0 INFO-VAX	Sat, 08 Jan 2005	Volume 2005 : Issue 15      Contents:0 Anyone have Sybase V10 'cpre' for OpenVMS Alpha?! Re: Basic virtual memory question ! Re: Basic virtual memory question ! Re: Basic virtual memory question ! Re: Basic virtual memory question ! Re: Basic virtual memory question ! Re: Basic virtual memory question ! Re: Basic virtual memory question ! Re: Basic virtual memory question ! Re: Basic virtual memory question ! Re: Basic virtual memory question ! Re: Basic virtual memory question ! Re: Basic virtual memory question ! Re: Basic virtual memory question ! Re: Basic virtual memory question ! Re: Basic virtual memory question ! Re: Basic virtual memory question = Re: CSWS 2.0 + MultiNet 5.0 - Problem Displaying Large Images 5 Re: Help - SYSUAF and $SETUAI on UAI$_USER_DATA field 5 Re: Help - SYSUAF and $SETUAI on UAI$_USER_DATA field  Here is $50 for signing up free * Re: How to have SEARCH match a "null" line Re: HP/Linux Tivo replacement  Re: HP/Linux Tivo replacement  Re: HP/Linux Tivo replacement 8 Re: Mark Gorham interview in Interex's HP World Magazine8 RE: Mark Gorham interview in Interex's HP World Magazine8 Re: Mark Gorham interview in Interex's HP World Magazine OT: Laptop hard disk upgrade) Q: How to have SEARCH match a "null" line - Re: Q: How to have SEARCH match a "null" line - Re: Q: How to have SEARCH match a "null" line  Re: Samsung! Re: Samsung! Re: Simple EDT or TPU init file  Re: Simple EDT or TPU init file  Re: Simple EDT or TPU init file  Re: Simple EDT or TPU init file  Sun shocka - take 9  Re: Sun shocka - take 9  Re: VMS ready laser printer? Re: VMS ready laser printer? Re: VMS ready laser printer? Re: VMS ready laser printer?  F ----------------------------------------------------------------------  % Date: Fri, 07 Jan 2005 21:49:47 -0000 1 From: wspencer@ap.dontspamme.org (Warren Spencer) 9 Subject: Anyone have Sybase V10 'cpre' for OpenVMS Alpha? 1 Message-ID: <95D7AF11Cwspenceraporg@216.168.3.30>   	 Hi Folks,   A I'm in historical mode here, trying to find a copy of Sybase's C  J precompiler, of Version 10 vintage, for OpenVMS Alpha.  We have the other E client-side files installed (dblib, etc), but for some reason didn't  L install the precompiler way back when.  Sadly, we cannot find a copy of the L original media, and Sybase Corp themselves seems to be struggling with this  request.  M Soooo, would anyone have a copy of the "cpre" executable for this platform?   K If so, please contact me at the enclosed address (minus the ".dontspamme")  + or respond to this newgroup.  Many thanks !    ws   --   Warren Spencer Senior Software Engineer The Associated Press   ------------------------------   Date: 7 Jan 2005 13:22:47 -0600 ; From: koehler@eisner.nospam.encompasserve.org (Bob Koehler) * Subject: Re: Basic virtual memory question3 Message-ID: <9uvxS3QTbl4c@eisner.encompasserve.org>   r In article <1105122665.245feb03fda7a96ead9c9055ba54b402@teranews>, JF Mezei <jfmezei.spamnot@teksavvy.com> writes:  F    Generally "memory" _is_ RAM, _not_ disk space, in the discussion ofD    virtual vs. physical memory.  This sometimes causes confusion forF    folks used to thinking of disk storage as "memory", confusion whichC    increases with respect to virtual memory OS which use disk space 1    to allow large virtual memory to be allocated.   O > OK, I browsed through the VAX/VMS internals book. But I couldn't find an easy 6 > answer to an age long question I have had about VMS. > L > On a conventional MAC, one had a very simple memory management config: YouL > told the MAC how much virtual memory it should have, and it would create aO > page file that big. Then, with each imager file (application), you configured 2 > how much virtual memory it was allowed to take.  > M > On VMS, the way I understand it, a process can allocate memory until it has P > used up all its pagefile quota and its working can't be grown anymore. Is that > correct ?   G    Per-process limits are in the UAF.  A process starts each image with G    WSDEF.  It can demand up to WSQUOTA.  Under normal conditions it can D    borrow up to WSEXT.  Page file allocation is limitted by the pageB    file quota,  WSQUOTA limits the working set.  There are lots ofB    things that can be in th working set that will never need to beD    paged out and therefor don't need page file quota, like read-only$    pages from on disk image files.    G    When the UAF is not appropriate, per-process limits come from SYSGEN     PQL parameters.  O > At the VMS level, are there SYSGEN parameters to tell the OS that it has over L > 100 megs of virtual address space ? (100 megs page file + however physical& > memory can be used as VIRTUALPAGES).  G    At the SYSGEN level things are limited by VIRTUALPAGECNT and WSMAX.  C    IIRC there are two parameters because some things count against  C    VIRTUALPAGECNT that are not a part of the working set and do not C    count against WSMAX.  (I don't have my internals manual handy, I '    might be remembering that backward.)   N > Or does VMS create a virtual memory page table for each process based on how@ > much pagfilquo and wsextent the process can potentially have ?  I    IIRC the process page table is sized at process creation according to      VIRTUALPAGECNT.   > M > From what I can gather, the SYSGEN parameters seem to split physical memory N > between non pageable memory and pageable memory (NPAGEDYN and VIRTUALPAGES),O > with soem other parameters probably taking up some physical memory apart from 5 > NPAGEDYN and VIRTUALPAGES (such as SPTREQ I think).   G    PAGEDYN and NPAGEDYN limit kernel memory pools, not process memory.  G    Lots of kernel and per-process things take up memory and are limited &    via a variety of SYSGEN parameters.  O > Is there a simple list of SYSGEN parameters that, when added up, equal to the # > physical memory on that machine ?   F    No.  There is a parameter PHYSICALPAGES which can be set to preventJ    VMS from mapping pages that look like RAM but are not actually memory, F    but is usually set to -1 to allow VMS to map everything that looks G    like RAM.  Some systems have this set to the actual size of RAM but  G    that's just a problem waiting to happen when the users install more      RAM and can't use it.  D    It is possible, and generally a good idea, to allow the total sumF    of all the virtual memory that all processes can allocate to exceedF    the total size of the pagefile, simply because much data never getsF    paged out so there is no need to waste disk space (requires VMS 3.0
    or later).   B    No process or combination of processes should be limited to the>    physical RAM size of the computer in any virtual memory OS.   ------------------------------   Date: 7 Jan 2005 13:52:30 -0600  From: briggs@encompasserve.org* Subject: Re: Basic virtual memory question3 Message-ID: <NwONKY0KaM2L@eisner.encompasserve.org>   r In article <1105122665.245feb03fda7a96ead9c9055ba54b402@teranews>, JF Mezei <jfmezei.spamnot@teksavvy.com> writes:O > OK, I browsed through the VAX/VMS internals book. But I couldn't find an easy 6 > answer to an age long question I have had about VMS. > L > On a conventional MAC, one had a very simple memory management config: YouL > told the MAC how much virtual memory it should have, and it would create aO > page file that big. Then, with each imager file (application), you configured 2 > how much virtual memory it was allowed to take.  > M > On VMS, the way I understand it, a process can allocate memory until it has P > used up all its pagefile quota and its working can't be grown anymore. Is that > correct ?   > Pretty much.  It can also have virtual memory corresponding to> image sections (read-only copies of program data backed by the> original image file), memory mapped files and global sections.  ; You could have hundreds of user processes sharing access to A hundreds of pages of shareable image sections.  This would amount > to tens of thousands of pages of virtual memory and only a few= hundreds of pages of physical memory and no page file at all.   H > Lets take the example of a *theoretical* all mighty Microvax II with aL > whopping 16 megabytes of physical RAM, but with a 100 megabytes page file. > O > At the VMS level, are there SYSGEN parameters to tell the OS that it has over L > 100 megs of virtual address space ? (100 megs page file + however physical& > memory can be used as VIRTUALPAGES).  B On VAX you used to have VIRTUALPAGECNT, but that was a per-processE limit (that affected the size of the portion of the balance set slots C allocated for P0 and P1 process page tables).  The balance set was, . in turn, mapped in system virtual memory (S0).  B On VAX, total virtual memory was not constrained by the page file.? It was constrained by VIRTUALPAGECNT*MAXPROCESSCNT and, to some ? extent, by sum(process=1 to MAXPROCESSCNT)(PGFLQUOTA(process)))   H If, at run time, you had all of your processes actually try to use theirD virtual memory allotment, you could run into what I referred to as aF "page file hang" with an empty free page list and bunches of processes waiting in COMO, FPW or RWMPB.   Rarely a fun experience.  N > Or does VMS create a virtual memory page table for each process based on how@ > much pagfilquo and wsextent the process can potentially have ?  E My internals knowledge is very dated at this point.  The process page C table on VAX was sized based on VIRTUALPAGECNT, completely ignoring  PGFLQUOTA and WSEXTENT.   D I believe that the virtual memory situation modern VMS systems underH Alpha is even more dynamic.  Process page table sizes are not statically knowable at all.  M > From what I can gather, the SYSGEN parameters seem to split physical memory N > between non pageable memory and pageable memory (NPAGEDYN and VIRTUALPAGES),  L NPAGEDYN controls the size of a particular pool of _dynamically allocatable_ non-paged system memory.   $ mcr sysgen SYSGEN>  SHOW VIRTUALPAGES$ %SYSGEN-E-NOPARAM, no such parameter  E I have always assumed that whatever physical memory is not devoted to E something else will be placed on the free list and made available for C virtual memory support.  There is no need for a SYSGEN parameter to ) control how many pages are used for this.   ; The system can dynamically allocate pageable system memory, = devote it to a loaded object and lock it into physical memory 2 if needed.  So can a user process for that matter.  O > with soem other parameters probably taking up some physical memory apart from 5 > NPAGEDYN and VIRTUALPAGES (such as SPTREQ I think).   F Parameters control the sizes of data structures in some cases.  It is D the data structures that occupy physical memory, not the parameters.A Not all physical memory is consumed by data structures with sizes   controlled by SYSGEN parameters.  O > Is there a simple list of SYSGEN parameters that, when added up, equal to the # > physical memory on that machine ?    Not as far as I know.    	John Briggs   ------------------------------  % Date: Fri, 07 Jan 2005 15:46:01 -0500 - From: JF Mezei <jfmezei.spamnot@teksavvy.com> * Subject: Re: Basic virtual memory questionB Message-ID: <1105130148.f8b79f6159b319c32dfc5320f50ffd8c@teranews>   Bob Koehler wrote:  H >    At the SYSGEN level things are limited by VIRTUALPAGECNT and WSMAX.  O SYSGEN's help tells me that VirtualPageCnt has been obsolete since version 7.0.   H What prevents me from giving a 200 meg PGFILQUO to a username and have aM process created that then does a MALLOC of 200 megabytes on a 16 meg Microvax  II ?  J >    IIRC the process page table is sized at process creation according to >    VIRTUALPAGECNT.    1 Ok, so each process has its own page table then ?   K How does VMS size that page table ? (seems it used to be VIRTUALPAGCNT that - answered that question, but is now obsolete).   H >    PAGEDYN and NPAGEDYN limit kernel memory pools, not process memory.  M Howeber, any physical memory that is alloacted to system structures is memory N that cannot be used by processes to map their virtual pages to physical pages,	 correct ?   H >    No.  There is a parameter PHYSICALPAGES which can be set to preventK >    VMS from mapping pages that look like RAM but are not actually memory,   J I thought PHYSICAL PAGES was to simulate a system with less memory that it really has.   D >    No process or combination of processes should be limited to the@ >    physical RAM size of the computer in any virtual memory OS.  L yes, but what determines the virtual memory limit a process can think it hasH access to ? If I give it a PAGFILQUO of 200 megs, (and there is 200 megsN available in the page file), is there another parameter that woudl prevent the procesxs from using 200 megs ?   ------------------------------  # Date: Fri, 07 Jan 2005 20:53:09 GMT % From: Rob Brown <mylastname@gmcl.com> * Subject: Re: Basic virtual memory questionE Message-ID: <Pine.LNX.4.61.0501071344350.17337@localhost.localdomain>   # On Fri, 7 Jan 2005, JF Mezei wrote:   D > What prevents me from giving a 200 meg PGFILQUO to a username and G > have a process created that then does a MALLOC of 200 megabytes on a   > 16 meg Microvax II ?  B *I* don't think anything does.  Nor can I think of any reason why G anything should.  My PGFILQUO on this 64Mb Vax 4000-90 is over 500,000   pages.  + I look forward to learning the real answer.      --    B Rob Brown                        b r o w n a t g m c l d o t c o mA G. Michaels Consulting Ltd.      (866)438-2101 (voice) toll free! 6 Edmonton                         (780)438-9343 (voice)5                                   (780)437-3367 (FAX) 2                                   http://gmcl.com/   ------------------------------   Date: 7 Jan 2005 15:55:36 -0600 ; From: koehler@eisner.nospam.encompasserve.org (Bob Koehler) * Subject: Re: Basic virtual memory question3 Message-ID: <SuimqAjfa5OV@eisner.encompasserve.org>   r In article <1105130148.f8b79f6159b319c32dfc5320f50ffd8c@teranews>, JF Mezei <jfmezei.spamnot@teksavvy.com> writes: > Bob Koehler wrote: > I >>    At the SYSGEN level things are limited by VIRTUALPAGECNT and WSMAX.  > Q > SYSGEN's help tells me that VirtualPageCnt has been obsolete since version 7.0.  > J > What prevents me from giving a 200 meg PGFILQUO to a username and have aO > process created that then does a MALLOC of 200 megabytes on a 16 meg Microvax  > II ?  F    MALLOC'ed pages are in the working set and therefor limitted by the    working set parameters.   > 3 > Ok, so each process has its own page table then ?   D    Yes, in the process header in P1 space up to (I think) VMS 7.1.  D    Now I think Alphas (and IA64s) can keep it in S2 space, but it's &    still a per-process data structure.  M > How does VMS size that page table ? (seems it used to be VIRTUALPAGCNT that / > answered that question, but is now obsolete).   F    Since my internals manual is pre VMS 7.0, I don't know that one.  IE    think since VIRTUALPAGCNT was eliminated there is some capacity to F    resize the map.  I think the map sizes are derived from the working
    set sizes.   I >>    PAGEDYN and NPAGEDYN limit kernel memory pools, not process memory.  > O > Howeber, any physical memory that is alloacted to system structures is memory P > that cannot be used by processes to map their virtual pages to physical pages, > correct ?   H    Processes do map system structures, they have to since they can call M    routines in the kernel and those routines must be able to manipulate that  F    data.  (A lot of those routines change to privileged mode but very E    few execute outside of process context, and even then some process I    is current so the system space map is valid).  Actually all processes  J    map system space via a shared map.  (The map is multi-level, the lower 9    levels of system space maps only actually exist once).   I >>    No.  There is a parameter PHYSICALPAGES which can be set to prevent L >>    VMS from mapping pages that look like RAM but are not actually memory, > L > I thought PHYSICAL PAGES was to simulate a system with less memory that it
 > really has.   E    That's another valid use.  But it used to be used to keep VMS from >    mapping UNIBUS addresses on VAX-11 series systems which hadE    peripherals that looked like RAM (VMS had to be able to map UNIBUS B    RAM in order to support VAX 11/730 and 725).  I think some QbusC    systems had the same issue.  Other systems generally had RAM and #    peripherals on different busses.   E >>    No process or combination of processes should be limited to the A >>    physical RAM size of the computer in any virtual memory OS.  > N > yes, but what determines the virtual memory limit a process can think it hasJ > access to ? If I give it a PAGFILQUO of 200 megs, (and there is 200 megsP > available in the page file), is there another parameter that woudl prevent the  > procesxs from using 200 megs ?  F    You'ld have to give it the required working set quota and make sureG    WSMAX was at least as big.  Then the process could map 200MB even if D    the pagefile free space was less than 200MB.  You can over-commitB    page files.  Just never let the pagefile run out of free space D    (i.e. actually page out too much).  For example, I routinely keepG    my page files about 25% free and use the negative "reservable" value H    as an estimate of how much to grow them when they get below 25% free.  C    By comparison, we routinely ran 3MB programs on a 1MB VAX 11/780 D    and 20MB programs on 4MB VAXen.  VMS is capable of running 200MB C    program on 16MB VAXen (I probably had some like that when I was  )    developing X11 based user interfaces).    ------------------------------  % Date: Fri, 07 Jan 2005 17:51:49 -0500 - From: JF Mezei <jfmezei.spamnot@teksavvy.com> * Subject: Re: Basic virtual memory questionB Message-ID: <1105137689.bf5615984a3cd0a563676c2236dd78d4@teranews>   John Laird wrote: S >> >SYSGEN's help tells me that VirtualPageCnt has been obsolete since version 7.0.  > J > You should have read on a bit further, where it says "on Alpha systems".B > Since your MVII is very much not an Alpha, it is still relevant.   SYSGEN> HELP SYS VIRTUALPAGECNT   < On VAX systems prior to Verison 7.0, VIRTUALPAGECNT sets theG maximum number of virtual pages that can be mapped for any one process.    [much stuff]  L Starting with O___VMS Alpha 7.0, VIRTUALPAGECNT is an obsolete parameter....   -------   A Nothing explains what VIRTUALPAGECNT does on VAX systems post 7.0    ------------------------------  % Date: Fri, 07 Jan 2005 18:03:47 -0500 - From: JF Mezei <jfmezei.spamnot@teksavvy.com> * Subject: Re: Basic virtual memory questionB Message-ID: <1105138402.43854f4997ae3b0878c9fce3c4b2e0e1@teranews>   Bob Koehler wrote:H >    You'ld have to give it the required working set quota and make sureI >    WSMAX was at least as big.  Then the process could map 200MB even if 2 >    the pagefile free space was less than 200MB.   J But isn't WSMAX/WSQUOTA/WSEXTENT a limit of how much virtual memory can be0 mapped to physical memory at any point in time ?  I If you malloc a chunk of memory and wrote to it, and then need some other N piece, nothing prevents that chunk from being written to the page file to make# room for what you now need, right ?     C >    page files.  Just never let the pagefile run out of free space   N Yes, those dreaded one line  "page file fragmented" messages on OPA0 that giveH you a few seconds to start killing processes to free up memory... But an, enourmous page file prevents those problems.   ------------------------------  % Date: Fri, 07 Jan 2005 23:33:55 +0000 - From: John Laird <nospam@laird-towers.org.uk> * Subject: Re: Basic virtual memory question8 Message-ID: <au6ut0dnlrfdmgg1hdmkp7ke6fgugkuuq0@4ax.com>  K On Fri, 07 Jan 2005 17:51:49 -0500, JF Mezei <jfmezei.spamnot@teksavvy.com>  wrote:   >John Laird wrote:T >>> >SYSGEN's help tells me that VirtualPageCnt has been obsolete since version 7.0. >>  K >> You should have read on a bit further, where it says "on Alpha systems". C >> Since your MVII is very much not an Alpha, it is still relevant.  >   >SYSGEN> HELP SYS VIRTUALPAGECNT > = >On VAX systems prior to Verison 7.0, VIRTUALPAGECNT sets the H >maximum number of virtual pages that can be mapped for any one process. > 
 >[much stuff]  > M >Starting with O___VMS Alpha 7.0, VIRTUALPAGECNT is an obsolete parameter....  >  >------- > B >Nothing explains what VIRTUALPAGECNT does on VAX systems post 7.0  J Nothing much different to what it did pre-7.0 would be my best guess.  The+ documentation you are reading is confusing.    On our 7.3-2 systems, it reads:   > "Starting with OpenVMS Version 7.0, VIRTUALPAGECNT has been an'   obsolete parameter on Alpha systems."    Is that clearer ?    --  8 Remember that you are special, just like everyone else.    Mail john rather than nospam...    ------------------------------  # Date: Fri, 07 Jan 2005 23:58:15 GMT & From: John Reagan <john.reagan@hp.com>* Subject: Re: Basic virtual memory question1 Message-ID: <rmFDd.5367$Iz1.578@news.cpqcorp.net>    JF Mezei wrote:   Q > SYSGEN's help tells me that VirtualPageCnt has been obsolete since version 7.0.  > J > What prevents me from giving a 200 meg PGFILQUO to a username and have aO > process created that then does a MALLOC of 200 megabytes on a 16 meg Microvax  > II ?   Intelligence. :-)   F As long as you have a large enough page file and a good book to read, K you can listen to your system page its brains out.  It should work however.      --   John Reagan / HP Pascal/{A|I}MACRO for OpenVMS Project Leader  Hewlett-Packard Company    ------------------------------  # Date: Fri, 07 Jan 2005 23:59:43 GMT & From: John Reagan <john.reagan@hp.com>* Subject: Re: Basic virtual memory question1 Message-ID: <PnFDd.5368$Iz1.983@news.cpqcorp.net>    JF Mezei wrote:  > Bob Koehler wrote: > H >>   You'ld have to give it the required working set quota and make sureI >>   WSMAX was at least as big.  Then the process could map 200MB even if 2 >>   the pagefile free space was less than 200MB.  >  > L > But isn't WSMAX/WSQUOTA/WSEXTENT a limit of how much virtual memory can be2 > mapped to physical memory at any point in time ?  G Yes, but that doesn't have anything to do with how much virtual memory   you can actually allocate.     --   John Reagan / HP Pascal/{A|I}MACRO for OpenVMS Project Leader  Hewlett-Packard Company    ------------------------------  # Date: Fri, 07 Jan 2005 23:53:23 GMT # From: Beach Runner <bob@nospam.com> * Subject: Re: Basic virtual memory question; Message-ID: <ThFDd.183344$8G4.9066@tornado.tampabay.rr.com>   C The size of page file quota has a limit in physical space. I don't  E remember the exact size at this second.  Virtual page count has been   made obsolete.  I If Working sets grow larger than page file quota, the process will crash.  This is common in sorts.   JF Mezei wrote:  > Bob Koehler wrote: >  > H >>   At the SYSGEN level things are limited by VIRTUALPAGECNT and WSMAX. >  > Q > SYSGEN's help tells me that VirtualPageCnt has been obsolete since version 7.0.  > J > What prevents me from giving a 200 meg PGFILQUO to a username and have aO > process created that then does a MALLOC of 200 megabytes on a 16 meg Microvax  > II ? >  > J >>   IIRC the process page table is sized at process creation according to >>   VIRTUALPAGECNT. >  >  > 3 > Ok, so each process has its own page table then ?  > M > How does VMS size that page table ? (seems it used to be VIRTUALPAGCNT that / > answered that question, but is now obsolete).  >  > H >>   PAGEDYN and NPAGEDYN limit kernel memory pools, not process memory. >  > O > Howeber, any physical memory that is alloacted to system structures is memory P > that cannot be used by processes to map their virtual pages to physical pages, > correct ?  >  > H >>   No.  There is a parameter PHYSICALPAGES which can be set to preventK >>   VMS from mapping pages that look like RAM but are not actually memory,  >  > L > I thought PHYSICAL PAGES was to simulate a system with less memory that it
 > really has.  >  > D >>   No process or combination of processes should be limited to the@ >>   physical RAM size of the computer in any virtual memory OS. >  > N > yes, but what determines the virtual memory limit a process can think it hasJ > access to ? If I give it a PAGFILQUO of 200 megs, (and there is 200 megsP > available in the page file), is there another parameter that woudl prevent the  > procesxs from using 200 megs ?   ------------------------------  # Date: Sat, 08 Jan 2005 00:27:56 GMT & From: John Reagan <john.reagan@hp.com>* Subject: Re: Basic virtual memory question2 Message-ID: <gOFDd.5375$TF1.4578@news.cpqcorp.net>   Beach Runner wrote: E > The size of page file quota has a limit in physical space. I don't  G > remember the exact size at this second.  Virtual page count has been   > made obsolete. > K > If Working sets grow larger than page file quota, the process will crash.  > This is common in sorts. >   H Not working set, virtual memory limit.  Your working set quota and page ! file quota aren't really related.   G If you exceed your page file quota, I wouldn't say "crash".  You might  I get a PGFILQUO exceeded message or you might just get an ACCVIO when you  H try to allocate more stack but you don't have enough page file quota to E back it up.  If you can make a process "crash" (ie, disappear), then  * you've found a bug that you should report.     --   John Reagan / HP Pascal/{A|I}MACRO for OpenVMS Project Leader  Hewlett-Packard Company    ------------------------------  $ Date: Fri, 7 Jan 2005 20:14:00 -0500) From: "Neil Rieck" <n.rieck@sympatico.ca> * Subject: Re: Basic virtual memory question; Message-ID: <GtGDd.31639$7n1.1279148@news20.bellglobal.com>   : "JF Mezei" <jfmezei.spamnot@teksavvy.com> wrote in message< news:1105122665.245feb03fda7a96ead9c9055ba54b402@teranews...J > OK, I browsed through the VAX/VMS internals book. But I couldn't find an > easy6 > answer to an age long question I have had about VMS. > L > On a conventional MAC, one had a very simple memory management config: YouL > told the MAC how much virtual memory it should have, and it would create aD > page file that big. Then, with each imager file (application), you > configured1 > how much virtual memory it was allowed to take.  > I > On VMS, the way I understand it, a process can allocate memory until it  > has K > used up all its pagefile quota and its working can't be grown anymore. Is  > that > correct ?  > H > Lets take the example of a *theoretical* all mighty Microvax II with aL > whopping 16 megabytes of physical RAM, but with a 100 megabytes page file. > J > At the VMS level, are there SYSGEN parameters to tell the OS that it has > overL > 100 megs of virtual address space ? (100 megs page file + however physical& > memory can be used as VIRTUALPAGES). > J > Or does VMS create a virtual memory page table for each process based on > how @ > much pagfilquo and wsextent the process can potentially have ? >  > F > From what I can gather, the SYSGEN parameters seem to split physical > memory? > between non pageable memory and pageable memory (NPAGEDYN and  > VIRTUALPAGES),J > with soem other parameters probably taking up some physical memory apart > from5 > NPAGEDYN and VIRTUALPAGES (such as SPTREQ I think).  > K > Is there a simple list of SYSGEN parameters that, when added up, equal to  > the # > physical memory on that machine ?    An overly simple response:  ? 1. IMHO, the two most important sysgen parameters are WSMAX and K MAXPROCESSCNT. They (and others) are used by AUTOGEN in setting the size of F the PAGEFILE and SWAPFILE. You can set WSDEF and WSQUO to any size youI want in SYSUAF but WSMAX places the final upper limit on what any process 
 can ever get.   J (p.s. In the old VAX-11/750 days AUTOGEN was too liberal in setting valuesL of WSMAX so I would override these values by making entries in MODPARAMS.DATL before invoking AUTOGEN. This was always a good way to squeeze 10 more entryJ clerks onto our system but wouldn't have worked if these folk were running large apps like compilers)  I 2. When a process is started, it is initially allowed WSDEF pages and can I demand more by faulting up to WSQUO (this amount is guaranteed if AUTOGEN F was done properly "and" no one was making manual changes to the SYSGENK params). If the system is lightly loaded a process may be allowed to borrow & more pages by faulting up to WSEXTENT.  I 3. How does OpenVMS processes know when to allow borrowing? Free pages in L OpenVMS are maintained in a structure known as the FREELIST. When the systemJ has more than GROWLIM pages, active "faulting" processes (at a rate higherL than PFRATH) can acquire more pages up to a maximum of WSEXTENT. On the flipL side, when the system gets busy and the FREELIST shrinks lower than FREELIM,K the system will start to trim back all processes (extended processes first, G then idle and low faulting processes (lower than PFRATL) next. This can F continue and may even involving trimming dormant (idle for longer thanI DORMANTWAIT) processes back to their DEFAULTS. After this the system will I resort to swapping out the process entirely.The system will continue this # until the FREELIST size = FREEGOAL.   $     $mon sys (to watch the freelist)  H BTW, GROWLIM and BORROWLIM work together to enable or disable borrowing.K The borrow feature is first enabled when the system has more than BORROWLIM D pages and is disabled when the system has fewer than GROWLIM (IIRC).I Together they provide a form of hysterisis which prevents the system from  acting erratically.   I 4. Program code and read-only information usually resides on the FREELIST K while program variables reside on a smaller structure known as the MODIFIED L LIST. This so-called dirty page data will need to be written to the PAGEFILE0 before these pages are given to another process.  F 5. When a processes needs more pages but is denied, it can get more byI giving up some existing pages back to the free list. If it needs the page F later and the desired page is on the FREELIST, it can get it back by aJ process of "soft faulting". However, if the page has been given to someone? else, the process will need to "hard fault" the data from disk.   L 6. Busy systems may want to enable Automatic Working Set (AWS) adjustment inI both directions (expansion and contraction). The value of PFRATL has been K defaulted to zero for many years now but can be enabled to ensure that more C free memory is usually available (at the expense of more faulting).   C 7. You should really read a tuning-guide before you mess with these K parameters. That said, if you must hack then check out the following online  help:    $mcr sysgen  SYSGEN>help sys freelim      sys freegoal     sys growlim      sys borrowlim      sys freelim      sys awstim     sys awsmin
     sys wsinc 
     sys wsdec      sys pfratl     sys pfrath  J That's all for now. Most of this info came from a Raxcom VMS tuning courseI more than 10 years ago but is still valid. If I've made any mistakes it's 4 because I was drinking a Heineken while typing this.  
 Neil Rieck Kitchener/Waterloo/Cambridge,  Ontario, Canada.8 http://www3.sympatico.ca/n.rieck/links/cool_openvms.html   ------------------------------    Date: 07 Jan 2005 20:36:05 -05003 From: Rich Alderson <news@alderson.users.panix.com> * Subject: Re: Basic virtual memory question. Message-ID: <mddhdlsr9bu.fsf@panix5.panix.com>  / JF Mezei <jfmezei.spamnot@teksavvy.com> writes:   J > What prevents me from giving a 200 meg PGFILQUO to a username and have aO > process created that then does a MALLOC of 200 megabytes on a 16 meg Microvax  > II ?   Only your innate good taste...   --  L Rich Alderson                                       | /"\ ASCII ribbon     |L news@alderson.users.panix.com                       | \ / campaign against |L "You get what anybody gets. You get a lifetime."    |  x  HTML mail and    |L                          --Death, of the Endless    | / \ postings         |   ------------------------------  % Date: Fri, 07 Jan 2005 22:20:41 -0500 ' From: Dave Froble <davef@tsoft-inc.com> * Subject: Re: Basic virtual memory question, Message-ID: <41DF5189.9060003@tsoft-inc.com>   John Reagan wrote:   > JF Mezei wrote:  > F >> SYSGEN's help tells me that VirtualPageCnt has been obsolete since  >> version 7.0.  >>K >> What prevents me from giving a 200 meg PGFILQUO to a username and have aVH >> process created that then does a MALLOC of 200 megabytes on a 16 meg  >> Microvaxe >> II ?n >  >  > Intelligence. :-)  > H > As long as you have a large enough page file and a good book to read, M > you can listen to your system page its brains out.  It should work however.o >  >   O Well, it's winter time up North.  Maybe JF's looking for a way to get the disk a' drives to produce a bit more heat.  :-)o   ------------------------------  $ Date: Sat, 8 Jan 2005 00:32:54 -0500: From: "Hein RMS van den Heuvel" <hein@Hewlett-Packard.com>* Subject: Re: Basic virtual memory question* Message-ID: <41df70b2@usenet01.boi.hp.com>  @ "Rich Alderson" <news@alderson.users.panix.com> wrote in message( news:mddhdlsr9bu.fsf@panix5.panix.com...1 > JF Mezei <jfmezei.spamnot@teksavvy.com> writes:  >eL > > What prevents me from giving a 200 meg PGFILQUO to a username and have aH > > process created that then does a MALLOC of 200 megabytes on a 16 meg Microvax > > II ? >x  > Only your innate good taste...   LOLl   John Reagan wrote: "Intelligence. :-)E As long as you have a large enough page file and a good book to read,lL you can listen to your system page its brains out.  It should work however."   LOL   G But... In the replies I have read on the topic I seem to be missing theeI distinction between virtual memory address space and instantiated memory. L That MALLOC might perform reasonably depending on implemenation details. Now/ if you had suggest CALLOC, now that would hurt.l  H I mean, it could just be just fine application design. You know, say 200G byte member records for a town band with 100 members stored in a simple K array indexed by their 7 digit local phone number. So you are only going to-I touch 100 pages or so, but you need space for 200MB. Perfectly fine! :-).5G Ok, I jest some, but the is very powerful to not to have to worry aboutjC storing dense data, but to be able to be very sparse when and where9 appropriate.   Hein.1   ------------------------------   Date: 7 Jan 2005 21:20:07 -0800f) From: "Rich Faust" <faustrj@fastspot.net> F Subject: Re: CSWS 2.0 + MultiNet 5.0 - Problem Displaying Large ImagesC Message-ID: <1105161607.318914.156700@z14g2000cwz.googlegroups.com>h  F I installed TCPIP V5.4-15 on my system and SWS V2.0 displays images ofE all sizes with no difficulty.  The large image files that were giving G me problems have fixed length 512 byte records.  I didn't change any oftE my config files before or after installing TCPIP Services for OpenVMSuD so I can only conclude the problem has something to do with MultiNet  V5.0 or how I had it configured.
 Rich Faust OpenVMS Hobbyist Richmond, TX   ------------------------------  % Date: Fri, 07 Jan 2005 11:20:20 -0800 , From: Ken Fairfield <my.full.name@intel.com>> Subject: Re: Help - SYSUAF and $SETUAI on UAI$_USER_DATA field+ Message-ID: <crmndk$alb$1@news01.intel.com>    briggs@encompasserve.org wrote:ae > In article <fc7rt0llet4t6vcirn2ibdemv3qessnadn@4ax.com>, Lorin Ricker <lorin@locktrack.com> writes:  > L >>Re: opening SYSUAF in Pascal (or Fortran, etc) -- my utility does not openN >>SYSUAF directly at all -- SETUAI doesn't have a single OPEN statement in it;Q >>open/closing SYSUAF is handled by the $GET/SETUAI interface.  My question aboutuF >>logicals was "Do these RTLs open by logical name, or by using a full >>path/filespec?"g >>P >> I suspected that they'd honor the "SYSUAF" logical name, and John's advice toP >>"Define your SYSUAF logical and use AUTHORIZE to see if you can read/write it.M >>It uses the same interfaces." reminds me that this is true, which is easilyhA >>verified.  When flailing, I don't always think to clearly!  ;-)m >  > @ > Testing reveals that AUTHORIZE does NOT use the same interface > as $GET/SETUAI.  > F > The AUTHORIZE utility respects process and outer mode logical names.B > The $GET/SETUAI routines respect only trustworthy logical names. > K > That is to say, AUTHORIZE behaves like a Pascal (or Fortran, etc) programoE > opening SYSUAF with a default file specification of SYS$DISK:[].DATU > ? > But $GET/SETUAI behaves like SYS$SYSTEM:LOGINOUT.EXE, openingoA > SYSUAF with a default file specification of SYS$SYSTEM:.DAT andw. > specifying the use of trusted logical names.  C I am a bit amazed at the misinformation being spread here by peopleeB I normally consider to be quite experienced and knowlegable... :-(  ? $GETUAIU/$SETUAI _will_ respect Process or other logical names.sD However, those logicals _must_ be define /EXECUTIVE, that's the key.  > I won't reproduce here a complete demo, but we have an old-ish> program (which happens to be written in Fortran but that's not> relevant) which calls $GETUAI to return various items from the? SYSUAF & RIGHTSLIST.  It happily pulls data from the non-systemIA "master" SYSUAF & RIGHTSLIST if those logicals are defined /EXEC. @ It fails to do so if the same logicals are define wihtout /EXEC.       Regards, Ken --  6 I don't speak for Intel, Intel doesn't speak for me...  
 Ken Fairfieldt! D1C Automation VMS System Supporte" who:   kenneth dot h dot fairfield where: intel dot com   ------------------------------  % Date: Fri, 07 Jan 2005 12:06:59 -0800R From: JBloggs@acme.com> Subject: Re: Help - SYSUAF and $SETUAI on UAI$_USER_DATA field8 Message-ID: <5rqtt017qs4ul7bvijebb9q3mdo3p6n83c@4ax.com>  F On Tue, 04 Jan 2005 21:34:27 -0700, Lorin Ricker <lorin@locktrack.com> wrote:   >PROGRAM DESCRIPTION:,D >I've written a utility, SETUAI.EXE in Pascal v5.8-92 on VMS v7.2-1,A >which can fetch (display with an application /SHOW), insert &/or ; >update (with /ADD="string"), and delete/remove data in theoB >UAI$_USER_DATA field of a SYSUAF.DAT user/account record, via theA >system RTLs $GETUAI and $SETUAI.  The basic utility command lines >syntax looks like:o  9 Somewhat OT, but I'd be interested to hear how many folkss= here on c.o.v, have run across 3rd party software that makes PA active use of the UAI$_USER_DATA fields?  but namely, which ones?    ------------------------------  $ Date: Fri, 7 Jan 2005 01:55:40 -0800) From: "me to you" <dishking@mwqrvdua.com>k( Subject: Here is $50 for signing up free0 Message-ID: <M-idndUONpsW7kLcRVn-iw@comcast.com>  M    Hello,  I'm trying out a new auto poster, and dont know how to use it veryiB    good, so if this end up in the wrong places, please forgive me,J    and disreguard, sorry for your incovenience, but everyone else, ENJOY!!.         KEEP READING TO GET YOUR $50.00 NOW!!!L   This is not spam, it is a legitament way to make money on the net, readingO   e-mail, and other things as well, plus you get $50.00 dollars for signing up. K   This is tottaly free, and requires no credit card on out of pocket money. J   You will get paid for various things on the net, hope you enjoy and findQ   some usefull info here that will help you out. Plus you will get fifty dollars.eP   this is not spam, its usefull news for all newsgroups readers, from me to you,M   hope you can use it for your benifit, please enjoy. If your not sure, get awH   new email address, and then try, you wont get spammed, and all info isL   carefully guarded, and not given out or traded or sold. Enjoy your $50.00,O   I did, so did my friends and family. You'll also be able to do survays if youd"   want,  P.S. forgive my spelling.   7    Just copy and paste into your browser, thanks again.n        cL        URL: http://www.CashRead.com/cgi-bin/signup.cgi?r=fogylight@yahoo.com        tI        Dont be mean and hatefull, if you dont want to Participate, DON'T.o:        If you don't want $50.00, let someone else have it.   ------------------------------  $ Date: Fri, 7 Jan 2005 23:54:08 -0500: From: "Hein RMS van den Heuvel" <hein@Hewlett-Packard.com>3 Subject: Re: How to have SEARCH match a "null" lineg* Message-ID: <41df679d@usenet01.boi.hp.com>  9 "Ken Fairfield" <my.full.name@intel.com> wrote in messageh% news:crn3ni$gsu$1@news01.intel.com...s  9 > $ SEARCH SYSUAF.LIS str1,str2/Match=NOR/Out=output-files >sH > This is simplified, and in fact, I manage to also strip off the header line in the process.B > However, the record following the original header line is empty.  F So you do not want the header line, the empty line following it, and a" collections of other lines. Right?$ In pure DCL I woudl solve that with:   $ open x sysuaf.list $ read x xx  $ read x xxi% $ sea x          Owner,str2/match=norl <output>	 $ close x   4 But in reality I would solve it with perl of course:  @ $ perl -ne "next if /^$/ || /Owner/ || /str2/; print" sysuaf.lis <output>  J -n = force loop through input equal to:  OPEN (FILE,stdin); WHILE (<FILE>) { block } CLOSE FILE" -e = immediate commands comming upC /^$/ = an empty string ( ^ anchors the record start, $ the end, andc nothingness in between.)  ! The reason I would choose perl isCC - the availabilty of regular expression for infinite flexibility inS. selecting (and in confusing future readers :-)L - because I probably needed additional record processing, for example column# selections, readily done with perl.d   hth, Hein.    ------------------------------   Date: 7 Jan 2005 13:30:52 -0600t4 From: kuhrt@nospammy.encompasserve.org (Marty Kuhrt)& Subject: Re: HP/Linux Tivo replacement3 Message-ID: <NgyIorATBH4M@eisner.encompasserve.org>g  _ In article <3f119ada050106103674def2a0@mail.gmail.com>, DeanW <dean.woodward@gmail.com> writes:hG > On Thu, 6 Jan 2005 06:35:27 -0500, John Smith <a@nonymous.com> wrote:C6 >> http://pcworld.com/news/article/0,aid,119173,00.asp >> vB >> Do you think that with the same amount of man-hours invested in@ >> customizing Linux for the task, HP could have made VMS do the >> same thing for them?r >  > Not even close.h- >    a) VMS doesn't run on commodity hardwarerD >    b) The device drivers (capture card) don't already exist on VMS  D The AV321 PCI card that I bought for $40 used has an RCA video inputD and can record from that source.  There are some demo apps that can D be used to record and play back video, but you have to have the DECWA server in 8bit mode for them to work (at least the ALPHAVCR app, e6 anyway).  It's not a tuner, but it will capture video.   ------------------------------   Date: 7 Jan 2005 13:34:14 -0600r4 From: kuhrt@nospammy.encompasserve.org (Marty Kuhrt)& Subject: Re: HP/Linux Tivo replacement3 Message-ID: <6ZO6gglZKyCE@eisner.encompasserve.org>t  n In article <e61rt0tjsvsgtner1bpv2ropskcp5plbqn@4ax.com>, Steve Lionel <Steve.Lionel@REMOVEintelME.com> writes:M > On Thu, 06 Jan 2005 12:43:31 -0500, JF Mezei <jfmezei.spamnot@teksavvy.com>L > wrote: > O >>Rela Tivo boxes have been hacked by Linux geeks who were able to add programseP >>and features to the standard Tivo software, enabling, for instance FTP service= >>so that you could copy moving to/from the Tivo and your PC.e > ' > I know - I've done it to my own TiVost > M >>Such hackjs became popular enough that Tivo is now starting to enable thosec$ >>features in the unhacked machines. > L > I doubt there's a cause-and-effect here. But clearly TiVo saw a demand andF > worked up a solution that would not get them sued into oblivion like > SonicBlue. > P >>Had Tivo chosen a more secure OS such as VMS for an embedded box, such hackersP >>would not likely have been able to penetrate the box which would have remained >>a big black box. > ; > The security of the OS had nothing to do with the hacks.   > M > My point was that none of the "DVR on Linux" (or even on Windows) solutionsaO > approaches a TiVo for feature set, ease of use, ease of integration, low cost  > and reliability.  F How about FreeVo, http://freevo.sourceforge.net/about.html ?  At firstD glance, it seems to have most of the features of a basic TiVo setup.   ------------------------------  % Date: Fri, 07 Jan 2005 14:58:48 -0500e3 From: Steve Lionel <Steve.Lionel@REMOVEintelME.com>n& Subject: Re: HP/Linux Tivo replacement8 Message-ID: <7aqtt0tg1ljusne8jseunk8sto9rlml33c@4ax.com>  L On 7 Jan 2005 13:34:14 -0600, kuhrt@nospammy.encompasserve.org (Marty Kuhrt) wrote:    N >> My point was that none of the "DVR on Linux" (or even on Windows) solutionsP >> approaches a TiVo for feature set, ease of use, ease of integration, low cost >> and reliability.t >oG >How about FreeVo, http://freevo.sourceforge.net/about.html ?  At firsttE >glance, it seems to have most of the features of a basic TiVo setup.r  L No name-based recording (Wishlists, Season Passes), no cable box control, noM elimination of duplicate episodes, no recognition of timeslot changes, ... ItrN also requires a pretty much dedicated high-end PC and a complex build process.  N See, this is the problem - people who don't have TiVo think TiVo is just a VCR2 with  a hard disk and an EPG.  It is so much more.   Steve    ------------------------------  $ Date: Fri, 7 Jan 2005 11:04:52 -0800* From: "Jack Peacock" <peacock@simconv.com>A Subject: Re: Mark Gorham interview in Interex's HP World MagazineN2 Message-ID: <l7-dnZQxqvPnQUPcRVn-ow@mpowercom.net>  ? "Keith Parris" <keithparris_NOSPAM@yahoo.com> wrote in message i* news:b4fDd.5239$1j.632@news.cpqcorp.net...$ > "Q: Is OpenVMS growing these days? > I > OpenVMS has been growing for the last year. Weve had four quarters of nJ > either single- or double-digit revenue growth. Its been quite exciting 
 > for us." >bK This is a very skillful PR answer, one that seemingly answers the question  K but misdirects in the process.  Growth is defined as revenue, not customer dK base.  I would not consider getting more money out of a shrinking customer  L base as a positive.  Exciting, yes, but not the kind of excitement I'd look M for.  Is the marketing budget for VMS increasing, are more salespeople being cM trained on VMS?  Or is this just the cash cow being milked before it becomes r	 dog food?@  L Just back from China, talked to an outfit there that used to support VMS in I process control systems (Basestar).  Head guy agreed with me, VMS has no cL visibility anymore.  Allen-Bradley has dropped all HP and Sun support.  For M process control systems using PLCs it's now a single OS, Windows or nothing. eL I'd hazard a guess that if there's any growth in that particular VMS sector I it's for help in converting from VMS, not to VMS.  (Interesting that Sun eH isn't far behind VMS in the perception of being a legacy, phase it out, ( system, at least with the people I met.)   Jack Peacock m   ------------------------------   Date: 7 Jan 2005 13:40:23 -0600 4 From: kuhrt@nospammy.encompasserve.org (Marty Kuhrt)A Subject: RE: Mark Gorham interview in Interex's HP World Magazines3 Message-ID: <Nn5A5wvqdPYm@eisner.encompasserve.org>M   In article <2E9EAC858F94A04ABAA9955F81C1B0B9E153FB@hadaltmail.althad.sa.gov.au>, "Barratt, Chris \(FMC\)" <Chris.Barratt@fmc.sa.gov.au> writes:pH > What is the ".NET integration technology" that is being brought to the > OpenVMS platform ?  H It's the follow on to the Affinity Program called the Antipathy Program.   ;^)    ------------------------------  $ Date: Fri, 7 Jan 2005 15:52:45 -0500# From: "John Smith" <a@nonymous.com> A Subject: Re: Mark Gorham interview in Interex's HP World MagazineP, Message-ID: <0t6dna0K0d4Ia0PcRVn-oQ@igs.net>   Jack Peacock wrote:,@ > "Keith Parris" <keithparris_NOSPAM@yahoo.com> wrote in message, > news:b4fDd.5239$1j.632@news.cpqcorp.net...% >> "Q: Is OpenVMS growing these days?, >>F >> OpenVMS has been growing for the last year. We've had four quartersD >> of either single- or double-digit revenue growth. It's been quite >> exciting for us." >>C > This is a very skillful PR answer, one that seemingly answers the ) > question but misdirects in the process.,   Absolutely correct.c     > Growth is defined asF > revenue, not customer base.  I would not consider getting more moneyE > out of a shrinking customer base as a positive.  Exciting, yes, buts* > not the kind of excitement I'd look for.  J More like the bucket shops selling worthless penny stocks whose salepeopleC rang a cash register each time customer....er..a sucker bought somey worthless stock.     > Is the marketing budget fora< > VMS increasing, are more salespeople being trained on VMS?  K Budget?? What budget? As for salespeople, HP has no control of who Best BuyI or OfficeMax hires.1     > Or iseA > this just the cash cow being milked before it becomes dog food?l   Very perceptive.     > F > Just back from China, talked to an outfit there that used to supportF > VMS in process control systems (Basestar).  Head guy agreed with me,F > VMS has no visibility anymore.  Allen-Bradley has dropped all HP andA > Sun support.  For process control systems using PLCs it's now aa  > single OS, Windows or nothing.  : Wait until a Bhopal-type disaster occurs on a Windows box.    & I'd hazard a guess that if there's anyG > growth in that particular VMS sector it's for help in converting from,E > VMS, not to VMS.  (Interesting that Sun isn't far behind VMS in thenG > perception of being a legacy, phase it out, system, at least with thea >   people I met.)  8 Would you like to supersize your Windows or Linux order?   ------------------------------  % Date: Fri, 07 Jan 2005 21:29:48 -0600t2 From: David J Dachtera <djesys.nospam@comcast.net>% Subject: OT: Laptop hard disk upgradee+ Message-ID: <41DF53AC.E24D3E9C@comcast.net>p   Folks,  H I was wondering if someone who's been through this would correspond withH me off-line (how to de-mung the reply-to should be obvious, but I'd also? recommend using the Comcast address since I get so much spam ata Earthlink).a  F I bought this machine last year near this time used from a company outG east. It's a ThinkPad T20 and only has a 5GB hard drive. Naturally, I'ms; looking for bigger without replacing this machine just yet.r  D The "how-to" I'm looking for is, of course, getting this hard disk's contents onto the new device.    Any help would be appreciated.   -- 8 David J Dachtera dba DJE Systemst http://www.djesys.com/  ) Unofficial OpenVMS Hobbyist Support Page:a" http://www.djesys.com/vms/support/  ( Unofficial Affordable OpenVMS Home Page: http://www.djesys.com/vms/soho/p  " Unofficial OpenVMS-IA32 Home Page: http://www.djesys.com/vms/ia32/r   Coming soon:& Unofficial OpenVMS Marketing Home Page   ------------------------------  % Date: Fri, 07 Jan 2005 14:50:25 -0800t, From: Ken Fairfield <my.full.name@intel.com>2 Subject: Q: How to have SEARCH match a "null" line+ Message-ID: <crn3ni$gsu$1@news01.intel.com>.   I'm taking the output of:l   $ Mcr Authorize List/Brief   and doing a:  7 $ SEARCH SYSUAF.LIS str1,str2/Match=NOR/Out=output-file   ; This is simplified, and in fact, I manage to also strip offv the header line in the process.u  9 However, the record following the original header line is88 empty.  The above SEARCH command retains it.  I've tried adding:    	" "/Key=(Posi:1,Size:1)  B which is how I am able to strip the header line (with /Match=NOR). I also tried this variant:   	""/Key=(Posi:1,Size:1)s  ; which matches _all_ lines _except_ the empty line followinga= the header.  So adding a /Match=NOR dumps everything _except_m the empty line.i  = It looks like I need to do two SEARCHs to accomplish what I'mt> after, or READ the file after the search and write it back out after skipping the empty line.  > Can anyone think of a combination of qualifiers & paramters to8 SEARCH that would _select_ the empty line so that adding /Match=NOR would dump it?a   	Thanks, Ken -- e6 I don't speak for Intel, Intel doesn't speak for me...  
 Ken Fairfield ! D1C Automation VMS System Support " who:   kenneth dot h dot fairfield where: intel dot com   ------------------------------  * Date: Sat, 8 Jan 2005 02:29:06 +0000 (UTC), From: lewis@OMEGA.MITRE.ORG (Keith A. Lewis)6 Subject: Re: Q: How to have SEARCH match a "null" line. Message-ID: <crnghi$stq$2@newslocal.mitre.org>   Ken Fairfield <my.full.name@intel.com> writes in article <crn3ni$gsu$1@news01.intel.com> dated Fri, 07 Jan 2005 14:50:25 -0800: 8 >$ SEARCH SYSUAF.LIS str1,str2/Match=NOR/Out=output-file  > >It looks like I need to do two SEARCHs to accomplish what I'm? >after, or READ the file after the search and write it back out  >after skipping the empty line.G  U $ pipe SEARCH SYSUAF.LIS str1,str2/Match=NOR | SEARCH SYS$INPUT: ""/key=(pos:1,siz:1)   ) It's even shorter if you use this symbol:n   GREP == "SEARCH SYS$INPUT:"   0 --Keith Lewis              klewis {at} mitre.org> The above may not (yet) represent the opinions of my employer.   ------------------------------  % Date: Fri, 07 Jan 2005 21:07:07 -0600e2 From: David J Dachtera <djesys.nospam@comcast.net>6 Subject: Re: Q: How to have SEARCH match a "null" line+ Message-ID: <41DF4E5B.258AC6F4@comcast.net>h   "Keith A. Lewis" wrote:A >  > Ken Fairfield <my.full.name@intel.com> writes in article <crn3ni$gsu$1@news01.intel.com> dated Fri, 07 Jan 2005 14:50:25 -0800:e: > >$ SEARCH SYSUAF.LIS str1,str2/Match=NOR/Out=output-file > @ > >It looks like I need to do two SEARCHs to accomplish what I'mA > >after, or READ the file after the search and write it back outr! > >after skipping the empty line.$ > W > $ pipe SEARCH SYSUAF.LIS str1,str2/Match=NOR | SEARCH SYS$INPUT: ""/key=(pos:1,siz:1)t  D That would be my suggestion, also. Use SEARCH in a PIPEline. just beF careful about placing the filters in the pipe so you get what you want at the output end.   -- g David J Dachtera dba DJE Systems? http://www.djesys.com/  ) Unofficial OpenVMS Hobbyist Support Page:e" http://www.djesys.com/vms/support/  ( Unofficial Affordable OpenVMS Home Page: http://www.djesys.com/vms/soho/   " Unofficial OpenVMS-IA32 Home Page: http://www.djesys.com/vms/ia32/t   Coming soon:& Unofficial OpenVMS Marketing Home Page   ------------------------------  % Date: Fri, 07 Jan 2005 21:22:45 -0600>2 From: David J Dachtera <djesys.nospam@comcast.net> Subject: Re: Samsung!8+ Message-ID: <41DF5205.A8E62AC7@comcast.net>o   GWDVMS::MOELLER wrote: > B > Seen on Korean (english language) TV <http://www.arirangtv.com>: > ? > In Tuesday's new report, they had a VERY LOW KEY ANNOUNCEMENTi> > of Samsung, the world's leading chipmaker, which mentioned a > company named INTEL.D > [Since I happened to look elsewhere, I did not get the full text.] > D > Do I remember correctly that these folks are free to build Alphas?  " I thought that was the deal, yes.   C I wonder if HP can fall back on Samsung to keep building Alpha CPUsO/ until the Itanic is raised and put to rights...p   --   David J Dachtera dba DJE Systemsa http://www.djesys.com/  ) Unofficial OpenVMS Hobbyist Support Page:m" http://www.djesys.com/vms/support/  ( Unofficial Affordable OpenVMS Home Page: http://www.djesys.com/vms/soho/h  " Unofficial OpenVMS-IA32 Home Page: http://www.djesys.com/vms/ia32/e   Coming soon:& Unofficial OpenVMS Marketing Home Page   ------------------------------   Date: 7 Jan 2005 21:42:43 -0600i+ From: young_r@encompasserve.org (Rob Young)o Subject: Re: Samsung!z3 Message-ID: <ShsfXc3c7PWh@eisner.encompasserve.org>   ` In article <41DF5205.A8E62AC7@comcast.net>, David J Dachtera <djesys.nospam@comcast.net> writes: > GWDVMS::MOELLER wrote: >> eC >> Seen on Korean (english language) TV <http://www.arirangtv.com>:F >>  @ >> In Tuesday's new report, they had a VERY LOW KEY ANNOUNCEMENT? >> of Samsung, the world's leading chipmaker, which mentioned aR >> company named INTEL.iE >> [Since I happened to look elsewhere, I did not get the full text.]  >>  E >> Do I remember correctly that these folks are free to build Alphas?s   	Yes.  g  D 	But here is the low key announcement you are no doubt referring to:    i http://www.realworldtech.com/forums/index.cfm?action=detail&PostNum=3011&Thread=1&entryID=44535&roomID=11    Topic: Hang together y  * Name: Paul DeMone (pdemone@igs.net) 1/5/05  3 In an earlier thread I predicted due to ever rising 0 costs in the semi industry smaller players would2 have to start teaming up in consortiums to stay in1 the wafer spinning business. It looks like it has  started already:  N "LONDON . George Pataki, the Governer of New York, is expected Wednesday (Jan.M 5) to unveil a plan for an IBM-led consortium that includes AMD, Infineon andNN Sony to invest more than US$2.5 billion in a wafer fab, according to a Reuters report.   I The report also said that lithography equipment maker ASML Holding NV haseH agreed to create its first research center outside Europe as part of theN project and that IBM had pledged to open up a semiconductor production line at% the Albany Nanotech research center. l  N IBM Corp., Sony Corp, Advanced Micro Devices Inc. and Infineon Technologies AGO have pledged $1.9 billion to build the core of the investment project . a waferuL fab at IBM's existing Fishkill, New York facility, the Reuters report said.   H Another report, from the Albany Times Union said that IBM would have sixM partners in the wafer fab project . Sony, Toshiba, Samsung, Infineon, AMD andV
 Chartered"   Here's the link:  ^ http://www.siliconstrategies.com/article/showArticle.jhtml?articleId=56900573&_requestid=72041  h ---   ( 	As Paul notes: "they'll hang together!"  A 	Why did you hear Intel?  No doubt something about Intel develops'> 	CPUs too, this appears to be an alliance opposing Intel, etc.   				Robr   ------------------------------   Date: 7 Jan 2005 12:58:43 -0600 ; From: koehler@eisner.nospam.encompasserve.org (Bob Koehler)d( Subject: Re: Simple EDT or TPU init file3 Message-ID: <Nd1dG+cwBH+p@eisner.encompasserve.org>p  i In article <newscache$n2gy9i$45a$1@news.sil.at>, peter@langstoeger.at (Peter 'EPLAN' LANGSTOEGER) writes:e > L > Or better a "$ DEFINE/SYSTEM EVE$KEYPAD EDT" in SYS$MANAGER:SYLOGICALS.COM  A    Anybody who defines that system wide better be a hobbyist on ay    single user system.  A    If you try that on one of my systems, you'll soon find out how8C    it feels to be disuser'ed, and loose any privileges you may have     been approved for.1   ------------------------------  % Date: Fri, 07 Jan 2005 11:34:06 -0800o, From: Ken Fairfield <my.full.name@intel.com>( Subject: Re: Simple EDT or TPU init file+ Message-ID: <crmo7e$b5a$1@news01.intel.com>n   Big John wrote:h  D > Over the holiday period, I was thinking about TPU (well, what elseB > is there to do?), and I thought about some possible solutions to > your problems..e >  > > >>First, using SHOW TERMINAL/FULL I find the keypad is usually> >>in numeric mode by default to start with.  It needs to be in< >>application mode before you call SIMPLE or else the keypad= >>doesn't work and you can't get out except with a CONTROL-Y.. >  > B > Since we can't 'fix' this (as we don't know how your emulator isF > working - clearly it's not a perfect emulation of a VT300), there isF > a nice simple workaround for it.  And this is to define any '1' key,C > the keypad 1, plus any other key that sends '1' to the system, totF > all do the same thing that kp1 does. Similarly for the other defined > kp keys, 2, 4, 6 and 8.  > ; > Thus we have 2 define_key commands for each command, e.g.c > 8 > define_key ("message(ascii(27)+'[?25h');QUIT (OFF,1)", > kp1, the_key_map);8 > define_key ("message(ascii(27)+'[?25h');QUIT (OFF,1)", > KEY_NAME ("1"), the_key_map);y > D > - so that whatever your keypad 1 key sends, it will be interpreted  > by the TPU file as you intend. [BIG SNIP!]l  ; John, have you confirmed this actually works as expected???V= I seem to recall that you can't redefine the "self-inserting" 8 keys, i.e., the typing keys, letters and numbers, etc...  < On a "real" VT, you couldn't tell the difference because TPU; puts the keypad in application mode on entry.  However, I'dr5 expect some complaint from the attempt to define_key.u  ( 	Admiting-I-haven't-tried-it-myself, Ken -- I6 I don't speak for Intel, Intel doesn't speak for me...  
 Ken FairfieldU! D1C Automation VMS System Support " who:   kenneth dot h dot fairfield where: intel dot com   ------------------------------  * Date: Fri, 7 Jan 2005 21:47:48 +0000 (UTC)6 From: peter@langstoeger.at (Peter 'EPLAN' LANGSTOEGER)( Subject: Re: Simple EDT or TPU init file0 Message-ID: <newscache$n7vy9i$36b$1@news.sil.at>  q In article <Nd1dG+cwBH+p@eisner.encompasserve.org>, koehler@eisner.nospam.encompasserve.org (Bob Koehler) writes:sj >In article <newscache$n2gy9i$45a$1@news.sil.at>, peter@langstoeger.at (Peter 'EPLAN' LANGSTOEGER) writes: >> nM >> Or better a "$ DEFINE/SYSTEM EVE$KEYPAD EDT" in SYS$MANAGER:SYLOGICALS.COMh >YB >   Anybody who defines that system wide better be a hobbyist on a >   single user system.s > B >   If you try that on one of my systems, you'll soon find out howD >   it feels to be disuser'ed, and loose any privileges you may have >   been approved for.  F How would I need to ? A good VMS manager would have already added this7 logical way back in the 80ties (with VMS V4 I think)...    --   Peter "EPLAN" LANGSTOEGER % Network and OpenVMS system specialistn E-mail  peter@langstoeger.atF A-1030 VIENNA  AUSTRIA              I'm not a pessimist, I'm a realist   ------------------------------  $ Date: Fri, 7 Jan 2005 20:24:13 -0500) From: "Neil Rieck" <n.rieck@sympatico.ca>n( Subject: Re: Simple EDT or TPU init file; Message-ID: <hDGDd.68228$P%3.2360044@news20.bellglobal.com>i  ; "JF Mezei" <jfmezei.spamnot@teksavvy.com> wrote in message .< news:1105113966.805431f1ffd2055b8d769fa39b395b66@teranews... > Neil Rieck wrote:s) >> $ cre  sys$library   EVE$INIT_NEIL.EVEd >> set key edt8 >> Now TPU will come up with most EDT commands in place. >i >iK > In this religious centuries old editor war, impersonating another editor h > isJ > punishable by public lapidation. How dare you pretend to be an EDT when  > you're > really a TPU. Sacrilege. :-) >gK > PS, you can also "set keypad wps" to emulate WPSPLUS. Only god knows whatnM > punishement is suitable when someone tries to emulate ALLIN1's editor with t > TPUF > :-) :-) :-) :-) :-) :-)d  H Hey, I've committed the EDT keystrokes to memory so long ago that it is + almost impossible to learn anything new :-)c  I Seriously, I would say that I do 95% of my editing with $EDIT/EDT and 5% tL with TPU impersonating EDT. While is it true that many EDT commands are not J available (eg. "EXIT file.ext") I sometime need it to search/inspect wide # log files (some > 1000 characters).   
 Neil Rieck Kitchener/Waterloo/Cambridge,a Ontario, Canada.9 http://www3.sympatico.ca/n.rieck/links/cool_openvms.html c   ------------------------------   Date: 7 Jan 2005 19:54:50 -0600r+ From: young_r@encompasserve.org (Rob Young)l Subject: Sun shocka - take 93 Message-ID: <NcG8UOKZWkgk@eisner.encompasserve.org>   _ In article <l7-dnZQxqvPnQUPcRVn-ow@mpowercom.net>, "Jack Peacock" <peacock@simconv.com> writes:e  K > it's for help in converting from VMS, not to VMS.  (Interesting that Sun kJ > isn't far behind VMS in the perception of being a legacy, phase it out, * > system, at least with the people I met.)  
 	Really?	    	Shocker.  News At 11.  H 	Sun has been in major decline for 3 years now (1), where have you been?3 	It is easy to move from one Eunuchs to another, ita? 	isn't a surprise that Sun is setting (that and Linux/Windows).   ! 	Here's an analyst's perspective:s  J "This marks the fourth time since 2002 that Sun's stock has seen a one- toM three-month return of 30 percent or greater. However, on each of the previouskK occasions, Sun has declined in the following months, as the company was not0> able to fulfill investors' heightened expectations," he wrote.  @ 	Scotty gets cheery, investors get happy, stock runs up, reality! 	sets in.  Rather, rinse, repeat!F   				Robe    (1) Go ahead - make my day . . .   ------------------------------  % Date: Fri, 07 Jan 2005 23:29:33 -0500a- From: JF Mezei <jfmezei.spamnot@teksavvy.com>2  Subject: Re: Sun shocka - take 9B Message-ID: <1105157930.51e316e2181df978f4b4dea0862bf273@teranews>   Rob Young wrote:Q >         Sun has been in major decline for 3 years now (1), where have you been?   K New sales may have been hurt but the .com bomb, but one important metric isi installed base.L  E Have there been many migrations of existing systems from Sun to other & platforms ? That is the real question.   ------------------------------   Date: 7 Jan 2005 13:22:41 -0600o4 From: kaplow_r@encompasserve.org.TRABoD (Bob Kaplow)% Subject: Re: VMS ready laser printer?.3 Message-ID: <0VAuY5S5z1qr@eisner.encompasserve.org>M  R In article <jZKdnS5LBZxFWkDcRVn-oQ@igs.net>, "John Smith" <a@nonymous.com> writes:E > I have an old Texas Instruments Microlaser LJIII/Postscript printerIF > connected via a parallel port back through an Emulex terminal serverJ > w/printer port (I think....it's been ages since I actually looked at how= > it's cobbled together (maybe it's a serial interface.....).a  I Hmmm, too much stuff here. Want tp be able to use it from VMS, Linux, and. Winblows...o  L > Brother makes some good inexpensive laser printers that may do the job forN > you. Go to the local Staples and get one....bring it back no questions asked > if it doesn't.  $ Nearest one is over 2 hours away :-(  1 	Bob Kaplow	NAR # 18L	TRA # "Impeach the TRA BoD"k& 		>>> To reply, remove the TRABoD! <<<K Kaplow Klips & Baffle:	http://nira-rocketry.org/LeadingEdge/Phantom4000.pdfnL     www.encompasserve.org/~kaplow_r/    www.nira-rocketry.org    www.nar.org  O ... One nation under surveillance, divisive, with liberty and justice for none.t   ------------------------------  % Date: Fri, 07 Jan 2005 11:54:19 -0800r From: JBloggs@acme.com% Subject: Re: VMS ready laser printer?r8 Message-ID: <ecptt05khv39gvhtkvn3s0q3qoi4532rm6@4ax.com>  D On 6 Jan 2005 13:07:22 -0600, kaplow_r@encompasserve.org.TRABoD (Bob Kaplow) wrote:  K >I'm looking for an INEXPENSIVE b/w laser printer for connecting to my home.L >office network. Everything in the mass market stores these days seems to beL >PC/USB only unless you buy an adapter. I want something that my VMS systemsM >can use. Something that understands PCL and maybe even PostScript? Somethingo  >with an ethernet port included. >hI >What's the current equivalent of the old standard Laserjet III / IV withi7 >built in JetDirect? And it doesn't have to be from HP!o > J >BTW, has VMS integrated HP Laserjet support yet so you don't have to roll5 >your own forms libraries for "standard" HP printers?    What's your budget?  t  4 I spent some $$$ for a HP4000TN (w/ JetDirect card) : some eight years ago.  Froogle suggests you might get one = nowadays for ~$600-$300+.  probably without toner cartridge. e  1 Overkill for home use, but it  works flawlessly, -A and I still have yet to replace the toner cartridge. PS support, a> and two paper trays is also a plus.  so far, I haven't needed > to add memory, but I don't use PostScript printing heavily ...  @ I concur that'd it be nice to get supported PCL form-libraries,  now that VMS is part of HP.  i  . I still periodically wrestle with the infamousB blank-page-at-beginning-of-print-jobs when tweaking form setups.     that said, I ought to look at 4 the PathWorks/AdvServer supplied forms for ideas ...   ------------------------------  * Date: Fri, 7 Jan 2005 21:41:47 +0000 (UTC)7 From: hamilton@Encompasserve.org (Bradford J. Hamilton)h% Subject: Re: VMS ready laser printer? . Message-ID: <crmvmr$tj$1@grandcanyon.binc.net>  j In article <0VAuY5S5z1qr@eisner.encompasserve.org>, kaplow_r@encompasserve.org.TRABoD (Bob Kaplow) writes:S !In article <jZKdnS5LBZxFWkDcRVn-oQ@igs.net>, "John Smith" <a@nonymous.com> writes:cF !> I have an old Texas Instruments Microlaser LJIII/Postscript printerG !> connected via a parallel port back through an Emulex terminal server K !> w/printer port (I think....it's been ages since I actually looked at howc> !> it's cobbled together (maybe it's a serial interface.....). !>J !Hmmm, too much stuff here. Want tp be able to use it from VMS, Linux, and !Winblows... !w  M Don't discount this solution out of hand; I know a company that hooked up two>N old HP LaserJets (without JetDirect cards) in a similar manner, with an EmulexO Print Server.  The printers were usable from both VMS and Windows in this way. iL Presumably, this solution would work for any Linux/Unix/BSD distro, as well.  N I have an old Emulex Print Server here at home, serving up my Epson dot-matrix( printer to VMS and Windows.  Works fine.   !snip!   ------------------------------  * Date: Sat, 8 Jan 2005 02:16:41 +0000 (UTC)- From: klewis@OMEGA.MITRE.ORG (Keith A. Lewis) % Subject: Re: VMS ready laser printer?a. Message-ID: <crnfq9$stq$1@newslocal.mitre.org>   kaplow_r@encompasserve.org.TRABoD (Bob Kaplow) writes in article <0VAuY5S5z1qr@eisner.encompasserve.org> dated 7 Jan 2005 13:22:41 -0600:aJ >Hmmm, too much stuff here. Want tp be able to use it from VMS, Linux, and >Winblows...  G If you've got Linux in the loop, you can probably buy a USB printer andmL enable LPR from there.  It will save a little cash, and the only drawback is6 your Linux system has to be running in order to print.  J My own setup uses a Powermac (OS X) as the print server.  Be sure to check3 linuxprinting.org to help choose an "open" printer.   0 --Keith Lewis              klewis {at} mitre.org> The above may not (yet) represent the opinions of my employer.   ------------------------------   End of INFO-VAX 2005.015 ************************