1 INFO-VAX	Tue, 26 Feb 2002	Volume 2002 : Issue 112       Contents:2 A lo hecho, ?pecho! (y te ayudamos un poco m?s...) Re: A problem with a patch Re: A problem with a patch Re: A problem with a patch4 Re: Advice wanted: separating input and output tasks2 Anyone See the full page ad letter to Mr. Hewlett?1 Re: BZIP2 for OpenVMS (was: Re: PGP for OpenVMS?) 1 Re: BZIP2 for OpenVMS (was: Re: PGP for OpenVMS?)  Computer waste. ' Re: DEQNA/DELQA - Checksum calculation?  Euro Symbol on VAX Re: Euro Symbol on VAX, Re: EV8 and McKinley analysis by Paul DeMone9 expensive procedure calls (was: Sv: Younger recruits ...)  Re: If you need PI....% inter-process communication and timer ) Re: inter-process communication and timer ) Re: inter-process communication and timer , Memory Upgrade for Alpha XP-1000 Workstation0 Re: Memory Upgrade for Alpha XP-1000 Workstation On 64-bit architectures  Re: On 64-bit architectures P Re: Oracle RDB  ABUSIVE PRICE$  was (Re: MPE/iX users form OpenMPE organization)P Re: Oracle RDB  ABUSIVE PRICE$  was (Re: MPE/iX users form OpenMPE organization) Re: Please help  Please help  Re: Powerpoint competition Re: PrintServer software) Re: Profitability and the survival of VMS  Re: Question for Backup Gurus   select() does not work for pipes$ Re: select() does not work for pipes$ Re: select() does not work for pipes$ Re: select() does not work for pipes$ Re: select() does not work for pipes$ Re: select() does not work for pipes Re: Shannon on the merger # Re: Shared SCSI Bus Cluster and TCQ # Re: Shared SCSI Bus Cluster and TCQ # Re: Shared SCSI Bus Cluster and TCQ # Re: Shared SCSI Bus Cluster and TCQ # Re: Shared SCSI Bus Cluster and TCQ # Re: Shared SCSI Bus Cluster and TCQ # Re: Shared SCSI Bus Cluster and TCQ # Re: Shared SCSI Bus Cluster and TCQ # Re: Shared SCSI Bus Cluster and TCQ ' Re: SHOW ENTRY /FULL doesn't show /PARA  Re: Singular they < RE: Singular they was Re: Sv: Younger recruits versus experiL Re: Singular they was Re: Sv: Younger recruits versus experienced veterans ($ TCPIP throuput for a VAXstation 3100( Re: TCPIP throuput for a VAXstation 3100( Re: TCPIP throuput for a VAXstation 3100( Re: TCPIP throuput for a VAXstation 3100( Re: TCPIP throuput for a VAXstation 3100& Re: tool to calc recursive dir sises ?& RE: tool to calc recursive dir sises ?& Re: tool to calc recursive dir sises ?& Re: tool to calc recursive dir sises ?& RE: tool to calc recursive dir sises ?& Re: tool to calc recursive dir sises ?& RE: tool to calc recursive dir sises ? Re: VAX SCAN Re: VAX SCAN Re: VAX SCAN Re: VAX SCAN Re: VAX SCAN Re: VAX SCAN Re: VAX SCAN Re: VAX SCAN Re: VAX SCAN Re: VAX SCAN Re: VAX SCAN Re: VAX SCAN Re: VAX SCAN RE: VAX SCAN VAX SCAN survey   F ----------------------------------------------------------------------  % Date: Tue, 26 Feb 2002 04:27:07 -0300  From: "BCA" <compus@nosend.com> ; Subject: A lo hecho, ?pecho! (y te ayudamos un poco m?s...) 9 Message-ID: <iss.569f.3c7b38cd.77b36.1@mx2.east.saic.com>    SDRAM 128 =09=09=09U$S  37.90  256 SDRAM =09=09=09U$S  81.90  PIII 750 BOX =09=09U$S 129.90  PIII 1100 =09=09=09U$S 189.90  PIV 1.5 BOX 478 =09U$S 205.- DURON 850 =09=09=09U$S  61.90  MOUSE OPTICO =09=09U$S  16.90  PC CHIPS 810 =09=09U$S  85.50  MONITOR LG 17" =09=09U$S 210.- HHDD 20 Gb 5400 =09U$S 104.- HHDD 40 Gb 5400 =09U$S 116.-  = ADEM=C1S, TENEMOS UNA LISTA CON CIENTOS DE PRODUCTOS Y VARIAS D CONFIGURACIONES DE MAQUINAS ESPECIALMENTE ARMADAS PARA VOS. SERVICIO- T=C9CNICO A DOMICILIO POR REDES, SOFT Y HARD.    BCA 
 011 4362 3371  busicomputers@yahoo.com   L Esta es una oferta para armadores, si no desea recibir m=E1s ofertas env=ED= e C un email vac=EDo a: webreadynow@yahoo.com con el subjet: "remove" o  "unsuscribe"   ------------------------------  % Date: Tue, 26 Feb 2002 09:30:24 +0100 1 From: "Tomasz Dryjanski" <tdryjanski@hotmail.com> # Subject: Re: A problem with a patch - Message-ID: <a5fh30$10a$1@news2.ipartners.pl>   F >   Your system disk is apparently corrupted.  Since you didn't make a BACKUPJ >   of your system disk before you applied the (now failed) ECO (who among usL >   really reads and follows the directions, after all? :-), please start byG >   applying the OpenVMS Alpha V7.2-1H1 kit (or better, apply V7.2-2 or  later), * >   and then roll on the appropriate ECOs.  K I had done a backup of the system disk before, and obviously I would prefer I to retain the previous instance. I have installed a fresh copy to try the B patch, and to develop a method to apply it to the previous system.  H >  I have seen a few DDIS errors reported within PCSI installations, andK >  these are almost always triggered by a PCSI file that was not downloaded J >  using correct FTP file transfer semantics -- webbrowser implementationsJ >  seemingly choose to use the binary or the text transfer mode at random.  I I downloaded an executable file, *.axpexe, and ran it on the server. So I ( expect that the patch is OK. Am I wrong?   T. D.    ------------------------------    Date: 26 Feb 2002 11:25:28 +0100* From: eplan@kapsch.net (Peter LANGSTOEGER)# Subject: Re: A problem with a patch * Message-ID: <3c7b6298$1@news.kapsch.co.at>  a In article <a5fh30$10a$1@news2.ipartners.pl>, "Tomasz Dryjanski" <tdryjanski@hotmail.com> writes: J >I downloaded an executable file, *.axpexe, and ran it on the server. So I) >expect that the patch is OK. Am I wrong?   H Not totally. But I had a handful occasions over the years, where a (FTP)H download aborted and left an incomplete *.pcsi-dcx_axpexe which unpackedE WITHOUT any error messages, but the contained PCSI file was of course D incomplete/unusable (that's where the checksum file makes sense ;-).  E I still don't understand why Q still insist on using DCX for VMS ECOs F and not also using ZIPSFX (like they do with eg. MOZILLA) because I soD far have never seen such a bad behaviour in ZIP and ZIP has a better compression, too...    --  < Peter "EPLAN" LANGSTOEGER           Tel.    +43 1 81111 2651; Network and OpenVMS system manager  Fax.    +43 1 81111 888 < KAPSCH AG      Wagenseilgasse 1     E-mail  eplan@kapsch.netH A-1121 VIENNA  AUSTRIA              "I'm not a pessimist, I'm a realist"   ------------------------------  % Date: Tue, 26 Feb 2002 15:57:59 +0100 1 From: "Tomasz Dryjanski" <tdryjanski@hotmail.com> # Subject: Re: A problem with a patch - Message-ID: <a5g7q1$kfe$1@news2.ipartners.pl>   L > >I downloaded an executable file, *.axpexe, and ran it on the server. So I+ > >expect that the patch is OK. Am I wrong?  > J > Not totally. But I had a handful occasions over the years, where a (FTP)J > download aborted and left an incomplete *.pcsi-dcx_axpexe which unpackedG > WITHOUT any error messages, but the contained PCSI file was of course F > incomplete/unusable (that's where the checksum file makes sense ;-).  = ... and this was the case. Thank you /so/ much for your hint! - BTW. In what way can I use the checksum file?    T. D.    ------------------------------  % Date: Tue, 26 Feb 2002 18:02:09 -0000 - From: wspencer@ap.nospam.org (Warren Spencer) = Subject: Re: Advice wanted: separating input and output tasks 7 Message-ID: <91C186A24warrenspencer1977@209.249.90.100>   ; owen@astro.washington.edu (Russell E. Owen) wrote in <owen- - 2BA5C8.15204519022002@news.u.washington.edu>:   F >We have a telescope control system that reads commands from multiple J >users and writes all replies to all users. Each user is a separate login G >running a separate interactive command interpreter image. Much of the  I >work is done by detached background processes and much of the output is  I >not solicited by commands. The control system is the only thing running   >on the computer.  > I >Originally I used sys$broadcast to send the output. It was dirt simple,  J >but any one user blocking output (with ^S) would cause a 20 second delay 1 >for output to all users, which was unacceptable.  > I >I then wrote a message hub. Any process wanting to send output sends it  G >to the hub (via a mailbox). The hub then copies it to a set of output  F >mailboxes, one per user. (It is smart enough to deal gracefully with J >full mailboxes, i.e. if a user gets behind). The question is how best to ; >get the data from those mailboxes to the individual users.  > D >Presently I set up an AST in the user's command interpreter. It is J >called whenever there is data in the user's output mailbox, it reads the E >mailbox and writes the data to the user's terminal. Mostly it works  I >fine, but it uses synchronous write, so if the user blocks output (with  I >^S), the user's interpreter is locked and this can cause problems under   >unusual circumstances.  > G >I'm trying to figure out the best way to fix this problem, best being  I >simple, robust and with adequate performance. Solutions I've thought of  	 >include:  > F >*** Have a separate subprocess for output and ditch all use of ASTs. G >This is as simple as it gets. The subprocess does nothing more than a  I >blocking read on the user's mailbox, then a blocking write of that data  0 >to sys$output. I tried to do this first, but... > I >But there's a catch: how do I make sure that this subprocess exits when  F >the user's command interpreter process exits? An exit handler in the G >interpreter will not suffice, because exit handlers are not called if  ' >the user does a ^Y followed by a STOP.   G If it's truly a sub-process (this is a proper name in OpenVMS) of each  L user's main process, it will exit when the parent process exits.  If it's a J detached process, or a subprocess of a different process (like your hub), % you'll have to handle exits yourself.    I >*** Have the output AST write via asynchronously (via SYS$QIO), instead  J >of blocking. I know I can make this work, but it sounds messy. I'll have C >to fire an AST when the write finishes (probably the same AST I'm  H >already firing). Otherwise, if new data comes in while the old data is H >being output, the existing AST will see that it cannot yet write, will 0 >give up for now, and may never get fired again.  J True - but since you'll get an AST when the write finishes, you can check  for new data then.   > J >This seems like a lot of ASTs to fire and it adds complexity to a system # >that already seems fairly complex.   L This may be overkill, but BEA Message Queue is a messaging middle-ware that E supports the notion of (non-guaranteed) broadcasts.  There are other  ( products available too, and all cost $$.    H >Any suggestions? I wish there was some way to simply connect a mailbox I >to a terminal, such that anything written to the mailbox is sent to the  6 >terminal. That's really all I'm trying to accomplish. > 	 >Regards,  >  >-- Russell      --     Warren Spencer' Senior Software Engineer (not a writer)  The Associated Press  < ** Time flies like an arrow.  Fruit flies like a bananna. **   ------------------------------  # Date: Tue, 26 Feb 2002 13:24:39 GMT ' From: "Jean Norton" <jean@staffing.com> ; Subject: Anyone See the full page ad letter to Mr. Hewlett? : Message-ID: <r0Me8.18956$dj3.712279@typhoon.austin.rr.com>    Talk about airing dirty laundry.   ------------------------------    Date: 26 Feb 2002 11:35:57 +0100* From: eplan@kapsch.net (Peter LANGSTOEGER): Subject: Re: BZIP2 for OpenVMS (was: Re: PGP for OpenVMS?)( Message-ID: <3c7b650d@news.kapsch.co.at>  f In article <hKBe8.167$fL6.4426@news.cpqcorp.net>, hoffman@xdelta.zko.dec.nospam (Hoff Hoffman) writes: > , >  I ported a version of bzip2 a while back. > = >    http://www.openvms.compaq.com/freeware/freeware50/bzip2/   H Steve, could you please provide ZIPfiles there (instead of single files)J or let the webserver fix, so that normal text file not named .TXT or .HTML< (like .,.C,.H,.1) are sent as text and not as octet-stream ?   TIA    --  < Peter "EPLAN" LANGSTOEGER           Tel.    +43 1 81111 2651; Network and OpenVMS system manager  Fax.    +43 1 81111 888 < KAPSCH AG      Wagenseilgasse 1     E-mail  eplan@kapsch.netH A-1121 VIENNA  AUSTRIA              "I'm not a pessimist, I'm a realist"   ------------------------------  # Date: Tue, 26 Feb 2002 15:14:00 GMT # From: "John Smith" <a@nonymous.com> : Subject: Re: BZIP2 for OpenVMS (was: Re: PGP for OpenVMS?)/ Message-ID: <YCNe8.2026$rG.1071@news2.bloor.is>   ? "Hoff Hoffman" <hoffman@xdelta.zko.dec.nospam> wrote in message + news:hKBe8.167$fL6.4426@news.cpqcorp.net...  > G >   If y'all get a decent/recent PGP port, please pass along a kit (for @ >   the next Freeware) or a pointer (for the next FAQ).  Thanks! >   H I suggest that the existence of a VMS PGP port also be made known to theJ following sites, as not all VMS users will necessarily come here or to the freeware CD's to find it.   L In addition, the fact that a VMS PGP implementation is found on the commonlyJ searched PGP sites will simply act as a form of advertising for VMS, ie. a= customer or potential VMS customer can point to the fact that I shareware/freeware does in fact exist for VMS, so it's one more reason to # consider buying VMS or keeping VMS.      http://www.ipgpp.com/ # http://web.mit.edu/network/pgp.html  http://www.pgpi.org/   ------------------------------  % Date: Tue, 26 Feb 2002 12:12:16 +0100 9 From: Jan-Erik =?iso-8859-1?Q?S=F6derholm?= <aaa@aaa.com>  Subject: Computer waste.' Message-ID: <3C7B6D90.F6F617FA@aaa.com>   8 Not 100% on topic, but it's at least about computers and I makes you think...  ' http://www.theinquirer.net/26020202.htm    /Jan-Erik Sderholm.   ------------------------------  % Date: Tue, 26 Feb 2002 13:06:38 +0100 % From: Harti Brandt <hbb@fokus.gmd.de> 0 Subject: Re: DEQNA/DELQA - Checksum calculation?> Message-ID: <20020226130437.E40774-100000@beagle.fokus.gmd.de>  , On Sun, 24 Feb 2002, Dennis Henriksen wrote:   DH>sword7@speakeasy.org wrote: DH> 8 DH>> In comp.sys.dec Mark Hittinger <bugs@pu.net> wrote: DH>>8 DH>>>DELQA was newer, faster, less buggy than the DEQNA. DH>>>  DH>>6 DH>>>Post 5.2 VMS starting making DEQNA's unsupported. DH>>>  DH>>I DH>> Ok, I get it.  I will see if OpenVMS v7.2 works with DEQNA.  If not, I DH>> I have modify my DEQNA emulation to funcion as DELQA.  Do both DEQNA - DH>> and DELQA have same functions/registers?  DH>  DH> E DH>Please excuse a dumb question, but didn't the DEQNA work with some E DH>of the pre vax systems pdp-11 and/or pdp-10 ? If so than the DEQNA F DH>simulation would be of value in those simulators or am I mistaken ?  F They work with PDP11s and I have a DEQNA emulation in my p11-emulator.  0 ftp://ftp.fokus.fhg.de/pub/cc/cats/usr/harti/p11   harti  --W harti brandt, http://www.fokus.gmd.de/research/cc/cats/employees/hartmut.brandt/private !               brandt@fokus.fhg.de    ------------------------------    Date: 26 Feb 2002 04:45:13 -08003 From: pa@it.singer-friedlander.com (Piyush Avichal)  Subject: Euro Symbol on VAX = Message-ID: <eb55ac2a.0202260445.4c330d74@posting.google.com>    Hello,  E We currently use Reflections 9.0 to access our VAX and UNIX machines. F I have managed to get the Euro symbol to come up on our UNIX machines,1 by changing to ISO LATIN 9 and typing CTRL-ALT-4. D However when I do the same on the VAX I get a $ symbol. We are usingG VMS 7.1. According to the openvms website we do not need to do anything / but it seems obvious that something is missing.    Anyone have any ideas?   Thanks,    Piyush.    ------------------------------  % Date: Tue, 26 Feb 2002 16:09:36 -0000 , From: "Bob Knowles" <bob.knowles@compaq.com> Subject: Re: Euro Symbol on VAX 1 Message-ID: <5rOe8.201$fL6.4972@news.cpqcorp.net>   K I haven't come across Reflections. Which text editor do you use (or is that J what Reflections does)? Is Office Server part of the configuration? If so,K you need a patch (ICF 13 for OffSvr V5 or ICF 1 for OffSvr V6). Euro Symbol & support in Office Server is covered atG http://www.openvms.compaq.com/commercial/newsletter/newsletter.html#tip    b   @ "Piyush Avichal" <pa@it.singer-friedlander.com> wrote in message7 news:eb55ac2a.0202260445.4c330d74@posting.google.com...  > Hello, > G > We currently use Reflections 9.0 to access our VAX and UNIX machines. H > I have managed to get the Euro symbol to come up on our UNIX machines,3 > by changing to ISO LATIN 9 and typing CTRL-ALT-4. F > However when I do the same on the VAX I get a $ symbol. We are usingI > VMS 7.1. According to the openvms website we do not need to do anythingi1 > but it seems obvious that something is missing.e >i > Anyone have any ideas? >:	 > Thanks,  >u	 > Piyush.t   ------------------------------  # Date: Tue, 26 Feb 2002 14:51:08 GMT # From: "John Smith" <a@nonymous.com> 5 Subject: Re: EV8 and McKinley analysis by Paul DeMonep- Message-ID: <whNe8.2015$rG.80@news2.bloor.is>a  H It was a deliberate omission. I know that your reply was more than a bit tongue-in-cheek.  L Which company makes 'industry standard' tires - Michelin, Pirelli, Goodyear,@ Bridgestone/Firestone, Cooper, Nokia, Continental, .....etc....?  L The only thing 'industry standard' about them is the size, speed rating, andI wear factors - eg. 16" 205/60 VR, wear factor = 400 .... all analogous tor the factors I listed.h  G Imagine the lawsuits that would fly, and public outcry,  if BridgestoneaK unilaterally declared their Potenza line of tires, or even all their tires,$ the sole 'industry standard'.e    K Why we sit still for the crap that MS and Intel push down our throats, I'lle never know.         @ "Fred Kleinsorge" <kleinsorge@star.zko.dec.com> wrote in message+ news:VTue8.142$fL6.2592@news.cpqcorp.net... G > You are missing one ingredient.  It also must be able to run Windows.s >e >e >t! > John Smith wrote in message ...eI > >Why don't we define an 'Industry Standary Server' to be what it reallyn8 > >is..... a 19" rack-mount box containing the following > >- one or more power suppliesC
 > >- a fan > >- mounting brackets > >- holes for ventilation$ > >- holes where connectors can fit. > > G > >The rest of it - what's inside - is proprietary in all cases (no pune
 > >intended).  > >dF > >Each of Dell, Compaq, HP, IBM, etc.... industry standard servers isH > >different - the bioses are different, the cpu's may be different, the > memoryF > >organization and physical form-factor thereof may be different, the number? > >of periphel controls and their mix may be different, etc....1 > >eK > >Just because it has an Intel cpu doesn't make it industry standard. It's1I > >like Ford saying that their Focus car is an industry standard car justtI > >because it uses the same kind of light bulb for its dashboard light asi > everys > >other car on the planet.i > >rK > >It is our job to beat the press over the head when they are fast-n-loosed > >with what they write. > >c > >e > >s > >?; > >"Rob Young" <young_r@encompasserve.org> wrote in messageA0 > >news:0k$edNglCJpf@eisner.encompasserve.org...H > >> In article <fbwd8.153867$Aw2.11063336@bin7.nnrp.aus1.giganews.com>, "BillN) > >Todd" <billtodd@metrocast.net> writes:  > >>J > >> > 2.  To avoid complete redundancy it seems reasonable to assume that anI > >> > 'industry standard box' is something more than simply any box thatc
 > >happens1 > >> > to contain an industry standard processor.. > >> > >>F > >> One wrinkle.. Industry Standard Server.  Since AMD really doesn'tI > >> do much with IA32 servers (their share of the server space is almost G > >> non-existent ... currently).  So Industry Standard Server has come G > >> to mean a server with an Intel processor in it and most observers,UG > >> writers and analysts would tend to agree with that.  By leveraging D > >> that built in marketing, the next step of course is to refer toA > >> an IPF box as an Industry Standard Server and of course your E > >> corporation wants to participate in using Industry Standards andDD > >> leverage all the cost savings and other benefits of being on an  > >> Industry Standard platform. > >> > >> Rob > >> > >= > >  >e >l   ------------------------------   Date: 26 Feb 2002 10:32:11 GMT3 From: anton@mips.complang.tuwien.ac.at (Anton Ertl),B Subject: expensive procedure calls (was: Sv: Younger recruits ...), Message-ID: <a5fo7b$5m5$1@news.tuwien.ac.at>  7 In article <name99-2502021512250001@handma2.apple.com>, )  name99@mac.com (Maynard Handley) writes:S@ >In article <3C72FD44.2058A1ED@ev1.net>, richmond@ev1.net wrote:D >> This "come from" constuct that uses the ASSIGNED GOTO variable isE >> useful to create a "simulated" recursive routine. I once wrote oneOC >> to traverse an AVL tree and print the nodes out in alphabetical  	 >> order.l  B No need to use COME FROM for that.  The second recursive call is aC tail call and can be converted into a goto without needing a return.E (and thus no COME FROM to simulate the return).  This leaves only onehD recursive call, so we don't need the return address to return to it,A we can do it with an ordinary (non-assigned) goto and only need alE return depth counter to tell us if we have another internal return tot2 perform or are done and can do the regular return.  ; >> It is blindingly *fast*...because there is *no* functionf >> call overhead.   A No function call in printing the node?  Apart from that, printing.= probably takes so much time that the tree-walking overhead is-7 comparatively small and optimizing it does not pay off.8  5 >> You have to stack and unstack the local variables c >> yourself. >wJ >Ah, once again the "blindingly fast because of no function call overhead"G >claim. If you look at any decent RISC machine (like a PPC) there IS no H >function call "overhead". The set of things that are done on a functionH >call is exactly the same as the set of things that are done in a set of >goto's faking a function call.h  B The last time I looked (a few days ago), the following things were done on a function call:   - save caller-saved registers  - set up arguments - perform the call - save callee-saved registersr - update stack pointer   and the reverse on return.  ? When emulating a call with a goto, you have to do some of theselE things, but you can optimize some of them away.  E.g., in the examplenC above, there is no need to save the return address of the recursiveoD call, but the compiler does save it.  Similarly, many registers needA not be saved because they stay the same or can be recomputed more  cheaply.  E There have been a number of papers about value prediction in the lastAF decade (e.g., [lipasti+96]), and some of them (sorry, I don't find theC references right now) looked at the reasons for the predictability;TA the loads that restore saved registers were found to be very wellwF predictable, and IIRC they found that the values in many registers are@ not changed for a long time, but still saved and restored a lot.   @InProceedings{lipasti+96,?   author = 	 "Mikko H. Lipasti and Christopher B. Wilkerson ando 		  John Paul Shen",7   title = 	 "Value Locality and Load Value Prediction",c   crossref =	 "asplos96",d   pages =	 "138--147",=   annote =	 "Starts with the surprising observation that moree3 		  than half of the dynamic loads (on both PPC and  		  Alpha) get the same values7 		  that they loaded the last time that static load wasc5 		  executed. They exploit this fact with a mechanism-5 		  that reduces load latency for correctly predicted0/ 		  values to zero (with a one-cycle penalty on 6 		  misprediction). They also propose a mechanism that7 		  completely avoids an access to the memory hierarchy.4 		  for highly predictable loads, but this mechanism6 		  does not look very cost-effective to me. They show5 		  how their mechanisms could be integrated into ther7 		  21164 and into the PPC~620, and that they (in theire8 		  simple forms) would provide a speedup of 6\% for the# 		  21164 and 3\% for the PPC~620."e }  @Proceedings{asplos96,@   title = 	 "Architectural Support for Programming Languages and$ 		  Operating Systems (ASPLOS-VII)",D   booktitle = 	 "Architectural Support for Programming Languages and$ 		  Operating Systems (ASPLOS-VII)",   year = 	 "1996",   key =		 "ASPLOS-VII" }o   Followups to comp.arch.t   - anton  --  K M. Anton Ertl                    Some things have to be seen to be believedIK anton@mips.complang.tuwien.ac.at Most things have to be believed to be seen 0 http://www.complang.tuwien.ac.at/anton/home.html   ------------------------------  % Date: Tue, 26 Feb 2002 08:31:16 -0500r' From: Howard S Shubs <howard@shubs.net>a Subject: Re: If you need PI....i< Message-ID: <howard-0C4E2A.08311626022002@enews.newsguy.com>  , In article <3C7B2244.22B3AB18@videotron.ca>,/  JF Mezei <jfmezei.spamnot@videotron.ca> wrote:   P > Note that on my MAC, that URL is too long and gets truncated to 65 characters. > in that case, you cans > http://202.224.192.18s  F Maybe it won't work on -your- MAC but it works fine on my Mac.  Maybe  it's your browser.     Hm.  Maybe not.k   -- t Howard S ShubsD "Run in circles, scream and shout!"  "I hope you have good backups!"' Aren't there any networked SJFs around?b   ------------------------------  % Date: Tue, 26 Feb 2002 10:39:02 +0100hB From: "Marcel van der Velden" <Marcel.vanderVelden@atosorigin.com>. Subject: inter-process communication and timerY Message-ID: <0E16861EE7BCD111BE9400805FE6841F1A80768B@c1s5x001.cor.srvfarm.origin-it.com>    Hi all,k  G I want to write an application (using 'C') which has serveral processes A (executables) using different priorities. These processes have toxG communicate with each other so I must sent messages from one process tor another.   Question 1:lG Can anybody tell me which message system I can use (TCP/IP or somethinga< else)? Which is the fastest? Can anybody sent me an example?   Question 2:eL Is there a way to trigger an application with a timer (for example every 200" milliseconds)? How must I do that?   Hope anyone can help.a   Thanks,p Marcel van der Velden " Marcel.vanderVelden@atosorigin.com   ------------------------------  % Date: Tue, 26 Feb 2002 05:23:33 -0500i- From: JF Mezei <jfmezei.spamnot@videotron.ca>w2 Subject: Re: inter-process communication and timer, Message-ID: <3C7B6221.F8BEEBF3@videotron.ca>   Marcel van der Velden wrote:I > Can anybody tell me which message system I can use (TCP/IP or somethingt> > else)? Which is the fastest? Can anybody sent me an example?   If on same node: mailboxes (MBA devices). DECNET ICCl TCPIPn) shared global sections with lock manager.   V If on separate nodes, but in a cluster, then remove the mailboxes and global sections.  O If on separate nodes, not in a cluster, then you are left with DECNET or TCPIP.     N > Is there a way to trigger an application with a timer (for example every 200$ > milliseconds)? How must I do that?     HELP SYSTEM $SETIMR  HELP SYSTEM $HIBER HELP SYSTEM $WAKEi HELP RTL LIB LIB$WAITo  X You can consult the full documentation on the vms web site http://www.openvms.compaq.com   ------------------------------    Date: 26 Feb 2002 06:10:23 -0600- From: Kilgallen@SpamCop.net (Larry Kilgallen)r2 Subject: Re: inter-process communication and timer3 Message-ID: <pf6e1oFE2D4E@eisner.encompasserve.org>.   In article <0E16861EE7BCD111BE9400805FE6841F1A80768B@c1s5x001.cor.srvfarm.origin-it.com>, "Marcel van der Velden" <Marcel.vanderVelden@atosorigin.com> writes:  I > I want to write an application (using 'C') which has serveral processes3C > (executables) using different priorities. These processes have to I > communicate with each other so I must sent messages from one process top
 > another.  F The nice thing about VMS is that System Services and Runtime LibrariesC are fairly language independent.  The answer for C will be the samee" as the answer for other languages.  I > Can anybody tell me which message system I can use (TCP/IP or something  > else)?  D Extra cost options you can use include TCP/IP, DECmessageQ (possibly< with a name change now), MQseries, ACMS and probably others.  H Bundled features you can use on a single node include DECnet, Mailboxes,E Intra-Cluster Communications, and for very small amounts of data, thec
 Lock Manager.   E Of the above, Mailboxes are the only one restricted to a single node,1G so you might want to avoid them in case later growth causes you to want 5 to run your processes on separate nodes in a cluster.1   > Which is the fastest?   B The Lock Manager is the fastest service, followed by Intra-Cluster Communications.,  ! > Can anybody sent me an example?h  D Yes.  Compaq has done that in the Programming Concepts manual on the? Documentation CD-ROM that came with your copy of VMS.  Look at:   ? 	file:///VMSDOC073/v73/5841/5841pro_contents.html#toc_chapter_3   
 > Question 2:6N > Is there a way to trigger an application with a timer (for example every 200 > milliseconds)?   Yes.   > How must I do that?   > A common method is to use the SYS$SETIMR system service.  See:  > 	file:///VMSDOC073/v73/5841/5841pro_073.html#114_timerrequests   ------------------------------    Date: 26 Feb 2002 05:41:02 -0800- From: pdafniotis@yahoo.com (Petros Dafniotis)h5 Subject: Memory Upgrade for Alpha XP-1000 Workstatione= Message-ID: <e54adf36.0202260541.287c404e@posting.google.com>M  ? I have an XP-1000 with a 667 MHz CPU and 4 SDRAMs of 128MB eachtC yielding a total of 512MB. I want to upgrade memory to max possible   without wasting RAM if possible.  B My question is what options do I have? Reading through the docs it? appears that I could add 1 GB more for a total of 1.5GB without-C wasting existing memory sticks. Please confirm and/or suggest other-B solutions. I am asking here because in Switzerland I have to use aE third party for the actual purchase and this third party has not beent- entirely reliable in information in the past.t  B Also, can I just go out and buy the SDRAM from a PC vendor? (ok, I need ECC registered SDRAM).    Thank you and kind regards,  Petros ---- Petros Dafniotis, PhD. pdafniotis@yahoo.com   ------------------------------  % Date: Tue, 26 Feb 2002 10:02:43 -05000* From: John Reagan <john.reagan@compaq.com>9 Subject: Re: Memory Upgrade for Alpha XP-1000 Workstationt) Message-ID: <3C7BA393.5040302@compaq.com>e   Petros Dafniotis wrote:>A > I have an XP-1000 with a 667 MHz CPU and 4 SDRAMs of 128MB each E > yielding a total of 512MB. I want to upgrade memory to max possibleH" > without wasting RAM if possible. > D > My question is what options do I have? Reading through the docs itA > appears that I could add 1 GB more for a total of 1.5GB withoutaE > wasting existing memory sticks. Please confirm and/or suggest other D > solutions. I am asking here because in Switzerland I have to use aG > third party for the actual purchase and this third party has not beenn/ > entirely reliable in information in the past.  > D > Also, can I just go out and buy the SDRAM from a PC vendor? (ok, I > need ECC registered SDRAM).r  I You are correct.  I have the same exact box as my desktop (I'm typing on eC it now).  I currently have 2GB RAM.  I took out the older DIMMs (I  E started with just 4 x 64M for 256MB) when I max'd it out to let some rL folks do Java coding on it.  My cost for the new memory was extremely cheap.  F As for buying "off the shelf", I can't help you there.  I bought from I Compaq of course and I don't know enough about the machine to talk about  
 other memory.d         -- u John Reaganc' Compaq Pascal/{A|I}MACRO Project Leader    ------------------------------  # Date: Tue, 26 Feb 2002 17:54:38 GMT # From: "John Smith" <a@nonymous.com>M  Subject: On 64-bit architectures/ Message-ID: <yZPe8.3571$jL.1447@news1.bloor.is>s  ? http://www.extremetech.com/article/0,3396,s=1005&a=23093,00.asph   ------------------------------  # Date: Tue, 26 Feb 2002 18:15:06 GMT # From: "John Smith" <a@nonymous.com>l$ Subject: Re: On 64-bit architectures/ Message-ID: <KgQe8.3665$jL.1989@news1.bloor.is>f  6 Perhaps a better link to this same set of articles is:  > http://www.extremetech.com/print_article/0,3428,a=23093,00.asp   ------------------------------    Date: 25 Feb 2002 23:12:30 -0800. From: mistdragon@zdnetonebox.com (mist dragon)Y Subject: Re: Oracle RDB  ABUSIVE PRICE$  was (Re: MPE/iX users form OpenMPE organization)o= Message-ID: <7500353b.0202252312.17105691@posting.google.com>a  E In fact (as ex. oracle employee), Oracle currently prices systems per4E CPU. They used to do this by megahertz where the price was multipliedfC by 1 for CISC CPU's and 1.5 for RISC CPU's. With every rising mhz'st@ the prices went thru the roof and therefore they went to per cpu pricing.  D During the last 12 years I remember, there has been a lot of schemesE of concurrent, deployment, terminal, row, etc. pricing schemes and ifeD you have an existing license that had one of those pricings, you canB still use them as Oracle does not preach contracts. However, it isD often cheaper to cancel the old license and buy new one, because the? prices have come down significantly during the years. One oftenRF overlooked, 'invisible' aspect of that is support, where support priceC for amount of users can far exceed the price of the new license andr support.  = In these days Oracle has wery simple pricing: 40.000$/CPU forr< enterprise version and 15.000$/CPU per standard version. TheC difference is scalability, expandability and amount of processors agB single system can have. If I recall correctly, the max. system youF could buy with standard edition was 4 cpu's, so your 6 cpu model falls> to enterprise edition whether you need the other extras or no.  D Oracle also has per user price, but that becomes more expensive whenF the amount of users exceeds 50 users so it is not usually worth of it.> It does not make sense to use that scheme on millions of usersF website, although I wery much suspect that the database would not take= that anyway. Most systems only handle few hundered CONCURRENTs? sessions, e.g. only wery few of web users are actually having a @ database transaction. A good website for millions of users needsD around 20-40 CPU's cluster depending on hw architecture (yes, I have; hands-on experience) and even with CPU pricing the price isf outrageous.-  D As for RDB, I believe it is true that it costs 40.000$/CPU. I do notA know if standard edition is even available for RDB. Oracle has no F interest of whatsoever selling cheap RDB when it has its own product -A it only wants RDB user base and convert the users to its flagshipuD product. However, if it would price RDB over its own product, it mayF cause riot, but as the prices are the same, there is not much that can be done.  E Now, RDB price was bundled with VMS price and VMS price with hardwareiF price. So it never was 'free'. This does not mean that I would believeF Oracle has right prices - they are wery expensive still, but I believeD it is important to notice that there is no free lunch and never was.  C That is, unless you switch from RDB to mySql, postGress or similar,cF which are open source/GNU based SQL databases on linux/unix. There mayE be other similar products for VMS that I am not aware of. Having used B mySql for example, I have noticed it to be wery similar to Oracle,E only that the huge amount of fat it had, has been sliced off and onlyfC the pure sql server is left. It is small, efficient and easy to usedE and supports easily those huge amounts of users. The downside is thate? since it is not commercial per se, you need to purchase supporttC separately, but as overall, it is of an excelent value. And you geteC source code for it so if you feel adventurous and VMS port does not  exist, you can make one :O)n   ------------------------------  % Date: Tue, 26 Feb 2002 13:54:14 +0100i$ From: "Dr. Dweeb" <Dweeb@NoSpam.com>Y Subject: Re: Oracle RDB  ABUSIVE PRICE$  was (Re: MPE/iX users form OpenMPE organization)s/ Message-ID: <nzLe8.80$6l2.2125@news.get2net.dk>r   Clip ...  E > That is, unless you switch from RDB to mySql, postGress or similar,>H > which are open source/GNU based SQL databases on linux/unix. There mayG > be other similar products for VMS that I am not aware of. Having usediD > mySql for example, I have noticed it to be wery similar to Oracle,G > only that the huge amount of fat it had, has been sliced off and onlytE > the pure sql server is left. It is small, efficient and easy to use-G > and supports easily those huge amounts of users. The downside is thatcA > since it is not commercial per se, you need to purchase supportlE > separately, but as overall, it is of an excelent value. And you geteE > source code for it so if you feel adventurous and VMS port does notn > exist, you can make one :O)   J MySQL (and I do not know about the others mentioned above) is a trival toyI database about as far removed from Rdb as a Trabant is from a Ferrari.  IuJ would suggest you read the MySQL manual.  It demonstrates clearly that theG authors do not have a clue about database systems per se, or relationaloJ databases in particular.  It is a glorified file system with a limited SQL access facility, nothing more.   Dweeb.   ------------------------------    Date: 26 Feb 2002 08:13:38 -0600- From: Kilgallen@SpamCop.net (Larry Kilgallen)  Subject: Re: Please help3 Message-ID: <9mdV1wPNwHQ3@eisner.encompasserve.org>s  c In article <20020226085924.05765.00000728@mb-cl.aol.com>, lilmjaraujo@aol.com (Lilmjaraujo) writes:e  H > If your company's IT department uses VMS, please could you mail me theL > contact details of the manager. Don't worry, I will not SPAM them I'm just > looking for work.   E If the manager has not requested such inquiries, email in that regardkD would be spam.  Of course responses providing "contact details" need not provide an email address.d   ------------------------------   Date: 26 Feb 2002 13:59:24 GMT' From: lilmjaraujo@aol.com (Lilmjaraujo)  Subject: Please help9 Message-ID: <20020226085924.05765.00000728@mb-cl.aol.com>    Hi, F If your company's IT department uses VMS, please could you mail me theJ contact details of the manager. Don't worry, I will not SPAM them I'm just looking for work.n Thanks.l P.S.E Any contact details for other companies would be greatly appreciated.e   ------------------------------  % Date: Tue, 26 Feb 2002 14:37:33 +0100 = From: Arne =?iso-8859-1?Q?Vajh=F8j?= <arne.vajhoej@gtech.com>a# Subject: Re: Powerpoint competitionr) Message-ID: <3C7B8F9D.D9385A4B@gtech.com>u   JF Mezei wrote: O > However, to download it, one needs "cvs" to the host toget the files. What isb > "CVS " ?????  > CVS is a source control system like CMS, SCCS, RCS, PVCS etc..  9 CVS is actually the dominant source control system in thee internet freeware world !e  ; You can get the CVS client for VMS, but not the CVS server.    Arne   ------------------------------  % Date: Tue, 26 Feb 2002 13:08:40 -0500a0 From: Paul Anderson <paul.r.anderson@compaq.com>! Subject: Re: PrintServer software ; Message-ID: <260220021308407473%paul.r.anderson@compaq.com>e  
 In articleD <18560B9BCCE902439194CA5C0AA9963F55F931@SGSINSMAPXMS10.SONYE2K.COM>,6 Quek, Hong Cheang <Hongcheang.Quek@ap.sony.com> wrote:  L > I need the printserver for solaris 7. I went to the link that you gave butO > it turned out to be a broken link. Searched the internet and found this site:V > ftp.digital.comoL > It does have the binaries for Solaris, but somehow cannot seem to download > the files. Advice please. :)  D I could get to kits for both Solaris and SunOS at the URL I posted.    The FTP locations are:  H <ftp.digital.com/pub/Digital/Alpha/printer/printserver/bsd/lps52sol.tar>H <ftp.digital.com/pub/Digital/Alpha/printer/printserver/bsd/lps52sun.tar>  D I've copied both tar files, so if you still have trouble downloading@ them, let me know which one you want and I can e-mail it to you.   Paul   -- e  Paul Anderson   OpenVMS Engineeringr   Compaq Computer Corporationb   ------------------------------    Date: 26 Feb 2002 11:06:38 +0100* From: eplan@kapsch.net (Peter LANGSTOEGER)2 Subject: Re: Profitability and the survival of VMS* Message-ID: <3c7b5e2e$1@news.kapsch.co.at>  R In article <3C7ABCB0.5D885E70@spam.net>, Brass Christof <welcome@spam.net> writes:C >In Switzerland the situation was very different as the DEC people o? >stayed in business. DEC was much stronger in Switzerland than o@ >Compaq and the buildings of DEC were much better and at better @ >places. In fact Compaq recently moved even their European head ? >quater from Munich (Germany) to Zurich (Switzerland) into the  B >very same buildings that were DEC once. The area is still called : >DECpark by most people I know although they replaced the ? >d|i|g|i|t|a|l logos all over the place by some Compaq writing.n; >The DEC call centre continued to exist and it is still no  1 >problem to get in contact with the right people.a >s- >But honestly I regard this as the exception.e  . The same 'exception' happened here in Austria.L ~50 Q vs ~500 DEC lead to integration of Q into DEC (but as the bad mgmt/mktK behaviour started in DEC, Q was no disadvantage - but no advantage either).n  . HP and Q would be a totally different story...   -- a< Peter "EPLAN" LANGSTOEGER           Tel.    +43 1 81111 2651; Network and OpenVMS system manager  Fax.    +43 1 81111 888 < KAPSCH AG      Wagenseilgasse 1     E-mail  eplan@kapsch.netH A-1121 VIENNA  AUSTRIA              "I'm not a pessimist, I'm a realist"   ------------------------------  % Date: Tue, 26 Feb 2002 17:50:10 -0000 - From: wspencer@ap.nospam.org (Warren Spencer)d& Subject: Re: Question for Backup Gurus7 Message-ID: <91C180F74warrenspencer1977@209.249.90.100>-  . p14175@email.sps.mot.com (Linda Luik) wrote in' <3C72A8CF.B91D0DCF@email.sps.mot.com>: l  
 -- snip --J >-------------------------------------------------------------------------E >-------------------------- The node was rebooted to clear the RWAST.i
 -- snip --  H I once cleared a tape drive/backup process hung in RWAST by cycling the  power on the tape drive.   ws -- m   Warren Spencer' Senior Software Engineer (not a writer)s The Associated Press  < ** Time flies like an arrow.  Fruit flies like a bananna. **   ------------------------------    Date: 26 Feb 2002 00:55:10 -0800, From: tony.cheung@asiayeah.com (Tony Cheung)) Subject: select() does not work for pipese= Message-ID: <f9dc0a5a.0202260055.2f3257f1@posting.google.com>,   Hi All,2  ? I just found out the select() does not work on file descriptorsu> created by pipe(). It returns an error of "socket operation on non-socket".   Is it the case?   > Basically, my process has two threads, one thread looping on aF select() call. I would like to be able to interrupt that select() callE from another thread. Originally I am planning to use a pipe to signalc that, but it does not work.h   Is there any other workaround?   Thank you very much.   Tony Cheunga   ------------------------------    Date: 26 Feb 2002 05:55:14 -0600- From: Kilgallen@SpamCop.net (Larry Kilgallen) - Subject: Re: select() does not work for pipes 3 Message-ID: <+du12bKHdKp3@eisner.encompasserve.org>y  l In article <f9dc0a5a.0202260055.2f3257f1@posting.google.com>, tony.cheung@asiayeah.com (Tony Cheung) writes:  A > I just found out the select() does not work on file descriptors @ > created by pipe(). It returns an error of "socket operation on > non-socket". >  > Is it the case?3  A Previous posts in this newsgroup have indicated that construct isi only for TCP/IP sockets.  @ > Basically, my process has two threads, one thread looping on aH > select() call. I would like to be able to interrupt that select() callG > from another thread. Originally I am planning to use a pipe to signalo > that, but it does not work.  >   > Is there any other workaround?  : Sorry, I did not understand what the first workaround was.   ------------------------------  % Date: Tue, 26 Feb 2002 08:13:15 -0500s4 From: John Malmberg <Malmberg@dskwld.zko.dec.compaq>- Subject: Re: select() does not work for pipese4 Message-ID: <3C7B89EB.4050809@dskwld.zko.dec.compaq>   Tony Cheung wrote:  	 > Hi All,a > A > I just found out the select() does not work on file descriptorsm@ > created by pipe(). It returns an error of "socket operation on > non-socket". >  > Is it the case?(    ) It is currently a documented restriction.M     @ > Basically, my process has two threads, one thread looping on aH > select() call. I would like to be able to interrupt that select() callG > from another thread. Originally I am planning to use a pipe to signalh > that, but it does not work.l >   > Is there any other workaround?    8 A. Use the SYS$QIO version of the TCP/IP calls.  See the+     documentation for TCP/IP from a link ath$       http://www.openvms.compaq.com.  ?     Also see the documentation for DECC$GET_SDC in the Compaq Cs      documentation from a link at/       http://www.openvms.compaq.com/commercial/a   B. (I have not tried this)>     Use the DECC$GET_SDC to obtain the OpenVMS channel for the@     socket.  Then use the SYS$CANCEL to abort the I/O.  This may?     cause the I/O the select() call is waiting for to complete,w      possibly with an error code.   C. (I have not tried this)?     Instead of a pipe, create and bind a pair of sockets to theR6     127.0.0.1 address, and have select() wait on that.  >     This also may be usable for applications being ported that=     expect to be able to use a select() statement to wait form:     terminal input.  Use ASTs and the SYS$QIO interface to6     obtain the data, send it to the local socket pair.   -Johnn malmberg@dskwld.zko.dec.compaq Personal Opinion Onlya   ------------------------------    Date: 26 Feb 2002 07:59:20 -0800, From: tony.cheung@asiayeah.com (Tony Cheung)- Subject: Re: select() does not work for pipeso= Message-ID: <f9dc0a5a.0202260759.28b866a0@posting.google.com>1  
 Thank you!  h Kilgallen@SpamCop.net (Larry Kilgallen) wrote in message news:<+du12bKHdKp3@eisner.encompasserve.org>...n > In article <f9dc0a5a.0202260055.2f3257f1@posting.google.com>, tony.cheung@asiayeah.com (Tony Cheung) writes: > C > > I just found out the select() does not work on file descriptorsaB > > created by pipe(). It returns an error of "socket operation on > > non-socket". > >  > > Is it the case?g > C > Previous posts in this newsgroup have indicated that construct is- > only for TCP/IP sockets.  E So is any other other call that could wait on both TCP/IP sockets andc> a pipe or even other file descriptors concurrently in OpenVMS?   > B > > Basically, my process has two threads, one thread looping on aJ > > select() call. I would like to be able to interrupt that select() callI > > from another thread. Originally I am planning to use a pipe to signalh > > that, but it does not work.  > > " > > Is there any other workaround? > < > Sorry, I did not understand what the first workaround was.  D I was about to create a pipe. The select() call waits on the pipe asB well as other TCP/IP sockets. Whenever the another thread wants toD wake up the thread which is waiting in the select() call, the threadD simply writes a single byte to the pipe. The select() call will then return.a  A Thus, the thread will be able to wait on the TCP/IP sockets whiler3 being able to be interrupted by an external thread.o  @ Howver, since select() cannot wait on a pipe. Is there any otherC workaround? I simply want to have a thread that is able to wait fort2 external interrupt as well as some TCP/IP sockets.  
 Thank you.   Tony Cheungv   ------------------------------  # Date: Tue, 26 Feb 2002 16:32:05 GMTrL From: winston@SSRL.SLAC.STANFORD.EDU ("Alan Winston - SSRL Admin Cmptg Mgr")- Subject: Re: select() does not work for pipest8 Message-ID: <00A0A228.88D61237@SSRL04.SLAC.STANFORD.EDU>   Tony Cheung wrote:  E >I was about to create a pipe. The select() call waits on the pipe aseC >well as other TCP/IP sockets. Whenever the another thread wants toOE >wake up the thread which is waiting in the select() call, the threadsE >simply writes a single byte to the pipe. The select() call will then  >return. >-B >Thus, the thread will be able to wait on the TCP/IP sockets while4 >being able to be interrupted by an external thread. >eA >Howver, since select() cannot wait on a pipe. Is there any other D >workaround? I simply want to have a thread that is able to wait for3 >external interrupt as well as some TCP/IP sockets.-  J You need to investigate ASTs (Asynchronous System Traps).  There are otherJ people on this newsgroup who can freehand the code better than I can, but J basically a process can hang out one or more ASTs - in this case, doing a H QIO read - and then go about its business (or go to sleep).  When the IOJ completes, the AST fires, and control goes to an interrupt service routineI that you've coded.  This can then set a flag that tells the main body of iH the program what to do and retires.  The main body, awake, does whatever it's supposed to do.       Roughlye  1        QIO #1 (specifying AST service routine #1)h1        QIO #2 (specifying AST service routine #2).          (forever) {           SYS$HIBERi/     --   If we got here, one of the ASTs fired.n            Was it #1?cH                Do the right thing, and hang another QIO on that channel.            Was it #2? I                Do the right thing, and hang another QIO on that channel.             }o              AST service routine #1               set flag #1e              AST service routine #2               set flag #2s  G And you'd probably use a mailbox rather than a pipe.  If you have to doPN different things depending on the input, you have to code so that each 'thing', can be successfully interrupted and resumed.  N This has the advantage over sockets that you don't have to waste time polling.   -- Alani    O ===============================================================================t0  Alan Winston --- WINSTON@SSRL.SLAC.STANFORD.EDUM  Disclaimer: I speak only for myself, not SLAC or SSRL   Phone:  650/926-3056 M  Physical mail to: SSRL -- SLAC BIN 69, PO BOX 4349, STANFORD, CA  94309-0210sO ===============================================================================e   ------------------------------    Date: 26 Feb 2002 11:24:24 -0600- From: Kilgallen@SpamCop.net (Larry Kilgallen)w- Subject: Re: select() does not work for pipest3 Message-ID: <J5wO3wzlYCj1@eisner.encompasserve.org>s  l In article <f9dc0a5a.0202260759.28b866a0@posting.google.com>, tony.cheung@asiayeah.com (Tony Cheung) writes: > Thank you! > j > Kilgallen@SpamCop.net (Larry Kilgallen) wrote in message news:<+du12bKHdKp3@eisner.encompasserve.org>...o >> In article <f9dc0a5a.0202260055.2f3257f1@posting.google.com>, tony.cheung@asiayeah.com (Tony Cheung) writes:o >> -D >> > I just found out the select() does not work on file descriptorsC >> > created by pipe(). It returns an error of "socket operation on  >> > non-socket".t >> > e >> > Is it the case? >> tD >> Previous posts in this newsgroup have indicated that construct is >> only for TCP/IP sockets.  > G > So is any other other call that could wait on both TCP/IP sockets ande@ > a pipe or even other file descriptors concurrently in OpenVMS?  B Use the $QIO interface to TCP/IP rather than the socket interface.   ------------------------------  % Date: Tue, 26 Feb 2002 18:04:36 +0100c$ From: "Dr. Dweeb" <Dweeb@NoSpam.com>" Subject: Re: Shannon on the merger0 Message-ID: <4ePe8.143$6l2.3373@news.get2net.dk>  J "Wayne Sewell" <wayne@tachysoft.xxx.120519.killspam.00c7> wrote in message( news:XgdE$1Pr04vc@tachxxsoftxxconsult...= > In article <3C79AB7D.781F4CC2@fsi.net>, "David J. Dachtera". <djesys.nospam@fsi.net> writes:t > > "Terry C. Shannon" wrote:s > >>3 > >> "John Smith" <a@nonymous.com> wrote in message - > >> news:hd6e8.407$1h5.386@news2.bloor.is...pL > >> > If the merger is voted down, it's important that HP shareholders vote NO,cL > >> > and Comapq shareholders vote yes. In this way, the deal 'breakup fee' ofL > >> > $675 million gets paid to Compaq by HP, and not the other way around. > >> >J > >> > Now if they'd only use that $675 million to properly market OpenVMS > >> instead4 > >> > of paying themselves fat bonuses for nothing. > >> > > >> > >> NOW *THAT'S* A CONCEPT!!! > >>K > >> CPQ currently spends close to $300M USD per year on "marketing." Maybe  they. > >> should outsource the whole wretched mess. > >eL > > Of course, the outsource provider would have to be some entity the Q canK > > control. Otherwise, they might end up getting effective exposure, mightyJ > > end up selling a good bit more VMS, and might end up making more money > > off of VMS there for.s > >l! > > Can't have that, now, can we?  > >r" > >> Of course, the outsourcing ofI > >> Accounts Payable and Accounts Receivable has been an utter disaster.p Trust  > >> me on this one! > > 1 > > I trust no one *REALLY* expected otherwise...e > >oG > >> What CPQ really needs is some new blood on the Board, not just thes	 same-old,mI > >> same-old candidates selected by the Houston Politburo. The Politburoi that$ > >> brought you Director Ken Lay... > >mK > > I'd be tempted to go run "Blazing Saddles" so I could quote a good linenI > > from Hedly LaMarr... (ala "My mind is raging torrent overflowing withtJ > > swirling nords of thought, an avalanche of creative alternatives.", or > > something like that...)t > >L >cK > Then I can play the Taggert (Slim Pickens) part and say "Ditto!"  (Or wasa the63 > reponse to that the "20-dollar whore" line?)  :-)   E IIRC it was "a mouth prettier than a $20 whore", "Ditto" was in therei% somewhere as well.  Very funny scene.<   Dweeb.  L ============================================================================ ===f: > Wayne Sewell, Tachyon Software Consulting  (281)812-0738 wayne@tachysoft.xxxe: > http://www.tachysoft.xxx/www/tachyon.html and wayne.htmlH > change .xxx to .com in addresses above, assuming you are not a spambot :-)e >lL ============================================================================ ===e? > Society Lady:  Are you familiar with the Great Wall of China?-7 >        Curly:  No, but I know a big fence in Chicago!o   ------------------------------  % Date: Tue, 26 Feb 2002 11:04:24 +00006( From: Nic Clews <sendspamhere@127.0.0.1>, Subject: Re: Shared SCSI Bus Cluster and TCQ) Message-ID: <3C7B6BB8.DC082AF9@127.0.0.1>m   Frederik Meerwaldt wrote:o >  > Hello there, > J > I have a question regarding the shared SCSI bus cluster functionality of
 > OpenVMS.G > I'm trying to create a shared SCSI bus cluster between 2 old DEC 20000 > model 300.7 > One controller is set to ID 7, the other one to ID 6.g> > The Disk Array in the middle consists of 6 HDDs from ID 0-5.  G We SCSI clustered betwqeen an Alphaserver 800 and a DEC 3000-200 at VMS, 7.1.  F Due to the config of the 3000-200 we bought an additional differentialD SCSI card purely for the shared disks in a storageworks shelf with aG DWZZ-B-va (?), and hooked that back into appropriate controller for the E '800. We got this data from the Systems and Options and it seemed ther< only [supported] way to do it, and it works absolutely fine.  F The key was the differential SCSI connection and supporting devices. I< don't think other controllers are supported on the 3000-200.   Hope this helps. -- m( Regards, Nic Clews CSC Computer Sciences nclews at csc dot comA   ------------------------------    Date: 26 Feb 2002 07:27:27 -0600- From: Kilgallen@SpamCop.net (Larry Kilgallen)7, Subject: Re: Shared SCSI Bus Cluster and TCQ3 Message-ID: <nk7Ng3ePCJzc@eisner.encompasserve.org>a   In article <rdeininger-2602020825500001@1cust11.tnt2.nashua.nh.da.uu.net>, rdeininger@mindspring.com (Robert Deininger) writes:.   > What is a DEC 3000-200?   H The server variant of the AXP PC 150, the first Alpha to run Windows NT.   ------------------------------  % Date: Tue, 26 Feb 2002 08:25:50 -0500s2 From: rdeininger@mindspring.com (Robert Deininger), Subject: Re: Shared SCSI Bus Cluster and TCQJ Message-ID: <rdeininger-2602020825500001@1cust11.tnt2.nashua.nh.da.uu.net>  3 In article <3C7B6BB8.DC082AF9@127.0.0.1>, Nic Clewsl <sendspamhere@127.0.0.1> wrote:M    H >We SCSI clustered betwqeen an Alphaserver 800 and a DEC 3000-200 at VMS >7.1.a > G >Due to the config of the 3000-200 we bought an additional differentialuE >SCSI card purely for the shared disks in a storageworks shelf with a$H >DWZZ-B-va (?), and hooked that back into appropriate controller for theF >'800. We got this data from the Systems and Options and it seemed the= >only [supported] way to do it, and it works absolutely fine.c >tG >The key was the differential SCSI connection and supporting devices. Ib= >don't think other controllers are supported on the 3000-200.e   What is a DEC 3000-200?a   ------------------------------    Date: 26 Feb 2002 08:11:28 -0600- From: Kilgallen@SpamCop.net (Larry Kilgallen)i, Subject: Re: Shared SCSI Bus Cluster and TCQ3 Message-ID: <1K+OhMX8UKSo@eisner.encompasserve.org>e  c In article <3C7B8F0D.90BDE193@aaa.com>, Jan-Erik =?iso-8859-1?Q?S=F6derholm?= <aaa@aaa.com> writes:o% > Wasn't that DEC 2000 modell 300" ??e   I stand corrected.  B > The DEC 3000 came as modell 600 ("desktop") and 800 ("deskside")  , And 400, 500, 700, 900, 300, 300L and 300LX.   > Jan-Erik Sderholm.t >  > Larry Kilgallen wrote: >> r >> In article <rdeininger-2602020825500001@1cust11.tnt2.nashua.nh.da.uu.net>, rdeininger@mindspring.com (Robert Deininger) writes: >> A >> > What is a DEC 3000-200? >> 1K >> The server variant of the AXP PC 150, the first Alpha to run Windows NT.H   ------------------------------  % Date: Tue, 26 Feb 2002 14:35:09 +0100R9 From: Jan-Erik =?iso-8859-1?Q?S=F6derholm?= <aaa@aaa.com>-, Subject: Re: Shared SCSI Bus Cluster and TCQ' Message-ID: <3C7B8F0D.90BDE193@aaa.com>.  # Wasn't that DEC 2000 modell 300" ??R  @ The DEC 3000 came as modell 600 ("desktop") and 800 ("deskside")   Jan-Erik Sderholm.1   Larry Kilgallen wrote: >  > In article <rdeininger-2602020825500001@1cust11.tnt2.nashua.nh.da.uu.net>, rdeininger@mindspring.com (Robert Deininger) writes:- >  > > What is a DEC 3000-200?- > J > The server variant of the AXP PC 150, the first Alpha to run Windows NT.   ------------------------------  % Date: Tue, 26 Feb 2002 15:24:48 +0100t9 From: Jan-Erik =?iso-8859-1?Q?S=F6derholm?= <aaa@aaa.com>n, Subject: Re: Shared SCSI Bus Cluster and TCQ' Message-ID: <3C7B9AB0.4047AD01@aaa.com>e   But no 200, right ? 	 Jan-Erik.-   Larry Kilgallen wrote: > e > In article <3C7B8F0D.90BDE193@aaa.com>, Jan-Erik =?iso-8859-1?Q?S=F6derholm?= <aaa@aaa.com> writes:  > >iD > > The DEC 3000 came as modell 600 ("desktop") and 800 ("deskside") > . > And 400, 500, 700, 900, 300, 300L and 300LX. >r   ------------------------------  % Date: Tue, 26 Feb 2002 15:32:34 +0000p( From: Nic Clews <sendspamhere@127.0.0.1>, Subject: Re: Shared SCSI Bus Cluster and TCQ( Message-ID: <3C7BAA92.C8AED7A@127.0.0.1>   Robert Deininger wrote:e > 5 > In article <3C7B6BB8.DC082AF9@127.0.0.1>, Nic Clews ! > <sendspamhere@127.0.0.1> wrote:c > J > >We SCSI clustered betwqeen an Alphaserver 800 and a DEC 3000-200 at VMSC                                                        ------------6 > >7.1.  >5 > What is a DEC 3000-200?@   Typo typo..... Sorry.l  3 DEC 3000 model 400. Desktop box. (And an Alpha 800)e  > I doubt you can get the right controller for the Jensen system2 (3000-200) EISA but I don't know I haven't looked. -- e( Regards, Nic Clews CSC Computer Sciences nclews at csc dot coms   ------------------------------  % Date: Tue, 26 Feb 2002 15:36:25 +0000 ( From: Nic Clews <sendspamhere@127.0.0.1>, Subject: Re: Shared SCSI Bus Cluster and TCQ) Message-ID: <3C7BAB79.8786FECF@127.0.0.1>n   Done it again.   Nic Clews wrote: >  > > What is a DEC 3000-200?  >  > Typo typo..... Sorry.n  @ > I doubt you can get the right controller for the Jensen system4 > (3000-200) EISA but I don't know I haven't looked.    --------p  9 2000-300. Maybe I should have stayed in bed this morning.e -- o( Regards, Nic Clews CSC Computer Sciences nclews at csc dot comr   ------------------------------  % Date: Tue, 26 Feb 2002 16:39:17 +0100:9 From: Jan-Erik =?iso-8859-1?Q?S=F6derholm?= <aaa@aaa.com>c, Subject: Re: Shared SCSI Bus Cluster and TCQ' Message-ID: <3C7BAC25.82759782@aaa.com>s  ! Just that Jensen was 2000-300 :-)r	 Jan-Erik.i   Nic Clews wrote: > @ > I doubt you can get the right controller for the Jensen system4 > (3000-200) EISA but I don't know I haven't looked. > --   ------------------------------    Date: 26 Feb 2002 05:43:11 -0600- From: Kilgallen@SpamCop.net (Larry Kilgallen) 0 Subject: Re: SHOW ENTRY /FULL doesn't show /PARA3 Message-ID: <bG9Da6DbDhOm@eisner.encompasserve.org>(  W In article <25FEB200223442139@gerg.tamu.edu>, carl@gerg.tamu.edu (Carl Perkins) writes:u  G > You can use DCPS and specify the desired parameters in the setup .COMSH > file. DCPS sets them up as logical names like DCPS$queuename_PARAMETER* > which the DCPS symbiont checks and uses.  C Since that is outside the scope of the queue manager, it thus makesm* sense that SHOW QUEUE would not report it.   ------------------------------    Date: 26 Feb 2002 13:52:16 +0100& From: "Ketil Malde" <ketil+@ii.uib.no> Subject: Re: Singular they. Message-ID: <eg1yf8csn3.fsf@sefirot.ii.uib.no>  6 Bob Cousins <bob@$NOSPAM$lintilla.demon.co.uk> writes:   > Ken Hagan wrote:  , >>>     Everyone took their coats with them.  ( >> "Everyone took their coat with them"?  3 > Jeez, whomever cares! English is a dumb language..  E I won't quote Dickens or that crowd, but there are *some* that managelC to speak of persons in an efficient and natural, yet gender-neutrals manner:   9         "We hates it, my precious!  We hates it forever."Y :-)    -kzm --  H If I haven't seen further, it is by standing in the footprints of giants   ------------------------------  % Date: Tue, 26 Feb 2002 09:12:57 -0500-* From: WILLIAM WEBB <WWEBB1@email.usps.gov>E Subject: RE: Singular they was Re: Sv: Younger recruits versus experin- Message-ID: <0033000054377119000002L092*@MHS>6  ? =0AThe parable of the loaves and fishes is not exactly obscure.o   WWWebb   -----Original Message-----/ From: Info-VAX-Request@Mvb.Saic.Com at INTERNET ) Sent: Tuesday, February 26, 2002 12:04 PMpB To: Webb, William W Raleigh, NC; Info-VAX@Mvb.Saic.Com at INTERNETE Subject: RE: Singular they was Re: Sv: Younger recruits versus experia    H In article <Jbl6KmZJoCQV@eisner.encompasserve.org>, koehler@encompasser= ve.orgH (Bob Koehler) writes... }In article <01KEMGOK42CI0058C8@tgmail.tg.nsw.g= ov.au>, / paddy.o'brien@zzz.tg.nsw.gov.au writes: }> Bob, - }>>      Fish is only singular in New Jersey.o }>- }> Is this some strange US attempt at humour?e }iD }   Everything in the state of New Jersey is strange.  Even the good }   things.e }rH }   As we moved around the country when I was a child, I was treated to=  aC }   great many variations on what was condidered education.  But ituE }   wasn't until we moved to NJ that someone tried to teach my sistere! }   fishes as the plural of fish.r  A If it wasn't, it wouldn't be possible to have someone "sleep with  the fishes". This, it must be.  	 --- Carl=t   ------------------------------    Date: 26 Feb 2002 08:18:48 -0600- From: koehler@encompasserve.org (Bob Koehler)nU Subject: Re: Singular they was Re: Sv: Younger recruits versus experienced veterans (a3 Message-ID: <4II2UF2taJXZ@eisner.encompasserve.org>l  W In article <25FEB200222520615@gerg.tamu.edu>, carl@gerg.tamu.edu (Carl Perkins) writes:o  C > If it wasn't, it wouldn't be possible to have someone "sleep witho  > the fishes". This, it must be.  E    It's possible they aint misbehavin'.  In the meantime I've broughtI4    home many fish without needing an extra syllable.   ------------------------------  % Date: Tue, 26 Feb 2002 05:29:22 -0500@- From: JF Mezei <jfmezei.spamnot@videotron.ca>m- Subject: TCPIP throuput for a VAXstation 3100w, Message-ID: <3C7B637E.BB1041A9@videotron.ca>   Got a vaxstation 3100-30.   K Tried to have the web server (OSU) serve one of those microsoft media files N (.wmv) (an 80 meg baby containing a star trek episode) to my mac that is right) next to the vax on the thinwire ethernet.   M The file was encoded at 239kbps. But with the current windows player, it justeH buffers about 4% of the file and then quits. With a previous version, itL streams the file correctly, but no image. (not sure if due to lost frames or; just incompatible video compression, however sound was ok).h  K Before I investigate further, is 239kbps something I should not ever expecttF from a vaxstation 3100-30 or would that be well within the real of its capability ?   ------------------------------  % Date: Tue, 26 Feb 2002 13:50:39 +0000p% From: Alan Greig <a.greig@virgin.net>I1 Subject: Re: TCPIP throuput for a VAXstation 3100 8 Message-ID: <es3n7uobt273snls9g72u379jij4rjmvhv@4ax.com>  , On Tue, 26 Feb 2002 05:29:22 -0500, JF Mezei% <jfmezei.spamnot@videotron.ca> wrote:o   >Got a vaxstation 3100-30. >uL >Tried to have the web server (OSU) serve one of those microsoft media filesO >(.wmv) (an 80 meg baby containing a star trek episode) to my mac that is rights* >next to the vax on the thinwire ethernet.  D I am beginning to think that Star Trek: Enterprise is the only thingF keeping ISPs afloat since Napster was taken offline! 293kb/s should be2 well within the capability of the vax/mac/ethernet  N >The file was encoded at 239kbps. But with the current windows player, it justI >buffers about 4% of the file and then quits. With a previous version, itnM >streams the file correctly, but no image. (not sure if due to lost frames or < >just incompatible video compression, however sound was ok).  E If this is the same encoding of early episodes of Enterprise as foundeE on Kazaa then  you need to do some detective work to find appropriatenA versions of decoders. I'm not enough into copyright violations toeD track this down myself but that's probably what you will have to do.D The pirate copies required a hacked decoder - at least for the earlyB episodes. If this decoder contacts FACT/FAST and/or reformats your hard-drive then tough!  E I know someone who has claimed to have found all the right bits so itL
 is doable.     >3L >Before I investigate further, is 239kbps something I should not ever expectG >from a vaxstation 3100-30 or would that be well within the real of its=
 >capability ?    -- Alan   ------------------------------  % Date: Tue, 26 Feb 2002 15:16:37 +0000p' From: Elliott Roper <elliott@yrl.co.uk>:1 Subject: Re: TCPIP throuput for a VAXstation 3100w2 Message-ID: <260220021516373195%elliott@yrl.co.uk>  5 In article <3C7B637E.BB1041A9@videotron.ca>, JF Mezeie% <jfmezei.spamnot@videotron.ca> wrote:    > Got a vaxstation 3100-30.p > M > Tried to have the web server (OSU) serve one of those microsoft media fileseP > (.wmv) (an 80 meg baby containing a star trek episode) to my mac that is right+ > next to the vax on the thinwire ethernet.- > O > The file was encoded at 239kbps. But with the current windows player, it justbJ > buffers about 4% of the file and then quits. With a previous version, itN > streams the file correctly, but no image. (not sure if due to lost frames or= > just incompatible video compression, however sound was ok).o > M > Before I investigate further, is 239kbps something I should not ever expectvH > from a vaxstation 3100-30 or would that be well within the real of its > capability ?  @ We just happen to have a VLC, which is not that different from aF 3100-30, running OSU here on thinwire. I can push a quicktime movie to( it with FTP from my Mac at 340KBytes/sec@ OSU serves it back to the Mac at speeds varying from 89KB/sec toG 30KB/sec for reasons I don't understand. The VLC shows the relevant OSUP= process using between 1% and 10% of CPU during the transfer. s  G (Ours might be slower, we have is a linux firewall standing between theCA Mac and the VLC with a different piece of ethernet in each of itsi hands.)   E It looks like the way we have OSU set up is one bottleneck, since the F VLC will happily ftp it back at 230 KBytes/sec with UCX barely raisingF a sweat. The firewall/ethernet seems to be another, because a parallelD ftp transfer from another web server, an alphastation 200/4, via the9 same route almost halves the rate to 140KB/sec from each.o  F The alpha serves the same file via OSU at speeds varying from 90KB/sec to 80KB/sec   = So you should be fine, with a possible need for OSU preening.>   ------------------------------  % Date: Tue, 26 Feb 2002 13:29:32 -0500h- From: JF Mezei <jfmezei.spamnot@videotron.ca>s1 Subject: Re: TCPIP throuput for a VAXstation 3100l, Message-ID: <3C7BD40B.E2DDFFF2@videotron.ca>   Alan Greig wrote:gH > keeping ISPs afloat since Napster was taken offline! 293kb/s should be4 > well within the capability of the vax/mac/ethernet  J Thanks.  may look into it further. ( I deleted the wmv file because it was taking up too much disk space)  G > If this is the same encoding of early episodes of Enterprise as foundsG > on Kazaa then  you need to do some detective work to find appropriater > versions of decoders.   M Seemed to work with the version 7 of the window media player. I tend to avoids5 that software and prefer real audio out of principle.h   ------------------------------  % Date: Tue, 26 Feb 2002 13:35:35 -0500t- From: JF Mezei <jfmezei.spamnot@videotron.ca>u1 Subject: Re: TCPIP throuput for a VAXstation 3100 , Message-ID: <3C7BD575.4689747C@videotron.ca>   Elliott Roper wrote:* > it with FTP from my Mac at 340KBytes/secB > OSU serves it back to the Mac at speeds varying from 89KB/sec to* > 30KB/sec for reasons I don't understand.  F Yeah I had a feeling it was the web server that was eating up a lot ofL resources. You'd think that once it has begun to spew the contents of a fileL that it wouldjust copy it out very efficiently, but perhaps there are issues> with the cache since the cache can't cop with such a big file.   ------------------------------  % Date: Tue, 26 Feb 2002 09:16:01 +0100/9 From: Jan-Erik =?iso-8859-1?Q?S=F6derholm?= <aaa@aaa.com>e/ Subject: Re: tool to calc recursive dir sises ?e' Message-ID: <3C7B4441.36E7934D@aaa.com>=  4 Sort of, but it dosn't calculate the total att every7 "node" in the directory tree on the DISK. Just the dirse under the [000000] level.   	 Jan-Erik.b   Don Sykes wrote: >  > You mean like :- >         $ LOOP:-2 >         $ check = f$search("DISK:[000000]*.dir")& >         $ if check.eqs.""  then exit6 >         $ check = check -"000000]" -".DIR;1" +"...]"! >         $ dir/siz/grand 'CHECK'r >         $ goto LOOPo >    ------------------------------  % Date: Tue, 26 Feb 2002 09:12:11 -0500s* From: WILLIAM WEBB <WWEBB1@email.usps.gov>/ Subject: RE: tool to calc recursive dir sises ?o- Message-ID: <0033000054376929000002L092*@MHS>   6 =0AI remember some DCL called TREE.COM which created a8 graphical representation of the directory structure of a disk on a VT screen display.  ; Study how the logic works in this, substitute DIRs with thee< appropriate qualifiers for the line drawing character parts,0 and you might just be able to get what you want.  > I'd have responded to your inquiry earlier, but was (and still< am confused) about exactly how you wanted that totalling and subtotalling to work.    WWWebb   -----Original Message-----/ From: Info-VAX-Request@Mvb.Saic.Com at INTERNETi( Sent: Tuesday, February 26, 2002 3:26 AMB To: Webb, William W Raleigh, NC; Info-VAX@Mvb.Saic.Com at INTERNET/ Subject: RE: tool to calc recursive dir sises ?m    4 Sort of, but it dosn't calculate the total att every7 "node" in the directory tree on the DISK. Just the dirs  under the [000000] level.y  	 Jan-Erik.    Don Sykes wrote: >I > You mean like :Q >         $ LOOP:g4 >         $ check =3D f$search("DISK:[000000]*.dir")& >         $ if check.eqs.""  then exit8 >         $ check =3D check -"000000]" -".DIR;1" +"...]"! >         $ dir/siz/grand 'CHECK'0 >         $ goto LOOPv >=   ------------------------------  % Date: Tue, 26 Feb 2002 15:33:14 +0100-9 From: Jan-Erik =?iso-8859-1?Q?S=F6derholm?= <aaa@aaa.com>r/ Subject: Re: tool to calc recursive dir sises ? ' Message-ID: <3C7B9CAA.F5FC6AC0@aaa.com>T  6 Actualy more or less as a DIR/TOT, *BUT* the total for4 each dir should include the file(s) in any subdir(s)/ to that directory. I'd like to find the largestw6 directory (incl any subdirs) on the [000000] level and: then "drill-down" from that to find the actual sub-sub-dir where all the files are.  	 Jan-Erik.l   WILLIAM WEBB wrote:a > @ > I'd have responded to your inquiry earlier, but was (and still> > am confused) about exactly how you wanted that totalling and > subtotalling to work.t >r   ------------------------------    Date: 26 Feb 2002 07:49:44 -0800. From: SPAMSINK2001@YAHOO.COM (Alan E. Feldman)/ Subject: Re: tool to calc recursive dir sises ?p= Message-ID: <343f30ae.0202260749.11955a2e@posting.google.com>d  U Jan-Erik Sderholm <aaa@aaa.com> wrote in message news:<3C7B4441.36E7934D@aaa.com>... 6 > Sort of, but it dosn't calculate the total att every9 > "node" in the directory tree on the DISK. Just the dirsd > under the [000000] level.S >  > Jan-Erik.s >  > Don Sykes wrote: > >  > > You mean like :> > >         $ LOOP:-4 > >         $ check = f$search("DISK:[000000]*.dir")( > >         $ if check.eqs.""  then exit8 > >         $ check = check -"000000]" -".DIR;1" +"...]"# > >         $ dir/siz/grand 'CHECK'0 > >         $ goto LOOPL > >   + So just add an ellipsis in the right place:r   	$ LOOP:, 	$ check = f$search("DISK:[000000...]*.dir") 	$ if check.eqs.""  then exit - 	$ check = check -"000000]" -".DIR;1" +"...]"i 	$ directory/siz/grand 'CHECK' 	$ goto LOOP  D This is certainly not the most efficient way to do it, but it should= work. If you use something like this, be sure that if you use0D something like DIR which you have defined as a personal symbol, that= it does not contain time-wasting extra qualifiers. E.g., with-A DIR:==DIR/SIZE/DATE/PROT, DIR/GRAND takes much longer than to run" DIRECTORY/GRAND.  D Just why do you want such multiple partially-redundant grand totals?   Disclaimer: JMHO Alan E. Feldman  afeldman bob gfigroup bob com04 "Help me help you help me help you"  --Bob Patterson   ------------------------------  % Date: Tue, 26 Feb 2002 11:42:38 -05000* From: WILLIAM WEBB <WWEBB1@email.usps.gov>/ Subject: RE: tool to calc recursive dir sises ?-- Message-ID: <0033000054404508000002L082*@MHS>2   =0AI reiterate my position.k  , (Here's the TREE.COM I remembered - it's old(  as the hills.  It will give you an ANSI graphical directory tree.)  3 It can walk the directory tree of a disk but unlessa( you modify it (which I'm sure you will),/ you'll have to set default to the MFD for it to? behave properly.   $ ESC[0,7] =3D 155 $ Gr_On =3D ESC + "(0" $ Gr_Off =3D ESC + "(B"" $ Tab =3D "                 "n $0 $ IF P1 .EQS. "" $  THEN  $   out =3D "WRITE SYS$OUTPUT" $  ELSE- $   OPEN/WRITE in 'P1' $   out =3D "WRITE in" $  ENDIF $l $ out ESC + "[2J"2 $ out ESC + "[1;1f": $a# $ Begin_Directory =3D F$DIRECTORY()- $ Pad =3D "tqq"E $@ $ OUT Tab + Begin_Directoryr $ CALL Get_Directories 'Pad' $ SET DEFAULT 'Begin_Directory'l $ IF P1 .NES. "" THEN CLOSE in $o $Get_Directories: SUBROUTINE $ Directory_Loop:e $   Dirct =3D F$search("*.dir")SE $   IF F$LOCATE("000000.000000",Dirct) .LT. F$LENGTH(Dirct) THEN Exit  $   IF Dirct .EQS. "" 	 $    THEN2
 $     EXIT	 $    ELSES5 $     Out_Dirct =3D F$EXTRACT(F$LOCATE("]",Dirct)+1,-a#                   F$LENGTH(Dirct),-h                   Dirct)-                    - ".DIR;1"0 $     OUT Tab + Gr_On + Pad + Gr_Off + Out_Dirct< $     New_Direct =3D F$EXTRACT(0,F$LOCATE("]",Dirct),Dirct)-F                    + "." + F$EXTRACT(0,F$LENGTH(Out_Dirct),Out_Dirct)-                    + "]" $     SET DEFAULT 'New_Direct' $     Pad =3D "x  " + Pads  $     CALL Get_Directories 'Pad', $     Pad =3D F$EXTRACT(3,F$LENGTH(Pad),Pad) $     GOTO Directory_Loop0	 $   ENDIFy $t $ENDSUBROUTINE $    That's gonna be some busy DCL-4 If it was electrical, the lights would probably dim.   :^)     & And I really think you do want to do a9 DIR/SIZ/TOT/GRAND disk[dir_where_you_are...] at each .DIRo1 you hit because you'll have less output to parse.a  0 $ dir/siz/grand/tot sys$sysdevice:[sys1...]*.*;*  < Grand total of 248 directories, 13402 files, 2582835 blocks.   WWWebb     -----Original Message-----/ From: Info-VAX-Request@Mvb.Saic.Com at INTERNETD( Sent: Tuesday, February 26, 2002 9:53 AMB To: Webb, William W Raleigh, NC; Info-VAX@Mvb.Saic.Com at INTERNET/ Subject: RE: tool to calc recursive dir sises ?n    6 Actualy more or less as a DIR/TOT, *BUT* the total for4 each dir should include the file(s) in any subdir(s)/ to that directory. I'd like to find the largesto6 directory (incl any subdirs) on the [000000] level and: then "drill-down" from that to find the actual sub-sub-dir where all the files are.  	 Jan-Erik.4   WILLIAM WEBB wrote:a >g@ > I'd have responded to your inquiry earlier, but was (and still> > am confused) about exactly how you wanted that totalling and > subtotalling to work.s >=   ------------------------------  % Date: Tue, 26 Feb 2002 18:01:24 +0100i9 From: Jan-Erik =?iso-8859-1?Q?S=F6derholm?= <aaa@aaa.com> / Subject: Re: tool to calc recursive dir sises ?.' Message-ID: <3C7BBF64.708A414D@aaa.com>e   Nice !   I did got a few :r  G "%DCL-W-DEFOVF, too many command procedure parameters - limit to eight"D   but otherwise it worked OK.b  	 Jan-Erik.o   WILLIAM WEBB wrote:i >  > I reiterate my position. > . > (Here's the TREE.COM I remembered - it's old* >  as the hills.  It will give you an ANSI > graphical directory tree.) >  [snip]   ------------------------------  % Date: Tue, 26 Feb 2002 12:50:41 -0500a* From: WILLIAM WEBB <WWEBB1@email.usps.gov>/ Subject: RE: tool to calc recursive dir sises ? - Message-ID: <0033000054415722000002L022*@MHS>    =0AI didn't write it,- it's practically apocryphal.  $ It's one of those standard exercises for VMS I or II classes.  ' It's right up there with "duplicate the2' output of the SHOW SYSTEM command using  DCL" exercise.   WWWebb     -----Original Message-----/ From: Info-VAX-Request@Mvb.Saic.Com at INTERNETg) Sent: Tuesday, February 26, 2002 12:04 PM B To: Webb, William W Raleigh, NC; Info-VAX@Mvb.Saic.Com at INTERNET/ Subject: RE: tool to calc recursive dir sises ?V     Nice !   I did got a few :E  H "%DCL-W-DEFOVF, too many command procedure parameters - limit to eight"=     but otherwise it worked OK.E  	 Jan-Erik.:   WILLIAM WEBB wrote:R >S > I reiterate my position. >r. > (Here's the TREE.COM I remembered - it's old* >  as the hills.  It will give you an ANSI > graphical directory tree.) >  [snip]=4   ------------------------------  % Date: Tue, 26 Feb 2002 10:08:30 +0300s4 From: "Ruslan R. Laishev" <laishev@smtp.deltatel.ru> Subject: Re: VAX SCANm0 Message-ID: <3C7B346E.6478D809@smtp.deltatel.ru>   Hello Hoff, H 	thanks for the info. Probably Tom will implement an Alpha version ? :-)  F >   Since Tom Linden has the source code for VAXscan from the FreewareE >   and likely also has a version of the GEM object libraries as partuD >   of his work supporting PL/I, Tom is probably in the best spot toA >   look into the GEM licensing and for an Alpha port of VAXscan.o   -- o Cheers, Ruslan. D +---------------------pure personal opinion------------------------+;       RADIUS Server for OpenVMS project - www.radiusvms.com@8         vms-isps@dls.net - Forum for ISP running OpenVMS*                  Mobile: +7 (901) 971-3222A    TKD (WTF) in Russia, St.-Petersburg - www.TaeKwonDo-WTF.SPb.RUo   ------------------------------  % Date: Tue, 26 Feb 2002 11:47:49 +0300a4 From: "Ruslan R. Laishev" <Laishev@SMTP.DeltaTel.RU> Subject: Re: VAX SCANn0 Message-ID: <3C7B4BB5.22264F47@SMTP.DeltaTel.RU>   Hi Hoff, 	some other question.aG >   There have been a number of folks that have looked at this port, asnH >   this is not the first request for an Alpha port of the VAXscan tool.F >   I and others have looked into this on various occasions -- VAXscanE >   is based on an old version of the VCG code generator for VAX, and1D >   would have to be heavily modified to get it to work with the GEM! >   code generator used on Alpha.iK 	Is there a way to simply redesign VAX SCAN to implement SCAN -> MACRO32/64t preprocessor ?   --   Cheers,hF +OpenVMS [Sys|Net] HardWorker .......................................+E  Russia,Delta Telecom Inc,                    Cel:  +7 (901) 971-3222eE  191119,St.Petersburg,Transportny per. 3                     116-3222yF +http://starlet.deltatel.ru ................. SysMan rides HailStorm +   ------------------------------    Date: 26 Feb 2002 05:51:50 -0600- From: Kilgallen@SpamCop.net (Larry Kilgallen)c Subject: Re: VAX SCAN13 Message-ID: <EFXIDyK13btL@eisner.encompasserve.org>a  g In article <3C7B4BB5.22264F47@SMTP.DeltaTel.RU>, "Ruslan R. Laishev" <Laishev@SMTP.DeltaTel.RU> writes:.  L > Is there a way to simply redesign VAX SCAN to implement SCAN -> MACRO32/64 > preprocessor ?  B That is possible, but certainly not "simple".  Doing so would alsoB reduce the degree of optimization possible, compared to a compiler& that called a code generator directly.  B I use SCAN a lot, but I have not illusions about the possibilitiesB of financial recovery of money invested in porting it to Alpha. ItB would possibly cause a small bit of increased interest in VMS, butB that is well below the level at which DEC thought funding it would be worthwhile.  B As an Ada fan, I note that GNAT offers a Spitbol (or is it Snobol)@ package for doing the same sorts of things that SCAN does.  That2 package probably has a larger user base than SCAN.   ------------------------------  % Date: Tue, 26 Feb 2002 13:01:04 +0100o9 From: Jan-Erik =?iso-8859-1?Q?S=F6derholm?= <aaa@aaa.com>  Subject: Re: VAX SCANo' Message-ID: <3C7B7900.A059FB67@aaa.com>f   Hi !) I'v read this thread with great interest.e6 Now, could someone in a few words desribe what SCAN is
 and/or does ?f   Jan-Erik Sderholm.o   ------------------------------    Date: 26 Feb 2002 07:24:35 -0600- From: Kilgallen@SpamCop.net (Larry Kilgallen)l Subject: Re: VAX SCANe3 Message-ID: <Oaje+hwG4sVf@eisner.encompasserve.org>e  c In article <3C7B7900.A059FB67@aaa.com>, Jan-Erik =?iso-8859-1?Q?S=F6derholm?= <aaa@aaa.com> writes:n  F > Now, could someone in a few words desribe what SCAN is and/or does ?  H VAX Scan is a compiled parsing language that results in high performance	 programs.e   ------------------------------    Date: 26 Feb 2002 05:41:36 -0800+ From: pierre.bru@spotimage.fr (Bru, Pierre)- Subject: Re: VAX SCAN < Message-ID: <1c0e37b1.0202260541.cfe2975@posting.google.com>  
 short answer:k ------------/ VAX-Scan is a text programable text processor. -   long answer: -----------t? VAX-Scan parses an input text coming from some input (a file oreA another language subroutine) using a lexical tokeniser + a syntax3E analyzer (think to Lex+Yacc in the same tool but IMHO more powerful).i  A recognized input text can be modified by the VAX-Scan program and C eventually re-scanned. input text, recognized or not, may be eithert@ pushed to some output (a file or another language subroutine) or ignored and dropped.  B of course, as you may call another language subroutine at any timeB during the scan process, anything else may be done like building aF memory syntax tree or compiling/interpreting the input but this is not the main goal of VAX-Scan.   Pierre.i  U Jan-Erik Sderholm <aaa@aaa.com> wrote in message news:<3C7B7900.A059FB67@aaa.com>...  > Hi !+ > I'v read this thread with great interest.i8 > Now, could someone in a few words desribe what SCAN is > and/or does ?  >  > Jan-Erik Sderholm.n   ------------------------------    Date: 26 Feb 2002 08:24:40 -0600- From: koehler@encompasserve.org (Bob Koehler)t Subject: Re: VAX SCANK3 Message-ID: <ukl+tFjR7KZY@eisner.encompasserve.org>e  c In article <3C7B7900.A059FB67@aaa.com>, Jan-Erik =?iso-8859-1?Q?S=F6derholm?= <aaa@aaa.com> writes:i > Hi !+ > I'v read this thread with great interest.e8 > Now, could someone in a few words desribe what SCAN is > and/or does ?t  G    VAX SCAN is sort of an intelligent replacement for lexx/yacc.  DEC's D    response to the desire to port SCAN was ue lexx and yacc instead.  @    In principle you could probably VEST SCAN, so it ran on AlphaA    creating VAX objects, cross link, then VEST the results to gettB    an Alpha program from SCAN.  Not exactly the best, but it would>    be an attractive alternative to rewriting good SCAN code in
    lexx/yacc.B   ------------------------------  # Date: Tue, 26 Feb 2002 14:38:34 GMT L From: winston@SSRL.SLAC.STANFORD.EDU ("Alan Winston - SSRL Admin Cmptg Mgr") Subject: Re: VAX SCANi8 Message-ID: <00A0A218.AD46CD3D@SSRL04.SLAC.STANFORD.EDU>  c In article <ukl+tFjR7KZY@eisner.encompasserve.org>, koehler@encompasserve.org (Bob Koehler) writes:nd >In article <3C7B7900.A059FB67@aaa.com>, Jan-Erik =?iso-8859-1?Q?S=F6derholm?= <aaa@aaa.com> writes: >> Hi !t, >> I'v read this thread with great interest.9 >> Now, could someone in a few words desribe what SCAN isd >> and/or does ? >iH >   VAX SCAN is sort of an intelligent replacement for lexx/yacc.  DEC'sE >   response to the desire to port SCAN was ue lexx and yacc instead.s >sA >   In principle you could probably VEST SCAN, so it ran on AlphaeB >   creating VAX objects, cross link, then VEST the results to getC >   an Alpha program from SCAN.  Not exactly the best, but it woulde? >   be an attractive alternative to rewriting good SCAN code in. >   lexx/yacc.  I Would it be a more attractive alternative than licensing the VAX emulatoroH (for Alpha) from SRI and just running VAX SCAN native on the emulator on Alpha?   -- Alan     O ===============================================================================d0  Alan Winston --- WINSTON@SSRL.SLAC.STANFORD.EDUM  Disclaimer: I speak only for myself, not SLAC or SSRL   Phone:  650/926-3056 M  Physical mail to: SSRL -- SLAC BIN 69, PO BOX 4349, STANFORD, CA  94309-0210 O ===============================================================================a   ------------------------------    Date: 26 Feb 2002 08:54:26 -0600- From: Kilgallen@SpamCop.net (Larry Kilgallen)? Subject: Re: VAX SCANc3 Message-ID: <r0kndZq6XSnc@eisner.encompasserve.org>e   In article <00A0A218.AD46CD3D@SSRL04.SLAC.STANFORD.EDU>, winston@SSRL.SLAC.STANFORD.EDU ("Alan Winston - SSRL Admin Cmptg Mgr") writes:Ke > In article <ukl+tFjR7KZY@eisner.encompasserve.org>, koehler@encompasserve.org (Bob Koehler) writes: e >>In article <3C7B7900.A059FB67@aaa.com>, Jan-Erik =?iso-8859-1?Q?S=F6derholm?= <aaa@aaa.com> writes:e >>> Hi !- >>> I'v read this thread with great interest..: >>> Now, could someone in a few words desribe what SCAN is >>> and/or does ?- >>I >>   VAX SCAN is sort of an intelligent replacement for lexx/yacc.  DEC's F >>   response to the desire to port SCAN was ue lexx and yacc instead. >>B >>   In principle you could probably VEST SCAN, so it ran on AlphaC >>   creating VAX objects, cross link, then VEST the results to get D >>   an Alpha program from SCAN.  Not exactly the best, but it would@ >>   be an attractive alternative to rewriting good SCAN code in >>   lexx/yacc.e > K > Would it be a more attractive alternative than licensing the VAX emulator J > (for Alpha) from SRI and just running VAX SCAN native on the emulator on > Alpha?  G It is certainly more attractive in that the pricing of the VAX emulator  from SRI is a big secret.-   ------------------------------  % Date: Tue, 26 Feb 2002 15:57:35 +0100 9 From: Jan-Erik =?iso-8859-1?Q?S=F6derholm?= <aaa@aaa.com>@ Subject: Re: VAX SCANT' Message-ID: <3C7BA25F.792450C0@aaa.com>    Hm...   B On "http://www.openvms.compaq.com/index.html", in the small box atC the left labeled "Search OpenVMS", I entered "VAX SCAN" and pressedl ENTER.  ( I got 1 (one) hit pointing at this URL :  H http://www.openvms.compaq.com/commercial/basic/basic_doc/bas_ref_015.htm  9 On that page there was one single reference to VAX SCAN :h     --- start pasted text ---;
 Example 3    !Conditional Statement Modifier  DECLARE INTEGER counter, &     STRING my_nameE INPUT "Try and guess my name";my_name FOR counter = 1 UNTIL my_name == "BASIC"- PRINT "You guessed it!"a   Output j  ! Try and guess my name? VAX PASCALi Try and guess my name? VAX SCANC Try and guess my name? BASIC You guessed it!/   --- end pasted text ---a   :-)s	 Jan-Erik-d   ------------------------------    Date: 26 Feb 2002 09:20:09 -0600- From: Kilgallen@SpamCop.net (Larry Kilgallen)  Subject: Re: VAX SCANn3 Message-ID: <fLpU29jgCNcF@eisner.encompasserve.org>r  c In article <3C7BA25F.792450C0@aaa.com>, Jan-Erik =?iso-8859-1?Q?S=F6derholm?= <aaa@aaa.com> writes:i > Hm...r > D > On "http://www.openvms.compaq.com/index.html", in the small box atE > the left labeled "Search OpenVMS", I entered "VAX SCAN" and pressed- > ENTER. > * > I got 1 (one) hit pointing at this URL :  ? How many references do you get to other discontinued products ?i   ------------------------------  % Date: Tue, 26 Feb 2002 15:35:27 +0100 6 From: =?iso-8859-1?Q?Jean=2DFran=E7ois=20PI=C9RONNE?=  Subject: Re: VAX SCAN.+ Message-ID: <3C7B9D2F.69FAA5DF@laposte.net>t   > H >   As I have indicated during these previous research projects into theE >   status of VAXscan, I would encourage the use of Perl or Python or  >   other approach.k >   O There is also PCCTS (http://www.polhode.com/home.html) which generate parser in 	 C or C++.cN PCCTS is a recursive descent parser generator supporting LL(k>=1) grammars and& arbitrary lookahead with backtracking., I have build many parser with it on OpenVMS.  O IMHO PCCTS is a superior tool to yacc/bison which are LALR-based parsing tools.*  N You can also take a look at ANTLR (http://www.antlr.org/) written in JAVA, but can generate parser in C++.'    D >   No, I know of no particular tools to port VAXscan code into Perl  >   or Python or other language. >    Jean-Franois Pironne   ------------------------------  % Date: Tue, 26 Feb 2002 16:38:11 +0100 9 From: Jan-Erik =?iso-8859-1?Q?S=F6derholm?= <aaa@aaa.com>i Subject: Re: VAX SCAN=' Message-ID: <3C7BABE3.C9B65FC9@aaa.com>l  : Well, for a discontinued product, I'd expect, at least, toB find a link to some page where discontinued products are listed...  	 Jan-Erik.a   Larry Kilgallen wrote: > e > In article <3C7BA25F.792450C0@aaa.com>, Jan-Erik =?iso-8859-1?Q?S=F6derholm?= <aaa@aaa.com> writes:u	 > > Hm...0 > >:F > > On "http://www.openvms.compaq.com/index.html", in the small box atG > > the left labeled "Search OpenVMS", I entered "VAX SCAN" and pressed8
 > > ENTER. > >r, > > I got 1 (one) hit pointing at this URL : > A > How many references do you get to other discontinued products ?p   ------------------------------  % Date: Tue, 26 Feb 2002 07:53:03 -0800s# From: "Tom Linden" <tom@kednos.com>  Subject: RE: VAX SCANs9 Message-ID: <CIEJLCMNHNNDLLOOGNJIOELIEDAA.tom@kednos.com>[  H I find google does a better job of finding thins on compaq's site rather than the "Search OpenVMS"   > -----Original Message-----/ > From: Jan-Erik Soderholm [mailto:aaa@aaa.com]i* > Sent: Tuesday, February 26, 2002 7:38 AM > To: Info-VAX@Mvb.Saic.ComS > Subject: Re: VAX SCAN_ >e >r< > Well, for a discontinued product, I'd expect, at least, toD > find a link to some page where discontinued products are listed... >l > Jan-Erik.S >" > Larry Kilgallen wrote: > >o4 > > In article <3C7BA25F.792450C0@aaa.com>, Jan-Erik4 > =?iso-8859-1?Q?S=F6derholm?= <aaa@aaa.com> writes: > > > Hm...G > > >tH > > > On "http://www.openvms.compaq.com/index.html", in the small box atI > > > the left labeled "Search OpenVMS", I entered "VAX SCAN" and pressed$ > > > ENTER. > > > . > > > I got 1 (one) hit pointing at this URL : > > C > > How many references do you get to other discontinued products ?D >E   ------------------------------  % Date: Tue, 26 Feb 2002 06:06:25 -08000# From: "Tom Linden" <tom@kednos.com>  Subject: VAX SCAN survey9 Message-ID: <CIEJLCMNHNNDLLOOGNJIOEKPEDAA.tom@kednos.com>   L OK, we will look into porting it to Alpha.  What we need to know is how many peopleL are still using SCAN code and would they be willing to pay an annual license fee (offA an amount yet to be determined, based on the number of responses)   F Send the response directly to me and append  the word YES or NO to the subject = and in the body put whatever you like which you think helpful    Tom/   ------------------------------   End of INFO-VAX 2002.112 ************************  