?         "Could not start input from VBN %0d to buffer #%0d.\n",f         ivbn, ibuf);       sys$putmsg(msgvec,0,0,0);s       goto write_return;     }      ivbn += nblocks_per_xfer;h     nreads_outstanding++;t   }n  
 write_return: >   if ((nreads_outstanding == 0) && (nwrites_outstanding == 0))     sys$setef(iefn);	   return;i } /* write_completion_ast */  ( int fast_io_copy(int argc, char *argv[]) #pragma nostandard main_program #pragma standard {-   register iosa *piosa = 0;0   register int status = 0;   register int ibuf = 0;"   volatile char *inadr[2] = {0,0};   char *retadr[2] = {0,0};   unsigned short int iosb[4];e   struct atrdef atrlst[2];   struct fibdef fib;!   struct dsc$descriptor_s fibdsc;t
   FAT fat;   /*0  * Require input and output file specifications.  */b   if (argc < 3)    {e>     printf("Usage: %s <input-file> <output-file>\n", argv[0]);
     return 1;w   }h   /*.  * Allow additional arguments to change either3  * the buffer count or the transfer size in blocks.   */-   nbuffers = MAX_BUFFERS;-)   nblocks_per_xfer = MAX_BLOCKS_PER_XFER;   !   for (ibuf=3; ibuf<argc; ibuf++)i   {e-     if (!strncmp(argv[ibuf], "-nbuffers", 9))      { 
       ibuf++;s       if (ibuf < argc)       {d$         nbuffers = atoi(argv[ibuf]);8         if ((nbuffers == 0) || (nbuffers > MAX_BUFFERS))!           nbuffers = MAX_BUFFERS;          else if (nbuffers < 0)           nbuffers = 1;h       }t     } 1     else if (!strncmp(argv[ibuf], "-nblocks", 8))y     { 
       ibuf++;-       if (ibuf < argc)       {0,         nblocks_per_xfer = atoi(argv[ibuf]);:         if ((nblocks_per_xfer == 0) || (nblocks_per_xfer > MAX_BLOCKS_PER_XFER))n1           nblocks_per_xfer = MAX_BLOCKS_PER_XFER;a&         else if (nblocks_per_xfer < 0)           nblocks_per_xfer = 1;t       }      }      else     {w;       printf("Unrecognized argument #%0d: \"%s\".\n", ibuf,t argv[ibuf]);     }    }a  8   nbytes_per_xfer = NBYTES_PER_BLOCK * nblocks_per_xfer;9   nbytes_per_buffer = NBYTES_PER_BLOCK + nbytes_per_xfer;t  <   printf("Using %0d buffers, with %0d blocks (%0d bytes) per
 transfer.\n",.5         nbuffers, nblocks_per_xfer, nbytes_per_xfer);t  %   for (ibuf=0; ibuf<nbuffers; ibuf++)l   {       inadr[0] = &buffer[ibuf][0];2     inadr[1] = &buffer[ibuf][nbytes_per_buffer-1];  A     status = sys$create_bufobj(inadr, retadr, 0, 0, iohan[ibuf]);k     if (!(status & 1))     { =       printf("Could not create buffer object #%0d.\n", ibuf);n       return status;     }   '     status = sys$io_setup(IO$_READVBLK,t:         iohan[ibuf], iohan[ibuf], &read_completion_ast, 0,
 &ifan[ibuf]);f     if (!(status & 1))     {,@       printf("Could not set up input for buffer #%0d.\n", ibuf);       return status;     }   (     status = sys$io_setup(IO$_WRITEVBLK,;         iohan[ibuf], iohan[ibuf], &write_completion_ast, 0, 
 &ofan[ibuf]);      if (!(status & 1))     {-A       printf("Could not set up output for buffer #%0d.\n", ibuf);c       return status;     }:   }r   /*-  * Fill in two RMS extended attribute blocks.B  */k(   bzero((char*)&xabfhc, sizeof(xabfhc));   xabfhc.xab$b_cod = XAB$C_FHC;a"   xabfhc.xab$b_bln = XAB$C_FHCLEN;$   xabfhc.xab$l_nxt = (char*)&xabsum;  (   bzero((char*)&xabsum, sizeof(xabsum));   xabsum.xab$b_cod = XAB$C_SUM;r"   xabsum.xab$b_bln = XAB$C_SUMLEN;   /*%  * Fill in the input file RMS blocks.e  */l$   bzero((char*)&ifab, sizeof(ifab));   ifab.fab$b_bid = FAB$C_BID;a   ifab.fab$b_bln = FAB$C_BLN;x   ifab.fab$l_nam = &inam;c)   ifab.fab$l_fop = FAB$M_NAM | FAB$M_UPI;n   ifab.fab$l_fna = argv[1];S#   ifab.fab$b_fns = strlen(argv[1]);0   ifab.fab$l_dna = 0;s   ifab.fab$b_dns = 0;:   ifab.fab$b_fac = FAB$M_GET;U    ifab.fab$b_shr = FAB$M_SHRGET;  $   bzero((char*)&inam, sizeof(inam));   inam.nam$b_bid = NAM$C_BID;0   inam.nam$b_bln = NAM$C_BLN;6   inam.nam$l_esa = iexpand;     inam.nam$b_ess = NAM$C_MAXRSS;   inam.nam$l_rsa = iresult;k    inam.nam$b_rss = NAM$C_MAXRSS;   /*&  * Fill in the output file RMS blocks.  */f$   bzero((char*)&ofab, sizeof(ofab));   ofab.fab$b_bid = FAB$C_BID;i   ofab.fab$b_bln = FAB$C_BLN;    ofab.fab$l_nam = &onam;05   ofab.fab$l_fop = FAB$M_CBT | FAB$M_MXV | FAB$M_OFP;n   ofab.fab$l_fna = argv[2];n#   ofab.fab$b_fns = strlen(argv[2]);-   ofab.fab$l_dna = 0;-   ofab.fab$b_dns = 0;-5   ofab.fab$b_fac = FAB$M_BIO | FAB$M_PUT | FAB$M_TRN;k   ofab.fab$b_shr = FAB$M_NIL; "   ofab.fab$l_alq = ifab.fab$l_alq;"   ofab.fab$w_deq = ifab.fab$w_deq;"   ofab.fab$b_org = ifab.fab$b_org;"   ofab.fab$b_rat = ifab.fab$b_rat;"   ofab.fab$b_rfm = ifab.fab$b_rfm;"   ofab.fab$w_mrs = ifab.fab$w_mrs;"   ofab.fab$l_mrn = ifab.fab$l_mrn;"   ofab.fab$b_bks = ifab.fab$b_bks;"   ofab.fab$b_fsz = ifab.fab$b_fsz;"   ofab.fab$w_gbc = ifab.fab$w_gbc;  $   bzero((char*)&onam, sizeof(onam));   onam.nam$b_bid = NAM$C_BID;M   onam.nam$b_bln = NAM$C_BLN;r   onam.nam$l_esa = oexpand;i    onam.nam$b_ess = NAM$C_MAXRSS;   onam.nam$l_rsa = oresult;o    onam.nam$b_rss = NAM$C_MAXRSS;   onam.nam$l_rlf = &inam;    /*:  * Fill in the invariant data for truncating output files.  */e'   atrlst[0].atr$w_size = ATR$S_RECATTR;e'   atrlst[0].atr$w_type = ATR$C_RECATTR;t #if __ALPHAn   atrlst[0].atr$l_addr = &fat; #elsec,   atrlst[0].atr$l_addr = (unsigned int)&fat; #endif     atrlst[1].atr$w_size = 0;t   atrlst[1].atr$w_type = 0;    atrlst[1].atr$l_addr = 0;e  &   fibdsc.dsc$w_length  = FIB$C_LENGTH;'   fibdsc.dsc$b_dtype   = DSC$K_DTYPE_T;i'   fibdsc.dsc$b_class   = DSC$K_CLASS_S;f%   fibdsc.dsc$a_pointer = (char*)&fib;z   /*  * Get an event flag.d  */w   lib$get_ef(&iefn);   if (iefn < 0) iefn = 32;   /*&  * Parse the input file specification.  */r   status = sys$parse(&ifab);   if (!(status & 1))   {      unsigned int msgvec[3];a     msgvec[0] = 2;     msgvec[1] = ifab.fab$l_sts;i     msgvec[2] = ifab.fab$l_stv; @     printf("Could not parse input file specification \"%s\".\n",	 argv[1]);c     sys$putmsg(msgvec,0,0,0);i     return status;   }    /*"  * Search for the next input file.  */w!   iexpand[inam.nam$b_esl] = '\0';o"   printf("Search: %s\n", iexpand);*   while ((status = sys$search(&ifab)) & 1)   {f#     iresult[inam.nam$b_rsl] = '\0';e$     printf("Reading %s\n", iresult);   /*/  * Open the input file (with implicit display).   */e     xabsum.xab$l_nxt = 0; $     ifab.fab$l_xab = (char*)&xabfhc;     ifab.fab$l_ctx = 0;M+     ifab.fab$l_fop = FAB$M_NAM | FAB$M_UPI;l     status = sys$open(&ifab);.     if (!(status & 1))     {a       unsigned int msgvec[3];t       msgvec[0] = 2;!       msgvec[1] = ifab.fab$l_sts;h!       msgvec[2] = ifab.fab$l_stv;w-       printf("Could not open input file.\n");r       sys$putmsg(msgvec,0,0,0);        continue;c     }w   /*>  * Get all available information on file allocation and areas.  */0:     ifab.fab$l_xab = xabsum.xab$l_nxt = (char*)&xaball[0];@     for (ibuf=0; (ibuf<xabsum.xab$b_noa) || (ibuf == 0); ibuf++)     { 8       bzero((char*)&xaball[ibuf], sizeof(xaball[ibuf]));)       xaball[ibuf].xab$b_cod = XAB$C_ALL;j,       xaball[ibuf].xab$b_bln = XAB$C_ALLLEN;$       xaball[ibuf].xab$b_aid = ibuf;6       xaball[ibuf].xab$l_nxt = (char*)&xaball[ibuf+1];     }e     if (xabsum.xab$b_noa) /       xaball[xabsum.xab$b_noa-1].xab$l_nxt = 0;      else       xaball[0].xab$l_nxt = 0;   /*C  * If there are any indexes, get all of the information about them.c  */,     if (xabsum.xab$b_nok)      {-       if (xabsum.xab$b_noa) A         xaball[xabsum.xab$b_noa-1].xab$l_nxt = (char*)&xabkey[0];(
       else0         xaball[0].xab$l_nxt = (char*)&xabkey[0];  1       for (ibuf=0; ibuf<xabsum.xab$b_nok; ibuf++)g       {l:         bzero((char*)&xabkey[ibuf], sizeof(xabkey[ibuf]));+         xabkey[ibuf].xab$b_cod = XAB$C_KEY;S.         xabkey[ibuf].xab$b_bln = XAB$C_KEYLEN;&         xabkey[ibuf].xab$b_ref = ibuf;+         xabkey[ibuf].xab$l_knm = knm[ibuf];t8         xabkey[ibuf].xab$l_nxt = (char*)&xabkey[ibuf+1];       }c/       xabkey[xabsum.xab$b_nok-1].xab$l_nxt = 0;h     }v   /*>  * Display the additional information about allocation, areas, indexes.  */       status = sys$display(&ifab);     if (!(status & 1))     {o       unsigned int msgvec[3];,       msgvec[0] = 2;!       msgvec[1] = ifab.fab$l_sts;V!       msgvec[2] = ifab.fab$l_stv;l<       printf("Could not display input file information.\n");       sys$putmsg(msgvec,0,0,0);z       sys$close(&ifab);s       continue;n     }    /*  * Close the input file.  */      ifab.fab$l_xab = 0;n     status = sys$close(&ifab);     if (!(status & 1))     {r       unsigned int msgvec[3];,       msgvec[0] = 2;!       msgvec[1] = ifab.fab$l_sts;o!       msgvec[2] = ifab.fab$l_stv;t.       printf("Could not close input file.\n");       sys$putmsg(msgvec,0,0,0);i       continue;o     }h   /*+  * Show area and index information, if any.t  */r/     for (ibuf=0; ibuf<xabsum.xab$b_noa; ibuf++) ;       printf("Area %03u; %0u blocks; %0u-block buckets.\n",r         xaball[ibuf].xab$b_aid,i         xaball[ibuf].xab$l_alq,           xaball[ibuf].xab$b_bkz);  /     for (ibuf=0; ibuf<xabsum.xab$b_nok; ibuf++)o:       printf("Key  %03u; LAN %03u; IAN %03u; \"%.*s\".\n",A         ibuf, xabkey[ibuf].xab$b_lan, xabkey[ibuf].xab$b_ian, 32,m knm[ibuf]);e   /*?  * Re-open the input file, UFO this time (accessed on user-moder	 channel).b  */       ifab.fab$l_fop |= FAB$M_UFO;     status = sys$open(&ifab);t     if (!(status & 1))     {n       unsigned int msgvec[3];-       msgvec[0] = 2;!       msgvec[1] = ifab.fab$l_sts;o!       msgvec[2] = ifab.fab$l_stv;i:       printf("Could not access input file on channel.\n");       sys$putmsg(msgvec,0,0,0);g       continue;/     }j   /*C  * Create the output file, using the RMS extended attribute blocks.t  */b$     ofab.fab$l_xab = (char*)&xabfhc;     ofab.fab$l_ctx = 0; 7     ofab.fab$l_fop = FAB$M_CBT | FAB$M_MXV | FAB$M_OFP; $     ofab.fab$l_alq = ifab.fab$l_alq;$     ofab.fab$b_org = ifab.fab$b_org;$     ofab.fab$b_rat = ifab.fab$b_rat;$     ofab.fab$b_rfm = ifab.fab$b_rfm;$     ofab.fab$w_mrs = ifab.fab$w_mrs;     status = sys$create(&ofab);g     if (!(status & 1))     {s       unsigned int msgvec[3];i       msgvec[0] = 2;!       msgvec[1] = ofab.fab$l_sts;g!       msgvec[2] = ofab.fab$l_stv;y0       printf("Could not create output file.\n");       sys$putmsg(msgvec,0,0,0);a       sys$close(&ifab);        continue;a     }r   /*F  * If the total allocation of the output file is less than that of theF  * input file, extend it to match the input file.  This can happen forD  * non-sequential files when the total blocks specifically allocated to2  * each area do not result in a large enough file.  */n     ofab.fab$l_xab = 0;l(     if (ofab.fab$l_alq < ifab.fab$l_alq)     {->       printf("Extending output file allocation from %0u to %0u blocks.\n",L(         ofab.fab$l_alq, ifab.fab$l_alq);7       ofab.fab$l_alq = ifab.fab$l_alq - ofab.fab$l_alq;a!       status = sys$extend(&ofab);T       if (!(status & 1))       {v         unsigned int msgvec[3];-         msgvec[0] = 2;#         msgvec[1] = ofab.fab$l_sts;n#         msgvec[2] = ofab.fab$l_stv;i2         printf("Could not extend output file.\n");!         sys$putmsg(msgvec,0,0,0);          sys$close(&ifab);          sys$close(&ofab);m         continue;.       }a     }e   /*  * Close the output file.h  */t     status = sys$close(&ofab);     if (!(status & 1))     {e       unsigned int msgvec[3];V       msgvec[0] = 2;!       msgvec[1] = ofab.fab$l_sts;e!       msgvec[2] = ofab.fab$l_stv; /       printf("Could not close output file.\n");a       sys$putmsg(msgvec,0,0,0);        sys$close(&ifab);r       continue;      }e  #     oresult[onam.nam$b_rsl] = '\0';a$     printf("Writing %s\n", oresult);   /*+  * Re-open the output file by NAM block andt1  * UFO this time (accessed on user-mode channel).   */g,     ofab.fab$l_fop |= FAB$M_NAM | FAB$M_UFO;     status = sys$open(&ofab);i     if (!(status & 1))     {a       unsigned int msgvec[3];e       msgvec[0] = 2;!       msgvec[1] = ofab.fab$l_sts; !       msgvec[2] = ofab.fab$l_stv;t;       printf("Could not access output file on channel.\n");.       sys$putmsg(msgvec,0,0,0);l       sys$close(&ifab);r       continue;V     }l   /*@  * Set the completion event flag.  Clear outstanding I/O counts.  */a     sys$setef(iefn);     nreads_outstanding = 0;e     nwrites_outstanding = 0;     max_reads_outstanding = 0;     max_writes_outstanding = 0;e   /*-  * Start as many as <nbuffers> read requests.u  */r     for (       ibuf=0, ivbn=1; >                 ((ibuf<nbuffers) && (ivbn <= ifab.fab$l_alq));                 ibuf++)d     {A       sys$setast(0);       if (ivbn == 1)       {R         sys$clref(iefn);         lib$init_timer(0);       }R  &       piosa = (iosa*)&buffer[ibuf][0];       piosa->iosa$l_status = 1; -       piosa->iosa$ih_count = nbytes_per_xfer;a(       piosa->iosa$q_context_q[0] = ibuf;(       piosa->iosa$q_context_q[1] = ivbn;  9       status = sys$io_perform(ifan[ibuf], ifab.fab$l_stv,A:         &buffer[ibuf][0], &buffer[ibuf][NBYTES_PER_BLOCK],         nbytes_per_xfer, ivbn);r       if (!(status & 1))       {[         sys$setast(1);F         printf("Could not start input from VBN %0d to buffer #%0d.\n",                 ivbn, ibuf);         return status;       }r       ivbn += nblocks_per_xfer;r       nreads_outstanding++;r       sys$setast(1);     }r   /*2  * Wait for the disk reads and writes to complete.  */x     sys$waitfr(iefn);      lib$show_timer(0);<     printf("     Total Fast I/O Reads: %0d; Writes: %0d.\n",(         ifab.fab$l_ctx, ofab.fab$l_ctx);<     printf("Maximum Outstanding Reads: %0d; Writes: %0d.\n",7         max_reads_outstanding, max_writes_outstanding);    /*  * Deassign the input channel.  */_(     status = sys$dassgn(ifab.fab$l_stv);     if (!(status & 1))     {t4       printf("Could not deassign input channel.\n");       return status;     }a   /*D  * Completely re-initialize the file information block.  The file ID andiA  * directory ID fields must be zeroed, or the XQP will attempt to_ truncateB  * and modify the attributes of the *original* or *previous* file, *not*n6  * the file that is currently accessed on the channel.  */($     bzero((char*)&fib, sizeof(fib));&     fib.fib$l_acctl = FIB$M_WRITETHRU;"     fib.fib$w_exctl = FIB$M_TRUNC;     fib.fib$l_exsz  = 0;'     fib.fib$l_exvbn = xabfhc.xab$l_ebk; ,     if (xabfhc.xab$w_ffb) fib.fib$l_exvbn++;     fib.fib$w_verlimit = 0;o     fib.fib$l_acl_status = 0;(   /*?  * Copy the file header characteristics to the file attributes.f  */u*     fat.fat$b_rtype    = xabfhc.xab$b_rfo;*     fat.fat$b_rattrib  = xabfhc.xab$b_atr;*     fat.fat$w_rsize    = xabfhc.xab$w_lrl;9     fat.fat$w_hiblkl   = xabfhc.xab$w_hbk0; /* ignored */ 9     fat.fat$w_hiblkh   = xabfhc.xab$w_hbk2; /* ignored */t+     fat.fat$w_efblkl   = xabfhc.xab$w_ebk0;f+     fat.fat$w_efblkh   = xabfhc.xab$w_ebk2;_*     fat.fat$w_ffbyte   = xabfhc.xab$w_ffb;*     fat.fat$b_bktsize  = xabfhc.xab$b_bkz;*     fat.fat$b_vfcsize  = xabfhc.xab$b_hsz;*     fat.fat$w_maxrec   = xabfhc.xab$w_mrz;*     fat.fat$w_defext   = xabfhc.xab$w_dxq;*     fat.fat$w_gbc      = xabfhc.xab$w_gbc;/     fat.fat$w_versions = xabfhc.xab$w_verlimit;    /*   * Modify and truncate the file.  */ C     status = sys$qiow(iefn, ofab.fab$l_stv, IO$_MODIFY, iosb, 0, 0, .                 &fibdsc, 0, 0, 0, &atrlst, 0);%     if (status & 1) status = iosb[0];      if (status & 1)t     {=A       printf("Output file truncated at VBN %0u by %0u blocks.\n",s1                 fib.fib$l_exvbn, fib.fib$l_exsz);c>       printf("New end-of-file block is %0u, first free byte is %0u.\n",4                 xabfhc.xab$l_ebk, xabfhc.xab$w_ffb);     }i     else     {]E       printf("Could not truncate output file to EBK %0u, FFB %0u.\n", 4                 xabfhc.xab$l_ebk, xabfhc.xab$w_ffb);     }    /*  * Deassign the output channel.f  */ (     status = sys$dassgn(ofab.fab$l_stv);     if (!(status & 1))     {i5       printf("Could not deassign output channel.\n");e       return status;     }     } /* input file search loop */  ,   if ((status != RMS$_NMF) && !(status & 1))   {1     unsigned int msgvec[3];s     msgvec[0] = 2;     msgvec[1] = ifab.fab$l_sts;o     msgvec[2] = ifab.fab$l_stv; 1     printf("Could not search for input file.\n");s     sys$putmsg(msgvec,0,0,0);    }    /*,  * Clean up after <nbuffers> buffer objects.2  * This apparently gets done at image exit anyway.  */=%   for (ibuf=0; ibuf<nbuffers; ibuf++)    {e(     status = sys$io_cleanup(ifan[ibuf]);     if (!(status & 1))     {iB       printf("Could not clean up input for buffer #%0d.\n", ibuf);       return status;     }p  (     status = sys$io_cleanup(ofan[ibuf]);     if (!(status & 1))     {eC       printf("Could not clean up output for buffer #%0d.\n", ibuf);t       return status;     }e  ,     status = sys$delete_bufobj(iohan[ibuf]);     if (!(status & 1))     {a=       printf("Could not delete buffer object #%0d.\n", ibuf);b       return status;     }    }f   return status; } /* fast_io_copy() */   ------------------------------  " Date: Fri, 5 May 2000 05:42:33 GMT- From: young_r@eisner.decus.org (Robert Young)s Subject: Re: faster gunzip+ Message-ID: <U7abK4B9M7Mu@eisner.decus.org>r  j In article <lPERORUnx4EysVkXpICGTSIuV3TI@4ax.com>, "Larry D Bohan, Jr" <LBohan@dbc.spam_less..com> writes:B > On 4 May 2000 18:34:25 GMT, mathog@seqaxp.bio.caltech.edu (David > Mathog) wrote: > L >>gunzip 1.2-4 on our OpenVMS 7.2-1 DS10 is very much slower than gunzip on I >>a DS10 running Linux/Alpha.  A big part of that is because on Linux youfM >>don't really get the data onto disk unless you force it there - instead it yN >>is cached into the free memory on the system.  We've hashed over the pluses L >>and minuses of that method before and I don't want to start another thread >>on  it here. > 3 > this was posted on C.O.V. a few weeks/months ago.E >   : 	Months ago.  November 1999 to be somewhat precise.  I was8 	trying to save some bandwidth.  But it is probably good7 	that it is out here again as the recently pitiful Deja = 	seems to be aging their archives and things are dropping offO! 	the face of the earth out there.r  = 	Hey Altavista folks, I like ragingsearch.altavista.com  ...  F 	do us all a favor and improve your Usenet archives and go after Deja.   	Just a suggestion.r  5 > the perceived difference in speed btw this example f9 > and a regular $COPY is impressive.  Even while already =  < 	Perceived?  I perceive it to be faster by watching a clock.   				Robr   ------------------------------  " Date: Fri, 5 May 2000 06:40:48 GMT- From: young_r@eisner.decus.org (Robert Young)i Subject: Re: faster gunzip+ Message-ID: <a6cxbhuG1n2B@eisner.decus.org>D  j In article <lPERORUnx4EysVkXpICGTSIuV3TI@4ax.com>, "Larry D Bohan, Jr" <LBohan@dbc.spam_less..com> writes:  ? Wait a second... this version below is different than what HeinfL forwarded and what Bob Sampson wrote.  Just looking at the includes and the  comments that Bob had in there:    #include <atrdef.h>0 #include <descrip.h>   #if __ALPHA&< #include <fatdef.h> /* +sys$library:sys$lib_c.tlb/library */ #elsenJ #include "fatdef.h" /* extract this header for VAX from an Alpha system */ #endif   #include <fibdef.h>  #include <iodef.h> #include <iosadef.h> #include <lib$routines.h>F #include <rms.h> #include <ssdef.h> #include <starlet.h> #include <stdio.h> #include <stdlib.h>s #include <string.h>       : 	I mean if you are going to change someone's code for your5 	own personal use, that's fine but it is pretty silly = 	to post that change back out here.  Worse than silly if whatB0 	you stick back out here won't compile properly.  C 	Someone stumbling upon your own whacked up version would read the lC 	brief comment and assume it would compile and run on a VAX and of dI 	course it wouldn't and then *may* not have the slightest idea how to getbA 	a fatdef.h file and so would be back out here asking about that.n  H 	But that is one nitpick.  More importantly , mentioned above, what you # 	stuck out here won't even compile.C  / 	Hummm.. maybe being too harsh.. best to check:b  
 $ cc /versionS+ Compaq C V6.2-003 on OpenVMS Alpha V7.2-1  k  
 $ type t.c  J #include "fatdef.h" /* extract this header for VAX from an Alpha system */ #include <stdio.h>   main()	         {n          printf("hello world\n");	         }V   $ cc t  J #include "fatdef.h" /* extract this header for VAX from an Alpha system */ .^O %CC-F-NOINCLFILEF, Cannot find file "fatdef.h" specified in #include directive.|/ at line number 2 in file USER_DISK:[YOUNGR]T.C;f   $ type t2.ci   #if __ALPHA < #include <fatdef.h> /* +sys$library:sys$lib_c.tlb/library */ #else;J #include "fatdef.h" /* extract this header for VAX from an Alpha system */ #endif   #include <stdio.h>   main()	         {=          printf("hello world\n");	         }$  * $ cc t2 +sys$library:sys$lib_c.tlb/library  : 	Yep, sure compiles a lot cleaner following the directionsB 	in the ORIGINAL comments.  I suspect it works for you because you- 	extracted your own private copy of fatdef.h.m     				RobC     >  > /*6 >  * FAST_IO_COPY.C - disk file copy Fast I/O example.7 >  * This program requires OpenVMS V7.0 (VAX or Alpha).r5 >  * The DEC C V5.2 (or newer) compiler is preferred.0 >  */_ > #include atrdefR > #include descrip > #include fibdef  > #include iodef   > #include iosadef r > #include lib$routines  > #include rms ( > #include ssdef   > #include starlet . > #include stdio   > #include stdlib  > #include string  > E > #include "fatdef.h"    <-------------------  Bzzzzzzzzzztttt!!!!!!!d >  > #define NBYTES_PER_BLOCK 512 > #define MAX_BUFFERS 16! > #define MAX_BLOCKS_PER_XFER 127rC > #define MAX_BYTES_PER_XFER (NBYTES_PER_BLOCK*MAX_BLOCKS_PER_XFER) D > #define MAX_BYTES_PER_BUFFER (NBYTES_PER_BLOCK+MAX_BYTES_PER_XFER) >    	[snip]s   ------------------------------  % Date: Thu, 04 May 2000 23:04:02 -0400]2 From: "Richard B. Gilbert" <DRAGON@compuserve.com>L Subject: Re: HP Laserjet Escape Sequences for VMS printing documented where?7 Message-ID: <200005042304_MC2-A3D7-3D11@compuserve.com>g           Gerald,e  .         Please consider posting your solution!  J         The solutions that used to work with older H-P printers no longer=  J work!  The <ESC>P<string><ESC>\ sequence no longer works because the newe= rmJ H-P printers (HP8100) do not ignore the ANSI DCS and ST escape sequences.=      & Message text written by "Gerald Marsh"J >The VMS string is only relevant to pathworks (I think!) - I've got aroun= dxJ the blank page problem with a combination of escape sequences and logical=  D names - Please send email to gerald.marsh@barclays.co.uk if you want detailss& of the text library modules I'm using.   (It's a swine of a problem!)   Gerald.<   ------------------------------   Date: 4 May 2000 20:44:56 GMTm# From: system@niuhep.physics.niu.edu : Subject: Re: Humble request for a modification to VMS Mail+ Message-ID: <8esng8$dbn$1@husk.cso.niu.edu>u  1 jlauret@?.chem.sunysb.edu (Jerome LAURET) writes:b >xE >	Then, you might consider PINE for VMS. A 'MUST' !! I use it all theaI >time and rarely use VMS mail (unless I have to move a bunch of selected L* >messages from folder to folders/drawers).  C >	In my opinion (don't blame/flame for this) : I don't think it is tJ >relevant to spend time on further developping the native Email interface C >as long as freeware exists and do the job just fine if not better _E >(attachment etc... in pine is there ... people always ask how to do  ; >that from native interface. I don't quite understand why).   	 No flame.b  8 Not all installations will have the freeware installed,   L Some installations won't allow freeware, (see the case of DELIVER as to why)  ? It takes vmsmail and DecMail under 10 seconds to access my MAIL <   folder, it takes pine 3.91-2 (vms) 30 seconds to a minute.  D To sell VMS to a corp. I don't think freeware is nearly as useful as native capability.  I That said, I am very glad that pine is available, I do use it on occasion]- and some of my users use it fairly regularly.b   But you did ask why. Morphis@physics.niu.edu_8 Real Women change tires			abuse@uu.net postmaster@uu.net7 Real Men change diapers                 security@uu.net    ------------------------------   Date: 4 May 2000 20:56:04 GMT # From: system@niuhep.physics.niu.edu : Subject: Re: Humble request for a modification to VMS Mail+ Message-ID: <8eso54$dbn$3@husk.cso.niu.edu>l  # Shane.F.Smith@healthnet.com writes:"H >David Spencer <spencer@recneps.spaamfree.com> on 05/02/2000 04:34:14 PM  I >>For years I have clung happily to VMS mail as my primary emailer. And InI >>wish to continue to do so. I just have one simple little wish: couldn'tiE >>it be enhanced to include a way to modify message subject lines? Sol  D >This can already be done with the DECwindows interface to VMS mail.  ) As I further expose my ignorance ... How?,   Morphis@physics.niu.edu 8 Real Women change tires			abuse@uu.net postmaster@uu.net7 Real Men change diapers                 security@uu.net    ------------------------------   Date: 4 May 2000 20:52:07 GMT # From: system@niuhep.physics.niu.edua: Subject: Re: Humble request for a modification to VMS Mail+ Message-ID: <8esntn$dbn$2@husk.cso.niu.edu>o  0 Michael Austin <maustin@nc.prestige.net> writes:  K >I usually don't use the message title to tell me what I am looking for.  InK >would generally use the search feature to locate documents, especially if m >I don't recall where I put it.f  @ I feel really stupid, how long has search been a part of mail?!?  > Yeesh!  Well, now I know how to look for those lost documents.   Thanks,  Robert Morphis@physics.niu.edu 8 Real Women change tires			abuse@uu.net postmaster@uu.net7 Real Men change diapers                 security@uu.nete   ------------------------------   Date: 4 May 2000 21:20:38 GMTl* From: helbig@astro.rug.nl (Phillip Helbig): Subject: Re: Humble request for a modification to VMS Mail. Message-ID: <8espj6$6bu$2@info.service.rug.nl>  + In article <8esntn$dbn$2@husk.cso.niu.edu>,.& system@niuhep.physics.niu.edu writes:   2 > Michael Austin <maustin@nc.prestige.net> writes: > M > >I usually don't use the message title to tell me what I am looking for.  IsM > >would generally use the search feature to locate documents, especially if ]! > >I don't recall where I put it.v > B > I feel really stupid, how long has search been a part of mail?!?  G For fancier stuff, especially multiple folders, multiple files, /MATCH  + and other goodies, check out the excellent    =    http://www2.wku.edu/scripts/fileserv/fileserv.com?MLSEARCHh     --M Phillip Helbig                       Email .............. helbig@astro.rug.nl M Kapteyn Instituut                    Email ................. helbig@man.ac.ukcM Rijksuniversiteit Groningen          Tel. ................... +31 50 363 4067 M Postbus 800                          Fax .................... +31 50 363 6100fM NL-9700 AV Groningen                 Web ... http://www.astro.rug.nl/~helbig/l  5 My opinions are not necessarily those of my employer.l  N <A HREF=" http://gladia.astro.rug.nl:8000/helbig/hire/hire.html ">HIRE ME!</A>   ------------------------------  $ Date: Thu, 4 May 2000 21:16:57 -0700- From: "Dann Corbit" <dcorbit@solutionsiq.com>b/ Subject: Mailbox and synchronization in OpenVMS ( Message-ID: <0XrQ4.2095$0F5.2416@client>  E If I use a mailbox for IPC, can I force all pending messages to flusht somehow?  B I want to be able to depend upon a previous transaction set having
 completed.  D If not, what is an alternative method for IPC that can be made fully	 reliable?0 --0 C-FAQ: http://www.eskimo.com/~scs/C-faq/top.html$  "The C-FAQ Book" ISBN 0-201-84519-91 C.A.P. Newsgroup   http://www.dejanews.com/~c_a_pmI C.A.P. FAQ: ftp://38.168.214.175/pub/Chess%20Analysis%20Project%20FAQ.htm    ------------------------------  % Date: Fri, 05 May 2000 01:11:15 -0400 0 From: JF Mezei <jfmezei.spamnot@vl.videotron.ca>3 Subject: Re: Mailbox and synchronization in OpenVMS*/ Message-ID: <391257F2.3BC144C4@vl.videotron.ca>    Dann Corbit wrote: > G > If I use a mailbox for IPC, can I force all pending messages to flush 
 > somehow?    M Can can simply read them. If you do not read transactions, they accumulate in H the mailbox up until the mailbox's buffer size, after which, any processI writing to the mailbox goes into the pesky RWMBX state. When your process*I reads transactions, then those processes in RWMBX become unstuck as theiry write operation completes.  F > If not, what is an alternative method for IPC that can be made fully > reliable?_  I DECnet offers a more robust method that also works across nodes, providesuL better buffering, identification of the sender, and ability to have multipleL processes talking to you at the same time and especially the ability of yourN process to respond to a transaction to a specific process. (you have a channelB created with each process so you can direct your output properly).  M There is also, with VMS 7.2, the IPC services which are similar to DECnet but  more efficient.o    L There are other methods such as shared global sections but they require that! you add your own synchronisation.s   ------------------------------  % Date: Thu, 04 May 2000 21:28:30 -0400b0 From: JF Mezei <jfmezei.spamnot@vl.videotron.ca> Subject: Marketing opportunity/ Message-ID: <391223B7.30ADDB07@vl.videotron.ca>%  N This afternoon, I called my mobile phone provider. There was a special message- before anything else. It went something like:r  J "Because a virus has infected most of our computers, our services are veryM limited, and waiting times will be longer than usual, we apoligize in advancer for this situation".  M If i were Compaq, I'd have a BIG full page ad in most newspapers the next daycM showing the advantages of a robust platform that is secure and not the target_ of so many viruses: VMS.   ------------------------------  % Date: Thu, 04 May 2000 22:52:51 -0400=. From: Michael Austin <maustin@nc.prestige.net>" Subject: Re: Marketing opportunity/ Message-ID: <39123783.FCB699DC@nc.prestige.net>u  , This is a multi-part message in MIME format.& --------------4E8078DE932E7D0A0AD41290* Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bitn  K no joke...   They wanted cheap, they got cheap and that's what you get withn3 cheap.  No security, no availability and no uptime.e   Michael AustinP DBA Consultant (currently on contract to one of those Wirless providers that got hit.)    JF Mezei wrote:   P > This afternoon, I called my mobile phone provider. There was a special message/ > before anything else. It went something like:= >oL > "Because a virus has infected most of our computers, our services are veryO > limited, and waiting times will be longer than usual, we apoligize in advance  > for this situation". >.O > If i were Compaq, I'd have a BIG full page ad in most newspapers the next dayxO > showing the advantages of a robust platform that is secure and not the targeti > of so many viruses: VMS.  & --------------4E8078DE932E7D0A0AD41290- Content-Type: text/x-vcard; charset=us-ascii;_  name="maustin.vcf"$ Content-Transfer-Encoding: 7bitx, Content-Description: Card for Michael Austin  Content-Disposition: attachment;  filename="maustin.vcf"    begin:vcard  n:Austin;Michael $ tel;work:704-947-1089b x-mozilla-html:FALSE org:Michael Austin, Inco
 adr:;;;;;; version:2.1t+ email;internet:michaelaustininc@hotmail.com* title:Presidents x-mozilla-cpt:;0 fn:Michael E. Austin	 end:vcard0  ( --------------4E8078DE932E7D0A0AD41290--   ------------------------------    Date: 04 May 2000 15:17:31 -0500- From: Graham Allan <allan@mnhep1.hep.umn.edu>V  Subject: Re: OpenVMS disk space?0 Message-ID: <w53ln1qoysk.fsf@lanark.spa.umn.edu>  , Arne Vajhj <arne.vajhoej@gtech.com> writes:   > Marco Shaw wrote: K > > I have a system with only 300MB of disk space.  Will I be able to put a 2 > > minimum Alpha OpenVMS 7.2 system on this disk? >  > I doubt that.  > > > How have you managed to get an Alpha with only 300 MB disk ? > A > (I thougth that they were always sold with at least 1 GB disks)l  6 166MHz Multia VX40, with hobbiest license, at a guess?A If so, the best thing is probably to forget the internal disk and ! use he external SCSI connector...    G. -- uI -------------------------------------------------------------------------s: Graham Allan - I.T. Manager - gta@umn.edu - (612) 624-50409 School of Physics and Astronomy - University of MinnesotatI -------------------------------------------------------------------------    ------------------------------  % Date: Thu, 04 May 2000 13:58:19 -0400a0 From: JF Mezei <jfmezei.spamnot@vl.videotron.ca> Subject: Re: OpenVMS marketing/ Message-ID: <3911BA39.B2C36249@vl.videotron.ca>    Arne Vajh=F8j wrote:6 > They do not start by a few employees doing some work6 > in their own spare-time. It is a part of the overall- > corporate strategies and treated like that.f    J Canadian Airlines, the first airline to have a web site, did exactly that= =2E One J employee, in his spare time, setup a web site for the airline that was mo= stlyC text based, but provided info such as live arrivals and departures,uJ information about airline etc. He developped it on a cheap machine that w= as at-
 his disposal.-  E When management saw the value, they "took it from him" and gave it tooH marketing which transformed the site into a image heavy (and poorly htmlJ designed) site that looked good on a large screen used to demo to the VPs= =2E,  J I have seen quite a few companies whose web site started off as a pilot o= n axJ cheap machine (linux these days) and once management has been convinced o= f the J potential, it gets included into the corporate environment and grows from=  0 there (so far: always on the original platform).   ------------------------------  % Date: Thu, 04 May 2000 14:06:09 -0400c0 From: JF Mezei <jfmezei.spamnot@vl.videotron.ca> Subject: Re: OpenVMS marketing/ Message-ID: <3911BC0F.9D77748E@vl.videotron.ca>t   "Ramon L. Tate" wrote:M > states, and my 25+ years of experience in the IT biz corroborate this, thathL > e-biz efforts emanate from the top down and are quite commonly planned and! > implemented by consulting firmst  N Yes, now is the case because the "top" realises that if their companies cannotH claim to be e-businesses, they will be sunk by Wall Street. But up until2 recently, the projects started from the bottom up.  L The first comers to the web were the "pilot project" types of companies. TheN late comers are the "quick, I don't care how much it costs and what it can do,F but hire some expensive consulting firsm so that we can claim to be an e-business too".   ------------------------------  $ Date: Thu, 4 May 2000 09:44:46 -0400  From: norm.raphael@jamesbury.com2 Subject: Re: Printing in Landscape - (Please help)4 Message-ID: <C22568D5.004AA96E.00@jklh21.valmet.com>  = If the library is open (some print symbiont has it open) thenI9 $LIBR/REPLACE will not work, so your module may not be ineB the library you are using. If you made a new version, the symbiont may have to old version open. 1 $SHOW DEVICE SYS$SYSDEVICE/FILES should show you.e# ...or it may be some other problem.t        / dgaunavou@hotmail.com on 05/04/2000 11:21:37 AMi   To:   Info-VAX@mvb.saic.com  cc:y/ Subject:  Printing in Landscape - (Please help)u        4 I have tried printing in Landscape but unfortunately  it came up with a error message.& I am having problems with the library./ Please shed some light to this as I am abit new. with the librarian utility.u& I am not sure where I have gone wrong.   Thanks In Advance . Davidt   Error message is as follows:  E Printer queue HEAD$2100, idle, on NODEA::"1.1.1.1:9100", mounted forme HP_LANDSCAPE (stock=A4) 2   <TCP Queue HP Laserjet Printer 2100 - 3rd Floor>?   /BASE_PRIORITY=4 /DEFAULT=(FEED,FORM=HP_LANDSCAPE (stock=A4))S, /LIBRARY=LANDSCAPE Lowercase /OWNER=[SYSTEM]F   /PROCESSOR=UCX$TELNETSYM /PROTECTION=(S:M,O:D,G:R,W:S) /RETAIN=ERROR  4   Entry  Jobname         Username     Blocks  Status4   -----  -------         --------     ------  ------>     524  PRT             SYSTEM           3  Retained on errorC        %PSM-E-MODNOTFND, library module !AS not found in record !ULu&        -LBR-E-KEYNOTFND, key not foundH          Submitted  4-MAY-2000 14:18:41.24 /FORM=HP_LANDSCAPE (stock=A4)
 /PRIORITY=100V'          File: _DKB0:[SYSTEM]PRT.TEST;4u=          Completed  4-MAY-2000 14:18:46.81 on queue HEAD$2100e        H ________________________________________________________________________H Get Your Private, Free E-mail from MSN Hotmail at http://www.hotmail.com   ------------------------------  # Date: Fri, 05 May 2000 00:17:17 GMT.5 From: David Warren <david.warren_deletethis@home.com>i Subject: Re: Reading VMS disks( Message-ID: <3911D7F5.EA00BFA7@home.com>  H If you mean the approximately 6" X 6" double-sided things with the metalN sliding door I have two of those drives connected to VAXen at work.  (I'd haveM to see one to identify it...)  If this is what you have email me and I'll see  what I can do to help you out.  L The other issue is exactly what you are trying to recover.  RMS files aren'tO flat files like unix or WNT, so unless we're talking text files there may be no - way to represent the files other than on VMS.     	 ma wrote:S   > Hello All,C >     I am not a VMS user and I haven't any VMS but I have some VMS.F > isks( optical disks) that I want to read (maybe by a LINUX system) . > ) > Is there any software that can do this?r > N >  what is the latest version of file system on a VMS? I tried to read them byJ > using a free software from OpenVMS site that could read ODS2 files but I, > couldn't read them. Is there any solution? >  > Any help will be appreciated.- >-	 > Regards-   ------------------------------  $ Date: Thu, 4 May 2000 18:03:31 -0400' From: "Bill Todd" <billtodd@foo.mv.com>X% Subject: Re: RMS indexed file cleanup(( Message-ID: <8ess0c$mtb$1@pyrite.mv.net>  D Larry Kilgallen <Kilgallen@eisner.decus.org.nospam> wrote in message% news:ZN7FNyFFa7Ox@eisner.decus.org...-: > In article <39113C46.B32E5CB8@vl.videotron.ca>, JF Mezei) <jfmezei.spamnot@vl.videotron.ca> writes:j
 > > Question:e > >EG > > If an application adds records to an indexed file, then deletes all2 records3; > > from it, and adds more, deletes all of them etc etc ...e > >iG > > In such an application, does the indexed file still require regularhE > > reorganisation, or does the file become "like new" every time alln records have > > been deleted ? > L > If you create your files with Prolog level 3, they require reorganization.? > If you create them with an earlier Prolog level, they do not.s  I That's not correct (at least for Prologue Version 1 files, and I'm prettypL sure for Prologue Version 2):  the deleted records leave deleted record RRVsL around forever in the buckets in which they were created, not only taking upH increasing amounts of space in those buckets (this could be avoided if aK bit-map were used to track in-use IDs within the bucket, but that certainlynJ wasn't the case in PV1 and I don't think was the case in PV2) but using upI bucket ID values as well (there were only 255 possible ID values in a PV1sK bucket, and once they were gone that bucket could never store any more userAI data, no matter how much space was free in it).  PV2 or PV3 increased thetJ number of ID values, but IIRC did not make a fundamental change in the wayL they were managed - for example, with a variable-length ID field up to, say,H 8 bytes in length, IDs would never have to be reused and deleted recordsD would consume no space at all (the only effect would be to graduallyL increase the size of the ID fields in undeleted records and RRVs pointing to$ undeleted records in other buckets).  H Hermann noted that just changing alternate key values gradually consumesH space, since the current-record positioning algorithms maintain positionJ context within duplicate alternate key values by ordinal duplicate number,F making it difficult to reclaim alternate key entries for records whereL alternate key values have changed (though, again, different algorithms couldK improve matters) unless duplicates aren't allowed (CHANGEABLE/NODUPES was an8 combination that RMS-11 didn't support, but RMS-32 may).   - bill   ------------------------------  % Date: Fri, 05 May 2000 06:59:53 +0800u4 From: Dave Sneddon - bigpond <dbsneddon@bigpond.com>% Subject: Re: RMS indexed file cleanup-+ Message-ID: <391200E9.9F8647E6@bigpond.com>6   JF Mezei wrote:t >  > Question:. > M > If an application adds records to an indexed file, then deletes all records89 > from it, and adds more, deletes all of them etc etc ...o > E > In such an application, does the indexed file still require regular:P > reorganisation, or does the file become "like new" every time all records have > been deleted ?  E If you are in fact deleting ALL records then you are best to create anG new file.  I came across an application years ago that did this type of G thing.  Each month it would delete all the records in the file, then dorE the processing for the month.  It used to take about 12 hours to run.cH By changing things so that a new file was created each time, the overall$ run time dropped to about 2.5 hours.   -- - Regards, Dave. I -------------------------------------------------------------------------jI David B Sneddon (dbs)  OpenVMS Systems Programmer   dbsneddon@bigpond.com$I DBS software at ...   http://www.users.bigpond.com/dbsneddon/software.htm:I "Life is what happens to you while you're busy making other plans" Lennon.   ------------------------------  % Date: Thu, 04 May 2000 20:19:09 -0400l0 From: JF Mezei <jfmezei.spamnot@vl.videotron.ca>% Subject: Re: RMS indexed file cleanupn/ Message-ID: <3912137B.71FE38BE@vl.videotron.ca>w   Dave Sneddon - bigpond wrote:.G > If you are in fact deleting ALL records then you are best to create a	I > new file.  I came across an application years ago that did this type ofuI > thing.  Each month it would delete all the records in the file, then do  > the processing for the month.P  E This is more of a queue type of file. Records are deleted as they arelN processed, and hopefully processed shortly after they have been inserted. (but/ must survive a crash, application restart etc).r  L I've had one such thing running for a few years without really any problems,K but volumes are not that great. If high volumes would result in large filesnN (block wise) with few records, then I will probably add a convert/reclaim done( by the application at an opportune time.   ------------------------------  % Date: Thu, 04 May 2000 16:33:55 -0400e! From: JGraf <sales@asoft-dev.com>e Subject: Re: TPU for unix ) Message-ID: <3911DEB3.45EF@asoft-dev.com>t   Scott McNeilly wrote:f > 1 > >From: Michael Austin <maustin@NC.PRESTIGE.NET> R > >Organization: WebUseNet Corp  http://www.usenetserver.com - Home of the fastest) > >              NNTP servers on the Net.l > >Subject: TPU for unix > >iS > >Is there a unix port of EVE/TPU?  I know there is emacs, but I would personnally R > >like to see a real port so when I am on Tru64 or Solaris or HPUX, I can use the$ > >same editor... boy, do I hate vi. > >~ > >Michael Austini > N > We are running a version of TPU on Digital Unix 4.0 and have been for years.L > We bought it from a/Soft Development, 24 Eastman Avenue, Bedford NH 03110. > Tel: 603-666-6699.9 > I don't know whether they are still in business or not.q > @ > --------------------------------------------------------------= > Scott Mc Neilly               email:  smcneilly@bridgew.edu 4 > Assistant Director            Phone:  508-697-12364 > Information Services          FAX:    508-697-1774 > Bridgewater State Collegen > Bridgewater, MA 02325eA > ---------------------------------------------------------------:   Hello Scott,  G a/Soft is still in business and providing nu/TPU.  We have relocated top the following address.   P.O. Box 707 Londonderry, NH 030534 603-432-3388 www.asoft-dev.comc  
 Best regards,-  
 James Graf sales@asoft-dev.come 603-432-3388   ------------------------------  % Date: Thu, 04 May 2000 14:02:29 -0400j0 From: JF Mezei <jfmezei.spamnot@vl.videotron.ca>+ Subject: Re: Web-enabling VMS reporting ...a/ Message-ID: <3911BB34.800BB8A9@vl.videotron.ca>x   Arne Vajh=F8j wrote: > What is the problem ?s > =   	 > If you: % >   - install any HTTP-server for VMS > >   - create a 5 line DCL CGI-script that displays a text-file  J Actually, I beleive that most web servers are fully capable of serving te= xtJ files, as are most clients capable of displaying text/plain mime types. O= SU issJ capable. Just define paths to the directories where the reports are locat= ed4 and then generate an HTML page that provides a menu.  J Alternatively, you can use ALL-IN-1 (office server) to store your reports=  in,J various formats and hav it serve the documents automatically in HTML. See=  the ALL-IN-1 SPD at :a# http://www.digital.com/info/SP6149/e   ------------------------------  # Date: Thu, 04 May 2000 18:18:42 GMTi# From: Mark Sterk <strong@chello.nl>a+ Subject: Re: Web-enabling VMS reporting ...o) Message-ID: <3911BFFD.C269FDF7@chello.nl>    Wayne Yung wrote:   F > I am looking for some web-enabling VMS reporting utilities which canF > convert the standard ASCII VMS reports to web-enabled reports.  Does > anyone have idea about that ?- >-E > REPORT.WEB has claimed itself with such capability, but till now, IM# > cannot get a copy of that to try.> >u > Thank you. >o > Wayne.  3 Maybe not quite you're looking for but it is cheap.h  G Simply write a DCL script that copies the log file to a ...htm file ande' add the following as a header / footer.      <HTML> <HEAD>    <TITLE>Logfile</TITLE>  </HEAD>t <PRE>r   ............  ascii  .......   </PRE> </BODY>  </HTML>     D The <PRE> tag instructs the browser not to mess with the space / tab format of the ascii text.a   Success,   Mark   ------------------------------  % Date: Thu, 04 May 2000 21:18:57 -0400u. From: Michael Austin <maustin@nc.prestige.net>+ Subject: Re: Web-enabling VMS reporting ...h/ Message-ID: <39122181.61B29CDE@nc.prestige.net>t  , This is a multi-part message in MIME format.& --------------BCAA664FE38E712743299A2E, Content-Type: text/plain; charset=iso-8859-1 Content-Transfer-Encoding: 8bit-       Arne Vajhj wrote:   > Wayne Yung wrote: H > > I am looking for some web-enabling VMS reporting utilities which canH > > convert the standard ASCII VMS reports to web-enabled reports.  Does! > > anyone have idea about that ?d >  > What is the problem ?n >B	 > If you:t% >   - install any HTTP-server for VMS > >   - create a 5 line DCL CGI-script that displays a text-file >t > what do you miss ? >e  H Yes, you can be that simple or you can be real creative to be able to doF full text search, ftp,tail,display or the most difficult  page throughJ very large  files.  And it is a lot more difficult than a 5 line DCL pieceJ of code (although the first DCL CGI I wrote 6 years ago did exactly that!)  * Plus, I am trying to make a living here...   Michael Austin DBA Consultant.h   >a > Arne  & --------------BCAA664FE38E712743299A2E- Content-Type: text/x-vcard; charset=us-ascii;-  name="maustin.vcf"E Content-Transfer-Encoding: 7bitx, Content-Description: Card for Michael Austin  Content-Disposition: attachment;  filename="maustin.vcf"C   begin:vcard  n:Austin;Michael p tel;work:704-947-1089f x-mozilla-html:FALSE org:Michael Austin, Incn
 adr:;;;;;; version:2.19+ email;internet:michaelaustininc@hotmail.comu title:President; x-mozilla-cpt:;0 fn:Michael E. Austin	 end:vcard@  ( --------------BCAA664FE38E712743299A2E--   ------------------------------   End of INFO-VAX 2000.250 ************************