1 INFO-VAX	Tue, 09 Dec 2003	Volume 2003 : Issue 681       Contents:G Re: Another DCL request: A way to get the .dir of the current directory G Re: Another DCL request: A way to get the .dir of the current directory G Re: Another DCL request: A way to get the .dir of the current directory G Re: Another DCL request: A way to get the .dir of the current directory  dcps Re: dcps Re: dcps8 DCPS  problem with /param=("page_orientation=landscape") Re: DEC$DOCUMENT v2.3 3 Gartner: HP is Leader in High-Availability Services  Re: Hairdoo Economics  RE: Hairdoo Economics  Re: Hairdoo Economics : Re: HP tops 1 million tpmC on Superdome with Itanium2 CPUs: Re: I wonder if this HP director will resign from HP's BOD: Re: I wonder if this HP director will resign from HP's BOD: Re: I wonder if this HP director will resign from HP's BOD It's simple... Re: lib$spawn particulars  Re: lib$spawn particulars  Re: lib$spawn particulars E linker collisions and porting (was Re: Freeware download corrupt (?))  Re: Mailbox deletion Mailbox deletion Re: Mailbox deletion: Re: OpenVMS clusters give Windows, Unix thorough thrashing: Re: OpenVMS clusters give Windows, Unix thorough thrashing: Re: OpenVMS clusters give Windows, Unix thorough thrashing: Re: OpenVMS clusters give Windows, Unix thorough thrashing" Re: OSU http server & http headers" Re: OSU http server & http headers" Re: OSU http server & http headers5 Re: OT Very scary: Cars running on Microsoft software 5 Re: OT Very scary: Cars running on Microsoft software 5 Re: OT Very scary: Cars running on Microsoft software 5 Re: OT Very scary: Cars running on Microsoft software 5 Re: OT Very scary: Cars running on Microsoft software 5 Re: OT Very scary: Cars running on Microsoft software   Re: porting problems encountered  Re: porting problems encountered& Re: printing 11x17 landscape documents$ Re: REPLY/STATUS in another VMS node Re: SEARCH enhancement Re: SEARCH enhancement Re: SEARCH enhancement Re: SEARCH enhancement Re: SEARCH enhancement Re: SEARCH enhancement" Re: SEARCH enhancement [type/tail]9 Re: Sun to use AMD Opteron - announcement expected Monday 9 Re: Sun to use AMD Opteron - announcement expected Monday 9 Re: Sun to use AMD Opteron - announcement expected Monday = Re: VMS clusters prove they are the best - Sun comes in last!   F ----------------------------------------------------------------------   Date: 9 Dec 2003 07:39:36 -0600 ; From: koehler@eisner.nospam.encompasserve.org (Bob Koehler) P Subject: Re: Another DCL request: A way to get the .dir of the current directory3 Message-ID: <xcJ9Z90jzH+B@eisner.encompasserve.org>   n In article <b096a4ee.0312081446.461d11d4@posting.google.com>, spamsink2001@yahoo.com (Alan E. Feldman) writes:# > Attention Guy Peleg! DCL Request:  > L > A lexical function that returns the .DIR version of the current directory: >   A    In general, build scripts would be easier to write and to port <    from UNIX if DCL had a lexical function that would append+    a subdirectory name to a directory name.   5    e.g. in UNIX make files we often find things like:        base = able        next_dir = $base/baker'       next_next_dir = $next_dir/charlie       which gives:        next_dir is able/baker)       next_next_dir is able/baker/charlie   D    To do that in DCL, you have to keep adding and removing . and [] H    or <> in all the right places.  It would be convinient if we could do    something like:        base = "[.able]" /      next_dir = f$subdirectory(base,"[.baker]") :      next_next_dir = f$subdirectory(next_dir,"<.charlie>")      from which you'ld get:        next_dir is "[.able.baker]"-      next_next_dir is "[.able.baker.charlie]"    ------------------------------   Date: 9 DEC 2003 15:46:40 GMT + From: Dave Greenwood <greenwoodde@ornl.gov> P Subject: Re: Another DCL request: A way to get the .dir of the current directory1 Message-ID: <9DEC03.15464099@feda01.fed.ornl.gov>   S In a previous article, koehler@eisner.nospam.encompasserve.org (Bob Koehler) wrote: p > In article <b096a4ee.0312081446.461d11d4@posting.google.com>, spamsink2001@yahoo.com (Alan E. Feldman) writes:% > > Attention Guy Peleg! DCL Request:  > > N > > A lexical function that returns the .DIR version of the current directory: > >  >   C >    In general, build scripts would be easier to write and to port > >    from UNIX if DCL had a lexical function that would append- >    a subdirectory name to a directory name.  >   7 >    e.g. in UNIX make files we often find things like:  >       base = able  >       next_dir = $base/baker) >       next_next_dir = $next_dir/charlie  >    >    which gives:  >       next_dir is able/baker+ >       next_next_dir is able/baker/charlie  >   F >    To do that in DCL, you have to keep adding and removing . and [] J >    or <> in all the right places.  It would be convinient if we could do >    something like: >    >      base = "[.able]" 1 >      next_dir = f$subdirectory(base,"[.baker]") < >      next_next_dir = f$subdirectory(next_dir,"<.charlie>") >    >    from which you'ld get: " >      next_dir is "[.able.baker]"/ >      next_next_dir is "[.able.baker.charlie]"   L Note that determining what "all the right places" are becomes more difficultH with ODS-5 since . and <> can now be imbedded in filenames.  Seems to meL that a VMS-supplied function to convert a directory to its .DIR version (andH vice versa) and/or to append subdirectories would improve the quality ofI VMS scripts.  It would be good to make the same capabilities available to  compiled languages also.   Dave --------------9 Dave Greenwood                Email: Greenwoodde@ORNL.GOV H Oak Ridge National Lab        %STD-W-DISCLAIMER, I only speak for myself   ------------------------------  # Date: Tue, 09 Dec 2003 16:10:43 GMT 0 From: "Guy Peleg" <guy.peleg@hp.com_remove_this>P Subject: Re: Another DCL request: A way to get the .dir of the current directory3 Message-ID: <7umBb.10375$hL5.6852@news.cpqcorp.net>   . I agree that F$FID_TO_NAME can be very useful.  E I wonder how many folks agree....please reply here or drop me a line.   G If there is a big enough of a demand..... you never know....maybe we'll  teach vaxes how to do this ;-)   Guy , <VAXman- @SendSpamHere.ORG> wrote in message* news:00A2A142.2BF1ACB9@SendSpamHere.ORG...? > In article <b096a4ee.0312081446.461d11d4@posting.google.com>, 0 spamsink2001@yahoo.com (Alan E. Feldman) writes:$ > >Attention Guy Peleg! DCL Request: > > B > >A lexical function that returns the .DIR version of the current
 directory: > >  > > / > >If you were in (or specified as a parameter)  > >  > >    DISK$DATA1:[FELDMAN.UTI]  > >  > >it would return > > " > >    DISK$DATA1:[FELDMAN]UTI.DIR > >  > > / > >If you were in (or specified as a parameter)  > >  > >    ROOTED_DISK:[000000]  > >  > >it would return > >  > >    DISK$DATA1:[1.2.3]4.DIR > > - > >if ROOTED_DISK were DISK$DATA1:[1.2.3.4.].  > >  > >  > >Alan E. Feldman > = > A better request might be for something like F$FID_TO_NAME. > > You could pass your filename to F$FILE_ATTRIBUTES to get the< > DID and then you could pass the DID to F$FID_TO_NAME.  I'd= > think a lexical like F$FID_TO_NAME would make DCL more use- < > ful than just a sinlge function to return directory names. >  > --B > http://www.legacy-2000.com  for the best OpenVMS system security
 solutions. > 2 > VAXman- A Bored Certified VMS Kernel Mode Hacker VAXman(at)TMESIS(dot)COM > 6 >   "Well my son, life is like a beanstalk, isn't it?"   ------------------------------  % Date: Tue, 09 Dec 2003 19:18:19 +0100 * From: Paul Sture <nospam@sture.homeip.net>P Subject: Re: Another DCL request: A way to get the .dir of the current directory0 Message-ID: <3FD61FFB.1690B1EA@sture.homeip.net>   Dave Greenwood wrote:  > U > In a previous article, koehler@eisner.nospam.encompasserve.org (Bob Koehler) wrote: r > > In article <b096a4ee.0312081446.461d11d4@posting.google.com>, spamsink2001@yahoo.com (Alan E. Feldman) writes:' > > > Attention Guy Peleg! DCL Request:  > > > P > > > A lexical function that returns the .DIR version of the current directory: > > >  > > E > >    In general, build scripts would be easier to write and to port @ > >    from UNIX if DCL had a lexical function that would append/ > >    a subdirectory name to a directory name.  > > 9 > >    e.g. in UNIX make files we often find things like:  > >       base = able   > >       next_dir = $base/baker+ > >       next_next_dir = $next_dir/charlie  > >  > >    which gives:   > >       next_dir is able/baker- > >       next_next_dir is able/baker/charlie  > > G > >    To do that in DCL, you have to keep adding and removing . and [] L > >    or <> in all the right places.  It would be convinient if we could do > >    something like: > >  > >      base = "[.able]" 3 > >      next_dir = f$subdirectory(base,"[.baker]") > > >      next_next_dir = f$subdirectory(next_dir,"<.charlie>") > >  > >    from which you'ld get: $ > >      next_dir is "[.able.baker]"1 > >      next_next_dir is "[.able.baker.charlie]"  >    I agree that would be useful.   E There's possibly an argument for converting instances of ".][" within   a dev/dir string to "." as well.  N > Note that determining what "all the right places" are becomes more difficultJ > with ODS-5 since . and <> can now be imbedded in filenames.  Seems to meN > that a VMS-supplied function to convert a directory to its .DIR version (andJ > vice versa) and/or to append subdirectories would improve the quality ofK > VMS scripts.  It would be good to make the same capabilities available to  > compiled languages also. >   E Maybe converting "[.able.baker.charlie]" to "able//baker/charlie" and H vice versa? Isn't that already available somewhere in either the CRTL or porting libraries?   --  
 Paul Sture   ------------------------------   Date: 9 Dec 2003 02:37:56 -0800 ( From: ed.janssen@globalknowledge.nl (Ed)
 Subject: dcps < Message-ID: <6cf15da0.0312090237.9468297@posting.google.com>  B I installed DCPS V2.2 and want to use it with a HP LaserJet 4000n.D I defined the logical dcps_lib related to a new created text library= file and include a lot of modules in the library, so also the  landscape module. C When I want to print a file using the page_orientation landscape it > just prints the file portrait. When I print the file using theF /setup=landscape it prints the file landscape. The problem is that theB OpenVMS server is the printserver for the Tru64 systems. The Tru64C system can print to the OpenVMS printserver but they always use the F qualifier /params=("page_orientation=landscape"), but the print server# does not use the qualifier /params.  What can be the problem?   ------------------------------  % Date: Tue, 09 Dec 2003 06:10:50 -0500 * From: JF Mezei <jfmezei.spamnot@istop.com> Subject: Re: dcps ) Message-ID: <3FD5AD85.C6DC0EAA@istop.com>   	 Ed wrote: E > When I want to print a file using the page_orientation landscape it @ > just prints the file portrait. When I print the file using the1 > /setup=landscape it prints the file landscape.      H I am not 100% sure of this, but I suspect that the page_orientation onlyN applies when DCPS actually takes your input file and converts it to postscript) (for instance, if you print a text file).   I /setup actually just inserts the postscript code in the specifided module  before your own.  M What is the format of the files you are printing ? Text, postscript ? other ?    ------------------------------  # Date: Tue, 09 Dec 2003 14:59:17 GMT * From: Paul Anderson <paul.anderson@hp.com> Subject: Re: dcps 5 Message-ID: <091220030959242617%paul.anderson@hp.com>   ? In article <6cf15da0.0312090237.9468297@posting.google.com>, Ed & <ed.janssen@globalknowledge.nl> wrote:  F > I installed DCPS V2.2 and want to use it with a HP LaserJet 4000n. ID > defined the logical dcps_lib related to a new created text library? > file and include a lot of modules in the library, so also the  > landscape module.   @ Is there a reason you created a new device control library?  TheF LaserJet 4000 should work fine with the DCPS device control library as shipped.  E > When I want to print a file using the page_orientation landscape it @ > just prints the file portrait. When I print the file using the0 > /setup=landscape it prints the file landscape.  E What module do you mean by the "landscape" module?  (There is no DCPS ( device control module called LANDSCAPE.)  E What is your definition of the DCPS_LIB logical name?  (Does it point 4 just to your new library or both it and DCPS$DEVCTL?  C > The problem is that the OpenVMS server is the printserver for the F > Tru64 systems. The Tru64 system can print to the OpenVMS printserver# > but they always use the qualifier G > /params=("page_orientation=landscape"), but the print server does not  > use the qualifier /params.  C So there are remote LPD queues on the UNIX system that use the DCPS D execution queues?  Are the UNIX systems passing the PAGE_ORIENTATION
 parameter?   Paul   --    Paul Anderson   OpenVMS Engineering    Hewlett-Packard Company    ------------------------------   Date: 9 Dec 2003 03:30:15 -0800 ( From: ed.janssen@globalknowledge.nl (Ed)A Subject: DCPS  problem with /param=("page_orientation=landscape") < Message-ID: <6cf15da0.0312090330.e64dbb6@posting.google.com>  ( I installed DCPS V2.2 on OpenVMS V7.2.2.A I created a new library file and inserted a lot of modules in the E file. So also the module landscape. I defined the logical dcps_lib to  my new created library file.E I defined the form dcps$default on the queue but when I want to print - a ascii text file to the queue with qualifier E /param=("page_orientation=landscape") it is just printed in portrait. C If I use the qualifier /setup=landscape in the print command, it is  printed landscape. What can be the problem?    
 Best regards, 
 Ed Janssen   ------------------------------   Date: 9 Dec 2003 01:56:56 -0800 . From: send_lotsa_spam_here@yahoo.com (Soterro) Subject: Re: DEC$DOCUMENT v2.3= Message-ID: <1a63f162.0312090156.3d4ab711@posting.google.com>   a Paul Repacholi <prep@prep.synonet.com> wrote in message news:<87brqj71ig.fsf@prep.synonet.com>... > > Did you run DOC$STARTUP? Installing the DOC message file may< > fix the NOMSG as well. It defines a shed load of logicals.   Did I run what??  $ *runs to seek any FM so he can RTFM*, Actually it worked fine after running that.    Thanks A LOT :)    S    ------------------------------  $ Date: Tue, 9 Dec 2003 09:00:39 +0100( From: "Rudolf Wingert" <win@fom.fgan.de>< Subject: Gartner: HP is Leader in High-Availability Services: Message-ID: <MCELKPMOKPMNDNKJNIONEECNCHAA.win@fom.fgan.de>   Hello,  I this may be possible. But the whole service must good to be a top service N seller. We did buy HP care paks for printer for three year onsite service. OurI feeling is: onsites means the customer is onsite and should qualifier the N problem and the highest level of service is: the customer repairs the printer.P For what do I buy onsite service? I dont know. Also every time you have a call,E the first solution of your problem is: you will have the wrong paper! K If HP would like to be a good service companie it should make there low end  service much much much better.   Best regards Rudolf Wingert    ------------------------------  % Date: Tue, 09 Dec 2003 03:57:31 -0500 * From: JF Mezei <jfmezei.spamnot@istop.com> Subject: Re: Hairdoo Economics) Message-ID: <3FD58E51.571293B5@istop.com>    Keith Parris wrote: H > Power is indeed NOT mainstream, nor an industry standard, and it neverG > will be.  It doesn't run Windows (it had its chance and blew it), and A > no other computer manufacturer (unless you count Apple with the 2 > PowerPC flavor) will ever make systems using it.  M Funny, didn't Microsoft just annouce they would adopting Power for its Xbox ? F Apple may not have a dominent market share, but it is mainstream. they? advertise to the public and its products are easily available.    M And yes, you can run Windows on Power. The mac has windows emulators. Heck, I  can run DOS on my vax :-)     C > IBM recently announced 16-way Itanium servers. An astute observer 3 > might ask why IBM is doing that if it has Power?     Two reasons:R 1- to please Intel and secure better prices for its industry standard line (8086).M 2- so they can claim "me too". And by being a customer if IA64, they probably I have access to advance information on the chip which then helps them keep  Power way ahead of IA64.   > IBM appears to be @ > hedging its bets because it looks like Itanium will become the$ > mainstream 64-bit CPU for servers,  V I would suggest you have your tap water tested for substances which may cloud reality.  A > The important word there is "was".  Now something "is" going to H > dislodge 80x86 from the "mainstream" title eventually, as the industry& > moves from a 32-bit focus to 64-bit.  K Yes, this was said early on when the 8086 was just a 16 bit chip with 8 bit M bus and <cough> segment registers. Today, it is a 64 bit architecture that is L quite respectable in performance and because of its simplicity, is poised toN move faster than IA64. 8086 can have every expectation to retain the "IndustryK standard" term for the foreseable future. The minute AMD start to steal too L many sales from Intel, you can bet that Intel will magically come out with a 64 bit 8086.  M Intel cannot afford to cannabalise its 8086 line  just for the sake of saving 2 the sale of a couple thousand IA64 chips per year.    $ > > Since when is IA64 'mainstream'? > ! > Not yet, but it's headed there.   J The expectation of IA64 making it to mainstream was completely killed when0 intel admitted it wouldn't be going to desktops.  H > Intel wants IA64 to be that mainstream, and AMD wants x86-64 to be the
 > mainstream.   < Nop. Intel has stated that IA64 wouldn't go for the desktop.  F In some ways, Intel reminds me of Boeing. When Airbus announced it wasI building the A380 super jumbo, Boeing said that it wasn't worried because M there wouldn't be a market for it. But in the last 4 years, how many new 747s I has Boeing sold ? Airbus has sold about 120 of those beasts, and the A380 L isn't flying yet. And now, you have the press talking about Boeing falteringN without a clear direction and Airbus getting more than half the market. BoeingM may be able to get back to health and prosperity, so it isn't a doomsday. But N Boeing has lost significant market share (especially if you combine the market7 share thet McD used to have (McD is now part of Boeing)     G Intel must not underestimate AMD'S 64 bit 8086 or it may become another  Boeing.   9   I think the signals at present indicate that Intel will D > get the majority of the design wins, as it did when competing withE > Motorola in the '80s. (Time to market beat technical elegance every 
 > time.)    ; IA64 isn't eactly a winner when it comes to time-to-market.    ------------------------------  $ Date: Tue, 9 Dec 2003 06:20:43 -0500' From: "Main, Kerry" <kerry.main@hp.com>  Subject: RE: Hairdoo EconomicsR Message-ID: <FD827B33AB0D9C4E92EACEEFEE2BA2FB1E09D9@tayexc19.americas.cpqcorp.net>   > -----Original Message-----, > From: Andrew Harrison SUNUK Consultancy=203 > [mailto:Andrew_No.Harrison_No@nospamn.sun.com]=20   > Sent: December 8, 2003 8:24 AM > To: Info-VAX@Mvb.Saic.Com   > Subject: Re: Hairdoo Economics >=20 > Main, Kerry wrote: > >>-----Original Message----- >=20H > >>>AIX and Solaris are struggling to remain relevant in light of Linux > >=20@ > > (and Solaris has the additional boat-anchor of poor SPARC=20 > performance).  > > << > >=20? > > Yep, and IBM (Senior VP of IBM Software) stated publicly=20  > earlier this=20 @ > > year that the likely future for their AIX base was likely=20 > going to be=20B > > Linux. So that decision, if in fact this is true, is likely=20
 > going to=20 B > > impact Sun as well. It also means some potential migrations=20 > are going=20G > > to be required for current AIX users at some point in the future=20 @ > > (albeit IBM will support them for some time - likely just=20 > like VAX users). > >=20 >=20= > Umm, if you had bothered to post or read the rest of the=20 & > article you would have noticed that: >=208 > 1.	There is no commitment from IBM to move to Linux it5 > 	from AIX it is just a logical option and of course  > 	it is a logical option.9 > 2.	The IBM exec interviewed is not responsible for OS's 9 > 3.	The timelines for this migration are multidecade e.g  > 	10-20 years minimum.  >=20> > 10-20 years is way too long for anyone to either take any=20D > notice of the comments or even think about planning for the event. >=20B > In addition I would also be really very interested in how you=20B > imagine that IBM's 10-20 year migration from AIX to Linux (if=20/ > it happens at all) is going to impact on Sun.  >=20< > This has been usefull though for one reason its a great=20  > example of real FUD in action. >=20@ > For anyone who doesn't know what real FUD is Kerry has just=206 > provided the almost perfect reference example of it. >=20A > Speculation piled on speculation linked with supposition and=20 @ > all designed to spread a little uncertainty, not a fact in it. >=20	 > regards  > Andrew Harrison  >=20   Andrew, Andrew -  C Yes, different VP's in the same company will disagree on timing and G strategies, but a key point in the article - the person stating this is C *the* Senior Exec VP of Software for IBM. This is not just some low  ranking IT manager.=20  H If this had been a HP quote, this specific newsgroup would have been lit> up like a X-mas tree with concerns about future OS strategies.  G While it is obviously going to take time, it illustrates the increasing C pressures that traditional UNIX's are going to have from a business H perspective i.e. how to differentiate themselves from Linux. This is notD something that I have just dreamed up - it has been an open question1 under discussion in many aspects of the industry.   B > Speculation piled on speculation linked with supposition and all? designed to spread a little uncertainty, not a fact in it. >=20   9 Oh, I see - and this is something you know nothing about?    :-)    Regards   
 Kerry Main Senior Consultant  HP Services Canada Voice: 613-592-4660  Fax: 613-591-4477  Email: kerryDOTmainAThpDOTcom . (remove the DOT's and AT for email address)=20   ------------------------------  % Date: Tue, 09 Dec 2003 12:12:50 +0000 O From: Andrew Harrison SUNUK Consultancy <Andrew_No.Harrison_No@nospamn.sun.com>  Subject: Re: Hairdoo Economics0 Message-ID: <br4e83$l1e$1@new-usenet.uk.sun.com>   Main, Kerry wrote: >    > Andrew, Andrew - >   > Kerry you should know from bitter experience repeating my name@ in a condescending way at the beginning of one of your responsesB is a sure fire measure of the amount of BS you are about to spout.  ( The more repetitions the greater the BS.  E > Yes, different VP's in the same company will disagree on timing and I > strategies, but a key point in the article - the person stating this is E > *the* Senior Exec VP of Software for IBM. This is not just some low  > ranking IT manager.  >   > Where in this article does any IBM VP talk about a strategy of migrating from AIX to Linux ?   = The Journo asked the VP a question to which the only sensible : answer was yes. Would Linux be an option to replace AIX to which the answer has to be yes.   A The fact that the VP who would actually be responsible for making D this descision and then implimenting it was much much more qualified> in his reponse should have helped you realise that there is no1 foundation in the article it is pure speculation.   J > If this had been a HP quote, this specific newsgroup would have been lit@ > up like a X-mas tree with concerns about future OS strategies. >   : And quite rightly given your track record of morphing your OS and platform strategy.   ? Imagine you are a convicted criminal released early on license. 8 You operate under a set of rules that don't apply to the: rest of society, you have to be home by 9, no fraternising9 with known criminals etc. Because of your previous crimes 9 you don't get to complain about this and if you do no one  takes any notice anyway.  < This is how it is for HP, there is a set of rules that apply8 to you because of your past behaviour which simply don't, apply to Sun, IBM, Dell etc. Get used to it.  I > While it is obviously going to take time, it illustrates the increasing E > pressures that traditional UNIX's are going to have from a business J > perspective i.e. how to differentiate themselves from Linux. This is notF > something that I have just dreamed up - it has been an open question3 > under discussion in many aspects of the industry.  >   E And so is Windows eventual evisceration of UNIX, its under discussion ' but its just speculation and thats all.    > B >>Speculation piled on speculation linked with supposition and all > ? > designed to spread a little uncertainty, not a fact in it. >   > ; > Oh, I see - and this is something you know nothing about?  >   / I note that you havn't even attempted to refute ? my claim that you have just produced a perfectly formed example > of FUD at its best, instead you appear to be trying to suggest that you are no worse than me.  3 So come back with some examples of the FUD that you & appear to be alledging I have spread ?     Regards  Andrew Harrison    ------------------------------  % Date: Tue, 09 Dec 2003 12:16:12 +0000 O From: Andrew Harrison SUNUK Consultancy <Andrew_No.Harrison_No@nospamn.sun.com> C Subject: Re: HP tops 1 million tpmC on Superdome with Itanium2 CPUs 0 Message-ID: <br4eec$l42$1@new-usenet.uk.sun.com>   Keith Parris wrote: f > "Bill Todd" <billtodd@metrocast.net> wrote in message news:<saidnW1SOPlYb06iRVn-vA@metrocast.net>... > G >>Oh, dear - another press release that's a bit moldy around the edges.  >  > C > A tablespoon of bleach in a quart of water should kill that mold.  > F > I'm even sorrier at my delay in getting this particular announcementD > out to you, as I know you were eagerly waiting to see which vendorD > would get the bragging rights to 1 million TPC-C tpm first, and itG > must have been even more of a disappointment than the late arrival of H > the news to see that it was HP, with that despicable Itanium chip, not& > IBM with Power, who got there first. > ` > I see at http://www.tpc.org/tpcc/results/tpcc_perf_results.asp?resulttype=noncluster&version=5E > that HP occupies the top 3 slots for TPC-C (non-clustered) results, G > and 5 out of the top 10 slots, vs. only 3 in the top 10 for IBM.  And @ > the spot in the top 10 held by NEC is also an Itanium2 system.  8 And the point of all this is ???????????????????????????  8 When you can explain why TPC-C is a worthwhile benchmark: then you also get to brag about having the fastest number.  6 TPC-C is the WonderBread of benchmarks, how much value; would you apply to being able to bake the best WonderBread.    Regards  Andrew Harrison    ------------------------------  % Date: Tue, 09 Dec 2003 03:24:59 -0500 * From: JF Mezei <jfmezei.spamnot@istop.com>C Subject: Re: I wonder if this HP director will resign from HP's BOD ) Message-ID: <3FD586B3.244AC402@istop.com>    Bill Todd wrote:M > plan (stated explicitly as a 'commitment' as well) was to release a faster, N > cooler, larger-cache, smaller-die EV79 product to give customers one more atE > least moderately significant upgrade to replace the cancelled EV8.    M As I recall, both EV7 and EV79 were designed to keep Compaq competitive until  IA64 could catch up.  K The original plan may have been to keep Alpha as fast as possible to retain L customers simply because competitors were slower, but it seems that now, theI plan is to make IA64 faster than Alpha as fast as possible and the easier M thing is to slow down Alpha. And by releasing a very slight speed bump a year K earlier, it will make Alpha inconsequential to the industry a year earlier, , with no risk of EV79 still outspeeding IA64.  L Is that is NOT the case, then HP has a very bad PR department because HP has0 taken no step, nor said anything to negate this.  K > 3.  At the time of the Alphacide, Compaq (knowing full well that a merger H > with HP was imminent) committed to porting Tru64 to Itanic.  So eitherE > Compaq lied knowingly, or HP 'agreed' and then immediately reneged    L You cannot fault Compaq or Curly/Carly for this. Curly could not in any way,L lead anyone to speculate on the future of Compaq.  Had he omitted Tru64 fromL the IA64 announcement, especially since Tru64's port to IA64 had begun underI Palmer, it would have lead to instant speculation on what strategy Compaq J would take for Unix since it would no longer have any commercial Unix. (InF hindsight, perhaps they could have announced a major effort for Linux,L dropping their proprietary Unix). Going Linux all the way could have been anL interesting diversion from any specualtion about Compaq being in discussions to be taken over.    ------------------------------  % Date: Tue, 09 Dec 2003 10:47:19 +0000 O From: Andrew Harrison SUNUK Consultancy <Andrew_No.Harrison_No@nospamn.sun.com> C Subject: Re: I wonder if this HP director will resign from HP's BOD 0 Message-ID: <br497n$j7t$2@new-usenet.uk.sun.com>   Keith Parris wrote:  > Andrew Harrison SUNUK Consultancy <Andrew_No.Harrison_No@nospamn.sun.com> wrote in message news:<br1po1$lm7$1@new-usenet.uk.sun.com>...  > , >>Last year the EBU lost 56 million dollars  >  > ) > The EBU figure was actually 54 billion.  >  > I >>So what you are saying in essence is without OpenVMS the EBU would have - >>lost 556 million dollars not including R&D.  >  > F > The $500 million in profits for VMS includes software (including VMSC > and VMS layered products), hardware (servers, storage, etc.), and C > services.  HP Services made a profit of $1.372 Billion last year.     = Ok take half of it out and give it to services, they now make  ~1.1 billion dollars.   < EBU lose 300 million dollars or 1.3 billion adjusted for R&D     Regards  Andrew Harrison    ------------------------------  % Date: Tue, 09 Dec 2003 10:45:35 +0000 O From: Andrew Harrison SUNUK Consultancy <Andrew_No.Harrison_No@nospamn.sun.com> C Subject: Re: I wonder if this HP director will resign from HP's BOD 0 Message-ID: <br494g$j7t$1@new-usenet.uk.sun.com>   Keith Parris wrote:  > Andrew Harrison SUNUK Consultancy <Andrew_No.Harrison_No@nospamn.sun.com> wrote in message news:<br1po1$lm7$1@new-usenet.uk.sun.com>...  > + >>Last year the EBU lost 56 million dollars  >  > 1 > According to HP's most-recent quarterly results G > (http://www.hp.com/hpinfo/investor/financials/quarters/2003/q4.html), D > the Enterprise Business Unit made a profit of $106 million for theF > quarter, returning to profitability before the end of the year, justH > as Carly predicted.  Profits for the EBU for the year improved by $610! > million over the previous year.     C And it lost 56 million for the full year. The EBU posted 70, 7, and 2 83 million dollar losses in the previous quarters.  ; These don't include EBU R&D which was removed from EBU P&L.   D HP Spend about 3.7 billion a year in R&D, assuming its spread evenly< across the 3 technology businesses, Printing, PC and EBU you; end up with a real loss of ~1 billion last year for the EBU  give or take 56 million.   Regards  Andrew Harrison    ------------------------------  $ Date: Tue, 9 Dec 2003 12:35:59 +0000( From: "Lariean Poringa" <Abyss@Dev.Null> Subject: It's simple... ? Message-ID: <dTkBb.22716$LI6.7062@news-binary.blueyonder.co.uk>             MAKE MONEY!!!     MAKE THOUSANDS!!!    I found this on a bulletin board and decided to try it: I don't care about the useless pre-fabricated crap this message usually says. All I say is, it works. Continue pre-fab crap.   WELL GUESS WHAT!!!  : Within seven days, I started getting money in the mail!! IB was shocked!! I figured it would end soon, but the money just keptU coming in. In my first week, I made about $25.00. By the end of the second week I had F made a total of more than $1000.00!! In the third week I had more than $10,000.00 and it's still growing!! This is now my fourth week and I have made a total of $42,000.00 and it's still coming rapidly. It's certainly worth $6.00 and six stamps, and I have spent more than that on the lottery without ever  winning!!!  I Let me tell you how this works and most important, why it works..........  also make sure you print this out NOW, so you can get the information off of it, as you will need it. I promise you that if you follow the directions exactly that youS will start making more money than you thought possible by doing something so easy!!:  N Suggestion: Read this entire message carefully!! (Print it out or download it)  E Follow the simple directions and watch the money come in!! It's easy. A It's legal. And, your investment is only $6.00 (Plus postage) !!!T  
 IMPORTANT:   This is not a rip-off, it is decent; it's legal; and it is virtually no risk - it really works!! If all the following instructions are adhered to, you will receive extraordinary dividends.   PLEASE NOTE:  F Please follow the directions EXACTLY, and $50,000 or more can be yoursL in 20 to 60 days. This program remains successful because of the honesty andS integrity of the participants. Please continue its success by carefully adhering to K the instructions. You will now become apart of the Mail Order business. YoutL are in the business of developing Mailing Lists. Many large corporations areJ happy to pay big bucks for quality lists. However, the money made from the mailing lists is secondary to income which is made from people like you and me asking to be included in that list. Here are the four easy steps to success.e  	 STEP ONE:r  ; Get six separate pieces of paper and write the following on 9 each piece of paper "PLEASE PUT ME ON YOUR MAILING LIST." Now get 6 U.S. $1.00 bills and place ONE inside of EACH of the six pieces of paper so the bill will not be seen through the envelope (to prevent thievery). Next, place one paper in each of the six envelopes and seal them. You now should have six sealed envelopes, each with a piece of paper stating the above phrase, your name and address, and a $1.00 bill. What you are doing is creating a service.d   THIS IS ABSOLUTELY LEGAL!!!!!t  H You are requesting a legitimate service and you are paying for it!! LikeN most of us I was a little skeptical and little worried about the legal aspectsR of it all. So I checked it out with the U.S. Post Office (1-800-238-5355) and they# confirmed that it is indeed legal!!r  2 Mail the six envelopes to the following addresses:   1) R. X. 3720 Royal Crest Street #26  Las Vegas, Nevadao   2) Andreas Swart Hoornschediep 11 9727 GA Groningen- The Netherlands-   3) W. Edens- 4829 Bud Ln- Lexington, KY  40514   4) L.Lessard 40 Martins Ferry Rde Hooksett,NH 03106o   5) J. Safian  6950 W. Forest Presrv. Dr., #115 Norridge, IL 60706-1324t   6) A. L. Alexander 12 Briggate 
 Elland BridgeJ Elland HX5 9DPe Englandr     STEP TWO:Now take the #1 name off the list that you see above, move the other names up (six becomes 5, 5 becomes 4, and etc.) and add YOUR NAME as number 6 on the list.   STEP THREE:a Change anything you need to but try to keep this article as close to original as possible. Now post your amended article to at least 200 news groups. :  (I think there are close to 24,000 groups) All you need is 200, but remember, the more you post, the more money you make!! This is perfectly legal!! If you have any doubts, refer to Title 18 Sec. 1302 & 1341 of the Postal Lottery laws. Keep a copy of these steps for yourself and whenever you need money, you can use it again, and again. PLEASE REMEMBER that this program remains successful because of the honesty and integrity of the participants and by their carefully adhering to directions. Look at it this w{ ay. If you were of integrity, the program will continue and the money that so many others have received will come your way.h  @ NOTE: You may want to retain every name and address sent to you,E either on a computer or hard copy and keep the notes people send you.e This VERIFIES that you are truly providing a service.  (Also, it might be a good idea to wrap the $1 bill in dark paper to reduce the risk of mail theft). So, as each post is downloaded and the directions carefully followed, all members will be reimbursed for their participation as a List Developer  with one dollar each. Your name will move up the list geometrically so that when your name reaches the #1 position you will be receiving thousands of dollars in CASH!!! What an opportunity for only $6.00 ( $1.00t  for each of the first six people listed above) Send it now, add your own name to the list and you're in business!!!    6 *****DIRECTIONS FOR HOW TO POST TO NEWS GROUPS!!!*****  F STEP ONE: You do not need to re-type this entire letter to do your ownM posting. Simply put your cursor at the beginning of this letter and drag yourrQ cursor to the bottom of this document, and select 'copy' from the edit menu. Thist7 will copy the entire letter into the computer's memory.h  	 STEP TWO: < Open a blank 'notepad' file and place your cursor at the topN of the blank page. From the 'edit' menu select 'paste'. This will paste a copyJ of the letter into the notepad so that you will add your name to the list.   STEP THREE:y Save your new notepad file as a text file. If you want to do your posting in different settings, you'll always have this file to go back to.  
 STEP FOUR:M You can use a program like "postXpert" to post to all the newsgroeps at once.l You can find this program at  <http://www.download.com>. If you don't understand how it works you can email me at: <mailto:andreass@orange.nl>P (this is only when my name is in the list, so send a copy of my address as well.@ put this in the header: make millions very easy + my full name )  C Use Netscape or Internet Explorer and try searching for various newuB groups (on- line forums, message boards, chat sites, discussions.)  
 STEP FIVE::Visit message boards and post this article as a new message by highlighting the text of this letter and selecting paste from the edit menu. Fill in the subject, this will be the header that everyone sees as they scroll through the list of postings in a particular group, click the post message button. You're done.   Congratulations!!!!!!n  6 THAT'S IT!! All you have to do, and It Really works!!!   Best Wishesn   ------------------------------  % Date: Tue, 09 Dec 2003 03:08:52 -0500 * From: JF Mezei <jfmezei.spamnot@istop.com>" Subject: Re: lib$spawn particulars) Message-ID: <3FD582EE.40A78294@istop.com>    Thien-Thi Nguyen wrote:A< >  - are these args somehow mutually exclusive in their use?@ >    that is, if i specify finish-ast and finish-code, does that@ >    mean the byte-wide-event-flag is ignored?  (or vice versa?)  E No, they are not exclusive. But remember that they only work when thetN subprocess has the NOWAIT flag bit set. Without the bit, the call to lib$spawnC returns control to your program when the suprocess has terminated. S  L With the NOWAIT flag set, then lib$spawn returns immediatly and continues toL execute while your supprocess executes. So your main process then has 2 ways to be notified.1  E >  - is finish-ast guaranteed to be called on subprocess termination?   H No. Consider the situation where the program running in the main processH exits, returning to DCL, but subprocess has not "logged out" yet. (whichM brings up the question: during image exit, does VMS automatically zap the ASTeL information in the subprocess since upon subpocess termination, the AST code no longer exists ?  G >    not called?)  should CLI (speficially DCL) subprocesses be treatedS >    specially in some way?   @ >  - what mechanisms are recommended for the lib$spawn caller toD >    insulate its event flags and mailbox channels from being munged7 >    by the subprocess?  shouldn't that be "automatic"?   J If the input and output of the subprocess are mailboxes controlled by yourY main process, then you need not worry since the user won't have access to the subprocess..  H > a bug whereby GNU Emacs 21.2 (work-in-progress) M-x shell (using PTYs)1 > works fine until i enter the "logout" command.    K does EMACS just give you a subprocess with the $ sign ? I would assume thenoK that emacs doesn't use the NOWAIT flag and while the subprocess is created,eM emacs is inactive in the background ? If so, there should be no event flag ornM AST specified since your main process "sleeps" while the subprocess is actived5 and will only wake up once subprocess is terminated ?C  K When the subprocess terminates and returns control to the main process, areeF you sure you take the proper steps to re-establish context of the mainI application ? (redraw screen, and re-issue the IO request to get the next  input key etc etc ?_    ! What is the subprocess used for ?    ------------------------------  % Date: Tue, 09 Dec 2003 12:17:43 -0500h% From: Thien-Thi Nguyen <ttn@glug.org> " Subject: Re: lib$spawn particulars( Message-ID: <7gu149vrk8.fsf@gnufans.net>  = koehler@eisner.nospam.encompasserve.org (Bob Koehler) writes:l  / > Just make sure sys$output and sys$command arex > what you want them to be.[  2 thanks for the tip.  i had not thought about this.   > Do you map the PTY to an > emacs window?o  H yes, this is the so-called shell (asynchronous DCL)-in-a-buffer feature.  E i notice that "logout/full" displays extra summary info.  would it be H correct to say that the ast call and event flag setting occur *strictly* after this output?  8 > You may need to clean up the PTY, use "SHOW SYSTEM" toD > make sure that the subprocess really exited.  In particular if youA > used a mailbox you never read the subprocess might be in RWMBX.o  G i see the subprocess has exited; it is the parent that is in LEF state.s   thit   ------------------------------  % Date: Tue, 09 Dec 2003 12:48:11 -0500e% From: Thien-Thi Nguyen <ttn@glug.org> " Subject: Re: lib$spawn particulars( Message-ID: <7gn0a1vq5g.fsf@gnufans.net>  , JF Mezei <jfmezei.spamnot@istop.com> writes:  F > during image exit, does VMS automatically zap the AST information inB > the subprocess since upon subpocess termination, the AST code no > longer exists ?   H the ast is only relevant to the originating process, so there's no pointF calling it if that process is gone.  so my guess is yes (and that suchE info really does not reside w/ the subprocess but in some other placebC but that is literally "neither here nor there" -- just a guess :-).u  9 > does EMACS just give you a subprocess with the $ sign ?w# > What is the subprocess used for ?i  C in this case the NOWAIT flag is used to be able to support multiple H asynchronous DCL subprocesses (the so-called shell-in-a-buffer feature).   thit   ------------------------------  % Date: Tue, 09 Dec 2003 01:15:07 -0600y6 From: "Craig A. Berry" <craigberry@mac.com.spamfooler>N Subject: linker collisions and porting (was Re: Freeware download corrupt (?))T Message-ID: <craigberry-F16959.01150709122003@dsl081-159-101.chi1.dsl.speakeasy.net>  ' In article <3fd474a9$1@cpns1.saic.com>,:2  Mark Berryman <Mark.Berryman@Mvb.Saic.Com> wrote:  H > As an aside, how is a freeware porter supposed to know what functions K > might one day find their way into the DECC RTL so (s)he can address this   > appropriately in the port?    F These collisions only happen when the compiler knows about a function @ and adds its default prefix of "decc$", but the presence of the H function varies in different versions of the CRTL.  So it comes down to G what the compiler has on its list, which appears to be based primarily  D on the ANSI C standard.  No doubt that list also varies by compiler G version.  If you check out the docs to CC/PREFIX, you can see that you nE have a choice of prefixing C89 or C99 functions.  However, there are  E plenty of functions that are not in either C standard that appear to nG get the same treatment.  Take utime() for example, which is common and oD may be in POSIX, but is not in C99.  Yet it gets the same prefixing F treatment.  My recommendation is that *any* function that replaces or A jackets a system-supplied function should have its symbolic name o= protected with a macro or prefix pragma, or at least use the d2 /PREFIX=EXCEPT=functionname in your build scripts.  1 The single UNIX specification is documented here:a    @ http://www.opengroup.org/products/publications/catalog/t031s.htm  D which may help in determining whether a function is part of a local D package or is a replacement function for a routine specified by the 	 standard.r   ------------------------------  % Date: Tue, 09 Dec 2003 14:32:49 +0000 0 From: Chris Sharman <chris.sharman@sorry.nospam> Subject: Re: Mailbox deletionP4 Message-ID: <br4mei$h5n$1$8300dec7@news.demon.co.uk>   Bhushan Narkhede wrote:P  F >   This may seem a basic question, but how do we delete the mailboxesD > that are getting created by applications and not cleaned up before
 > exiting.   One of:rH (a) make them temporary, so that they go away when no longer referenced.F (b) if they must be 'permanent' (to get a logical name etc), 'delete' G them immediately after creation, so that they behave as temporary. The e; mailbox & its logical will stick around until dereferenced. H (c) do the cleanup, either (i) inline, or (ii) with an exit handler, or  (iii) with a plv exit routine.5 (d) write a program to delete them (using sys$delmbx)n (e) live with them.- (f) reboot!-  E (a) or (b) are easiest, bullet-proof, and adequate for most purposes. + c(iii) is also bullet-proof, but more work.g= Others are tacky, and would suggest second-rate applications.4   ------------------------------   Date: 9 Dec 2003 06:14:06 -0800 - From: bhushann@hotmail.com (Bhushan Narkhede)  Subject: Mailbox deletiono< Message-ID: <8a3b834.0312090614.60c17397@posting.google.com>   Hi All,e  D   This may seem a basic question, but how do we delete the mailboxesB that are getting created by applications and not cleaned up before exiting.   Rgds,t Bhushand   ------------------------------   Date: 9 Dec 2003 08:40:44 -0600u From: briggs@encompasserve.org Subject: Re: Mailbox deletionh3 Message-ID: <GUCoO71bTOTc@eisner.encompasserve.org>n  l In article <8a3b834.0312090614.60c17397@posting.google.com>, bhushann@hotmail.com (Bhushan Narkhede) writes:	 > Hi All,j > F >   This may seem a basic question, but how do we delete the mailboxesD > that are getting created by applications and not cleaned up before
 > exiting.  F Temporary mailboxes are deleted automatically when the last channel isA deassigned.  So one assumes that the applications in question ared creating permanent mailboxes.   F Your first challenge is identifying these mailboxes and verifying that- the application is, in fact, done using them.   B Perhaps the application uses a particular naming convention on theC mailbox logical names.  If so, the output of $ SHOW LOG /SYSTEM maye
 be of use.  ( Perhaps the mailbox owner UIC is unique.  A Having found a list of mailboxes that are ripe for deletion, whatb< you could do is write a short program to call SYS$DELMBX for	 each one.b   	John Briggs   ------------------------------  $ Date: Tue, 9 Dec 2003 08:49:26 +0100( From: "Rudolf Wingert" <win@fom.fgan.de>C Subject: Re: OpenVMS clusters give Windows, Unix thorough thrashinga: Message-ID: <MCELKPMOKPMNDNKJNIONOECMCHAA.win@fom.fgan.de>   Hello,   John N. did wrote:   >>> J I really really really hate to say this, but I have to agree with at leastK part of what Andrew is saying.  When doing a TCO study, you have to comparevK systems of similar capacity.  To do anything else is ludicrous.   If I have-L an application that needs to be able to handle a specific application, and IG need to select a platform on which to run that system,   I will look at8I systems that will be able to handle my needs.  And they will obvisouly benA systems of similar capacity.  To compare the needs of Boeing to an* supermarket makes very little sense to me.  L Cany anyone verify Andrews assertion that the Sun system he spoke of and theL GS320 are such performance mismatches?  And is that indeed what the Techwise study compared?o <<<d  N Thats not right. You cant compare TCO of an cluster with the TCO of a singleM system. Is Sun able to handle a cluster of 128 nodes with automatic failover?lD AFAIK no!!! Also is Sun able to mirror a cluster within a wide (overP 10kilometres) area? AFAIK also no. As bigger the cluster is, the downtime of theN whole cluster  becomes smaller then of a little cluster. If you will see, thatM an internet seller like Ebay wont have any downtime since his start and know I how many downtime Ebay did have since they use Sun, then you will see the  difference. G On the other side, if you will have a look to the top 500, you will seehN AlphaServer at place four. Which place for a Sun? If you look around, you willH see: the Sun is going down, it is the evening of Sun. I hope for all SunO employer that this will be a short time and the will still have there work in at
 few month.   Best regards Rudolf Wingerto   ------------------------------  % Date: Tue, 09 Dec 2003 04:18:14 -0500f* From: JF Mezei <jfmezei.spamnot@istop.com>C Subject: Re: OpenVMS clusters give Windows, Unix thorough thrashinga) Message-ID: <3FD5932B.7FC741E8@istop.com>    Rudolf Wingert wrote: M > part of what Andrew is saying.  When doing a TCO study, you have to comparee > systems of similar capacity. u  L If discussing TCO, then yes. But when discussing ther cost of downtime, thenI there are many examples where the cost of downtime is not proportional toe system size.  F I believe that availability (no downtimes) was the roginal discussion.   ------------------------------  % Date: Tue, 09 Dec 2003 10:51:50 +0000iO From: Andrew Harrison SUNUK Consultancy <Andrew_No.Harrison_No@nospamn.sun.com>4C Subject: Re: OpenVMS clusters give Windows, Unix thorough thrashing 0 Message-ID: <br49g6$j7t$4@new-usenet.uk.sun.com>   Bob Ceculski wrote:si > "John N." <JNixon@cfl.rr.com> wrote in message news:<Ga%Ab.31537$b01.697588@twister.tampabay.rr.com>...  > L >>I really really really hate to say this, but I have to agree with at leastM >>part of what Andrew is saying.  When doing a TCO study, you have to compare M >>systems of similar capacity.  To do anything else is ludicrous.   If I havenN >>an application that needs to be able to handle a specific application, and II >>need to select a platform on which to run that system,   I will look atbK >>systems that will be able to handle my needs.  And they will obvisouly behC >>systems of similar capacity.  To compare the needs of Boeing to a , >>supermarket makes very little sense to me. >>N >>Cany anyone verify Andrews assertion that the Sun system he spoke of and theN >>GS320 are such performance mismatches?  And is that indeed what the Techwise >>study compared?  >  > ; > why are we stuck on an old gs320 system?  why not comparee9 > a gs1280 ev7 system ... equal number of processors and  * > watch Andrews garbage get blown away ...    < Because you i***t thats what the TechWise study was based on: it didn't compare GS1280's with Sun's or IBM's it compared GS320's.  - Havn't you been paying any attention at all ?l   Regardsw Andrew Harrisoni   ------------------------------  % Date: Tue, 09 Dec 2003 10:56:38 +00006O From: Andrew Harrison SUNUK Consultancy <Andrew_No.Harrison_No@nospamn.sun.com>oC Subject: Re: OpenVMS clusters give Windows, Unix thorough thrashinge0 Message-ID: <br49p7$je2$1@new-usenet.uk.sun.com>   Rudolf Wingert wrote:  > Hello, >=20 > John N. did wrote: >=20 >=20J > I really really really hate to say this, but I have to agree with at le= astiJ > part of what Andrew is saying.  When doing a TCO study, you have to com= pareJ > systems of similar capacity.  To do anything else is ludicrous.   If I = haveJ > an application that needs to be able to handle a specific application, = and IPJ > need to select a platform on which to run that system,   I will look at=  J > systems that will be able to handle my needs.  And they will obvisouly = beC > systems of similar capacity.  To compare the needs of Boeing to ah, > supermarket makes very little sense to me. >=20J > Cany anyone verify Andrews assertion that the Sun system he spoke of an= d the J > GS320 are such performance mismatches?  And is that indeed what the Tec= hwisee > study compared?B > <<<s >=20J > That=B4s not right. You can=B4t compare TCO of an cluster with the TCO = of a single J > system. Is Sun able to handle a cluster of 128 nodes with automatic fai= lover?F > AFAIK no!!! Also is Sun able to mirror a cluster within a wide (overJ > 10kilometres) area? AFAIK also no. As bigger the cluster is, the downti=	 me of theIJ > whole cluster  becomes smaller then of a little cluster. If you will se= e, thatoJ > an internet seller like Ebay won=B4t have any downtime since his start = and knowJ > how many downtime Ebay did have since they use Sun, then you will see t= he
 > difference.   @ Look that isn't what the TechWise study measured, read it before= posting. None of your points have any relevance at all to theX  value or otherwise of the study.  F Incedentally eBays availability over the last couple of years has been& very good, they now have Sun clusters.  J > On the other side, if you will have a look to the top 500, you will see=  J > AlphaServer at place four. Which place for a Sun? If you look around, y= ou willtJ > see: the Sun is going down, it is the evening of Sun. I hope for all Su= neJ > employer that this will be a short time and the will still have there w= ork in a > few month.  C What on earth has this got to do with the TechWise study apart fromnF anything else the TechWise study was measuring OpenVMS TCO and not oneD of the Alpha entries in the Top 500 run OpenVMS, they all run Tru64.   Regardsa Andrew Harrison=   ------------------------------  % Date: Tue, 09 Dec 2003 09:52:28 +0000l0 From: Chris Sharman <chris.sharman@sorry.nospam>+ Subject: Re: OSU http server & http headersn4 Message-ID: <br460s$oc4$1$8300dec7@news.demon.co.uk>  * Alan Winston - SSRL Admin Cmptg Mgr wrote:  i > In article <br210m$554$1$8302bc10@news.demon.co.uk>, Chris Sharman <chris.sharman@sorry.nospam> writes:h > 3 >>We generate PDFs for customer approval/amendment. F >>If customers request amendments, we do them & create new PDFs (same L >>filename). Customers frequently don't see the changes because of cacheing. >>I >>Can anyone suggest what http headers we should be using to ensure that tD >>browsers (especially ie) at least check the file date every time ?: >>And how we should get the OSU HTTP server to send them ? >>6 >>I've got "fileexpire /proofs/*.pdf CDT+00:04:30" in H >>www_root:[system]http_paths.conf, which I thought would do it (expire J >>all PDFs 4.5 mins after creation, forcing a check-back and/or re-fetch). >>	 >>Thanks,n >>Chris  >>! >>Here's an example pdf document:g@ >>http://services.ccagroup.co.uk/proofs/tn5604_devicen_color.pdf >>+ >>Fetch_http returns the following headers:k. >>HTTP/1.0 200 CGI script output data follows. >>Accept-ranges: bytes >>Allow-ranges: bytes . >>Last-Modified: Mon, 08 Dec 2003 13:59:10 GMT >>Content-Length: 111047 >>MIME-version: 1.0r >>Server: OSU/3.9c;UCX >>Content-type: application/pdfn% >>Date: Mon, 08 Dec 2003 14:00:23 GMTn >  > K > You'll note that you're not actually getting an Expires: header, which is @ > one of the two things FileExpire does for you.  (The other is K > make the document expire from OSU's internal cache, forcing the server toaH > do a recheck of the document headers on disk every time you look it up > after the expiration.) > D > But you've got the "Accept-ranges" and "Allow-ranges" headers, so A > I think you've got the byterange presentation script set up foru( > PDF (in HTTP_SUFFIXES.CONF, probably). > E > I don't have time to look at the code now and determine whether therN > presentation script processing circumvents the FileExpire processing, but I  > think it's likely. > M > Try disabling the byterange script - much less important if your customers  K > aren't on dialup links, anyway - and run the Fetch-http test again to seeu$ > whether you get an Expires header.   Thanks - you've got it.f  That's the trick: commenting out0 .EXPAND $pdfbrenab application/pdf $pdfbyterange* from http_suffixes.conf & restarting gets: HTTP/1.0 200 Sending datat MIME-version: 1.0o Server: OSU/3.9c;UCX Content-type: application/pdfe! Content-transfer-encoding: binary  Content-length: 111047, Last-Modified: Mon, 08 Dec 2003 13:59:10 GMT& Expires: Thu, 17 Apr 2003 09:41:59 GMT# Date: Tue, 09 Dec 2003 09:30:30 GMTo  J Don't know how to send the cache-control header (hopefully not important).I Refreshes correctly in ie6, ie5.0, ie4.0, moz1.5 (all on some flavour of l	 Windows).eI But there probably are still a lot of dial-up lines among customers, and bH some of the PDFs are quite large (124k average, up to 7M - 2% over 1M), 1 so I can see byterange being a useful feature ...  Is there a way to get both ?F I could probably hardcode something into byterange.c for my purposes, > but honouring whatever's set in configuration looks beyond me.   ------------------------------  # Date: Tue, 09 Dec 2003 11:14:42 GMTuL From: winston@SSRL.SLAC.STANFORD.EDU ("Alan Winston - SSRL Admin Cmptg Mgr")+ Subject: Re: OSU http server & http headerse6 Message-ID: <00A2A18B.0D7363EB@SSRL.SLAC.STANFORD.EDU>  g In article <br460s$oc4$1$8300dec7@news.demon.co.uk>, Chris Sharman <chris.sharman@sorry.nospam> writes: + >Alan Winston - SSRL Admin Cmptg Mgr wrote:e >tj >> In article <br210m$554$1$8302bc10@news.demon.co.uk>, Chris Sharman <chris.sharman@sorry.nospam> writes: >> p4 >>>We generate PDFs for customer approval/amendment.G >>>If customers request amendments, we do them & create new PDFs (same 3M >>>filename). Customers frequently don't see the changes because of cacheing.o >>>aJ >>>Can anyone suggest what http headers we should be using to ensure that E >>>browsers (especially ie) at least check the file date every time ?s; >>>And how we should get the OSU HTTP server to send them ?s >>>e7 >>>I've got "fileexpire /proofs/*.pdf CDT+00:04:30" in eI >>>www_root:[system]http_paths.conf, which I thought would do it (expire  K >>>all PDFs 4.5 mins after creation, forcing a check-back and/or re-fetch).  >>>s
 >>>Thanks, >>>Chris >>>n" >>>Here's an example pdf document:A >>>http://services.ccagroup.co.uk/proofs/tn5604_devicen_color.pdf  >>> , >>>Fetch_http returns the following headers:/ >>>HTTP/1.0 200 CGI script output data follows.i >>>Accept-ranges: bytese >>>Allow-ranges: bytes/ >>>Last-Modified: Mon, 08 Dec 2003 13:59:10 GMTi >>>Content-Length: 111047w >>>MIME-version: 1.0 >>>Server: OSU/3.9c;UCX/  >>>Content-type: application/pdf& >>>Date: Mon, 08 Dec 2003 14:00:23 GMT >>   >> tL >> You'll note that you're not actually getting an Expires: header, which isA >> one of the two things FileExpire does for you.  (The other is  L >> make the document expire from OSU's internal cache, forcing the server toI >> do a recheck of the document headers on disk every time you look it upi >> after the expiration.)h >> uE >> But you've got the "Accept-ranges" and "Allow-ranges" headers, so eB >> I think you've got the byterange presentation script set up for) >> PDF (in HTTP_SUFFIXES.CONF, probably).  >> wF >> I don't have time to look at the code now and determine whether theO >> presentation script processing circumvents the FileExpire processing, but I   >> think it's likely.l >> gN >> Try disabling the byterange script - much less important if your customers L >> aren't on dialup links, anyway - and run the Fetch-http test again to see% >> whether you get an Expires header.  >9 >Thanks - you've got it.! >That's the trick: commenting outg1 >.EXPAND $pdfbrenab application/pdf $pdfbyterange + >from http_suffixes.conf & restarting gets:t >HTTP/1.0 200 Sending data >MIME-version: 1.0 >Server: OSU/3.9c;UCXp >Content-type: application/pdf" >Content-transfer-encoding: binary >Content-length: 111047e- >Last-Modified: Mon, 08 Dec 2003 13:59:10 GMT ' >Expires: Thu, 17 Apr 2003 09:41:59 GMT $ >Date: Tue, 09 Dec 2003 09:30:30 GMT >sK >Don't know how to send the cache-control header (hopefully not important).eJ >Refreshes correctly in ie6, ie5.0, ie4.0, moz1.5 (all on some flavour of 
 >Windows).J >But there probably are still a lot of dial-up lines among customers, and I >some of the PDFs are quite large (124k average, up to 7M - 2% over 1M), e2 >so I can see byterange being a useful feature ... >Is there a way to get both ?eG >I could probably hardcode something into byterange.c for my purposes, h? >but honouring whatever's set in configuration looks beyond me.r  K I looked at the code. send_document3.c shows that if there's a presentation H script, the cache (and fileexpire processing) is indeed bypassed, which  makes sense.  I I think you're hosed as far as honoring the configuration from byterange, L because it's a separate script and doesn't have ready access to the compiled configuration.    L Presentation scripts go entirely by filetype, not by path, so any change you4 make for PDFs affects all PDFs for the whole server.  J It's not a really happy answer, but I think the best you can easily do is E stick code into byterange.c to compute now plus 04:30 and generate an I Expires: header; you can probably lift that code from send_document3.c or0 http_session.c  L That'll get every .PDF on the system.  You can't get around the suffix thingD by using different suffixes for byterange and non-byterange downloadI because IE ignores the mime-type and goes by suffix, so your .PDFX won't e. fire up Acrobat for your dialup Adobe readers.  K If for some reason you really need to have some PDFs have Expires: headers uF and some not, and still want to use byterange, you might be reduced toH running a second instance of OSU on another port - this involves editingF the startup stuff to use job logicals instead of system logicals - andF configuring the second one to use an unmodified byterange.  It's ugly.  H Or maybe you should just offer your users a choice of non-byteranged PDFH (with path-based expiration) or non-byteranged but faster ZIP (also withI path-based expiration).  I think PDFs have some compression already, so IvL don't know how much you'd win, but it'd show your dial-up customers you were trying.l   --Alan -- rO ===============================================================================u0  Alan Winston --- WINSTON@SSRL.SLAC.STANFORD.EDUM  Disclaimer: I speak only for myself, not SLAC or SSRL   Phone:  650/926-3056wM  Paper mail to: SSRL -- SLAC BIN 99, 2575 Sand Hill Rd, Menlo Park CA   94025@O ===============================================================================c   ------------------------------  % Date: Tue, 09 Dec 2003 14:04:15 +0000$0 From: Chris Sharman <chris.sharman@sorry.nospam>+ Subject: Re: OSU http server & http headersd4 Message-ID: <br4kp0$pdg$1$830fa7b3@news.demon.co.uk>  * Alan Winston - SSRL Admin Cmptg Mgr wrote:  i > In article <br460s$oc4$1$8300dec7@news.demon.co.uk>, Chris Sharman <chris.sharman@sorry.nospam> writes:t > , >>Alan Winston - SSRL Admin Cmptg Mgr wrote: >> >>j >>>In article <br210m$554$1$8302bc10@news.demon.co.uk>, Chris Sharman <chris.sharman@sorry.nospam> writes:5 >>>>We generate PDFs for customer approval/amendment.7H >>>>If customers request amendments, we do them & create new PDFs (same N >>>>filename). Customers frequently don't see the changes because of cacheing. >>>>K >>>>Can anyone suggest what http headers we should be using to ensure that rF >>>>browsers (especially ie) at least check the file date every time ?< >>>>And how we should get the OSU HTTP server to send them ? >>>>8 >>>>I've got "fileexpire /proofs/*.pdf CDT+00:04:30" in J >>>>www_root:[system]http_paths.conf, which I thought would do it (expire L >>>>all PDFs 4.5 mins after creation, forcing a check-back and/or re-fetch). >>>> >>>>Thanks,!	 >>>>ChrisE >>>># >>>>Here's an example pdf document:nB >>>>http://services.ccagroup.co.uk/proofs/tn5604_devicen_color.pdf >>>>- >>>>Fetch_http returns the following headers: 0 >>>>HTTP/1.0 200 CGI script output data follows. >>>>Accept-ranges: bytes >>>>Allow-ranges: bytesi0 >>>>Last-Modified: Mon, 08 Dec 2003 13:59:10 GMT >>>>Content-Length: 111047 >>>>MIME-version: 1.0f >>>>Server: OSU/3.9c;UCX! >>>>Content-type: application/pdf ' >>>>Date: Mon, 08 Dec 2003 14:00:23 GMT0 >>> L >>>You'll note that you're not actually getting an Expires: header, which isA >>>one of the two things FileExpire does for you.  (The other is lL >>>make the document expire from OSU's internal cache, forcing the server toI >>>do a recheck of the document headers on disk every time you look it upo >>>after the expiration.)t >>>wE >>>But you've got the "Accept-ranges" and "Allow-ranges" headers, so aB >>>I think you've got the byterange presentation script set up for) >>>PDF (in HTTP_SUFFIXES.CONF, probably).h >>>uF >>>I don't have time to look at the code now and determine whether theO >>>presentation script processing circumvents the FileExpire processing, but I o >>>think it's likely.w >>>sN >>>Try disabling the byterange script - much less important if your customers L >>>aren't on dialup links, anyway - and run the Fetch-http test again to see% >>>whether you get an Expires header.i >> >>Thanks - you've got it.l" >>That's the trick: commenting out2 >>.EXPAND $pdfbrenab application/pdf $pdfbyterange, >>from http_suffixes.conf & restarting gets: >  >>HTTP/1.0 200 Sending data, >>MIME-version: 1.0u >>Server: OSU/3.9c;UCX >>Content-type: application/pdfa# >>Content-transfer-encoding: binaryt >>Content-length: 111047. >>Last-Modified: Mon, 08 Dec 2003 13:59:10 GMT( >>Expires: Thu, 17 Apr 2003 09:41:59 GMT% >>Date: Tue, 09 Dec 2003 09:30:30 GMTn >>L >>Don't know how to send the cache-control header (hopefully not important).K >>Refreshes correctly in ie6, ie5.0, ie4.0, moz1.5 (all on some flavour of t >>Windows).yK >>But there probably are still a lot of dial-up lines among customers, and  J >>some of the PDFs are quite large (124k average, up to 7M - 2% over 1M), 3 >>so I can see byterange being a useful feature ...  >>Is there a way to get both ?H >>I could probably hardcode something into byterange.c for my purposes, @ >>but honouring whatever's set in configuration looks beyond me.  6 Created a modified version: byterange_with_expires.exe Modified pdf's to use it.xH It has access to filespec, cdt, etc, so I can encode my requirements in G there. Not at all general/portable, but easy to do & functional for me.   B As far as PDF compressibility goes, it's largely dependent on the F embedded graphics (mostly jpeg in our case) - zip saves mostly around A 6-20%, which doesn't seem worth the extra complication & loss of i accessibility.   Thanks all,p Chrisf   ------------------------------  $ Date: Tue, 9 Dec 2003 09:16:02 +0100- From: "Martin Vorlaender" <mv@pdv-systeme.de> > Subject: Re: OT Very scary: Cars running on Microsoft software9 Message-ID: <br40cb$27t5v6$1@ID-56200.news.uni-berlin.de>e   Michael Moroney wrote: > William Webb wrote:hE >> All these posts and nobody came up with "Blue Dashboard of Death?") >> s >> I'm disappointed. > " > How about Blue Cockpit of Death? > G > I saw on someone's jokes page a photo of some modern aircraft cockpitaD > with all kinds of knobs and displays -- including several screens,D > each displaying a BSOD.  The caption was something like "We assure > you, flying is very safe."  @ Reminds me of a cartoon in c't (famous german computer magazine)? where the screen in the cockpit in approach to the runway readsa7 "Please update to version 2.0 to use landing gear". ;-)e   cu,t   Martin -- tF   OpenVMS:                | Martin Vorlaender  |  VMS & WNT programmer3    The operating system   | work: mv@pdv-systeme.derF    God runs the           |   http://www.pdv-systeme.de/users/martinv/:    earth simulation on.   | home: martin@radiogaga.harz.de   ------------------------------   Date: 9 Dec 2003 14:52:52 GMT,( From: bill@cs.uofs.edu (Bill Gunshannon)> Subject: Re: OT Very scary: Cars running on Microsoft software: Message-ID: <br4nk4$288qts$1@ID-135708.news.uni-berlin.de>  * In article <br4n7v$5ma$1@reader10.wxs.nl>,/ 	"Patrick Coulier" <patrick@delight.be> writes:i >  > No MicroSh*t in my car ever !   E I drive a 1978 MGB and it has an older distributor so it doesn't evennC have the electronic ignition module.  Just drove it to GA and back,eE 1500 miles.  Of course, I did take my laptop with me in the car.  :-)i   bill   -- (J Bill Gunshannon          |  de-moc-ra-cy (di mok' ra see) n.  Three wolvesD bill@cs.scranton.edu     |  and a sheep voting on what's for dinner. University of Scranton   |A Scranton, Pennsylvania   |         #include <std.disclaimer.h>   e   ------------------------------  $ Date: Tue, 9 Dec 2003 15:43:34 +0100, From: "Patrick Coulier" <patrick@delight.be>> Subject: Re: OT Very scary: Cars running on Microsoft software* Message-ID: <br4n7v$5ma$1@reader10.wxs.nl>  7 "JF Mezei" <jfmezei.spamnot@istop.com> wrote in messagew# news:3FCD35BC.836E30AC@istop.com...n1 > Unfortunatly, this is NOT an April Fool's joke.  >UE > http://wired.com/news/autotech/0,2554,61412,00.html?tw=wn_tophead_6o >  > ##J > Cars with the Microsoft software will speak up when it's time for an oilG > change. They'll warn drivers about wrecks on the road ahead and scout F > alternative routes. They'll pay freeway tolls automatically. ****The software: > running their brakes will upgrade itself wirelessly **** >eI > The Microsoft platform already is in 23 different car models, includingt theoK > BMW 7 series, Citroen, Daimler, Fiat, Volvo, Hyundai,  Mitsubishi, Subaru. andT	 > Toyota.  > ##  ; This is an interesting article concerning these matters ...t  C http://the.honoluluadvertiser.com/article/2003/Nov/23/bz/bz08a.html    No MicroSh*t in my car ever !a   Ce   ------------------------------    Date: 09 Dec 2003 07:12:35 -0800( From: Javier Henderson <javier@KJSL.COM>> Subject: Re: OT Very scary: Cars running on Microsoft software- Message-ID: <86k7562ffg.fsf@skylane.kjsl.com>e  9 moroney@world.std.spaamtrap.com (Michael Moroney) writes:h   > William Webb wrote:lF > > All these posts and nobody came up with "Blue Dashboard of Death?" > >  > > I'm disappointed.e > " > How about Blue Cockpit of Death? > G > I saw on someone's jokes page a photo of some modern aircraft cockpithI > with all kinds of knobs and displays -- including several screens, eacheK > displaying a BSOD.  The caption was something like "We assure you, flyingt > is very safe."  F There is a number of cockpit devices running Windows these days. I hadF an UPSAT MX20 MFD in my Grumman Tiger, for example, which runs embeded" NT. I never did experience BSOD...  D I would not, however, install a device on which I'd have to rely forB navigating inside clouds, which would be Windows based. No way, no how.   -jav   ------------------------------  % Date: Tue, 09 Dec 2003 15:21:03 +0000 O From: Andrew Harrison SUNUK Consultancy <Andrew_No.Harrison_No@nospamn.sun.com>i> Subject: Re: OT Very scary: Cars running on Microsoft software0 Message-ID: <br4p90$osi$1@new-usenet.uk.sun.com>   Bill Gunshannon wrote:, > In article <br4n7v$5ma$1@reader10.wxs.nl>,1 > 	"Patrick Coulier" <patrick@delight.be> writes:I >  >>No MicroSh*t in my car ever !c >  > G > I drive a 1978 MGB and it has an older distributor so it doesn't eveneE > have the electronic ignition module.  Just drove it to GA and back, G > 1500 miles.  Of course, I did take my laptop with me in the car.  :-)s >   . I have a 1962 Austin Healey 3000 in the garage  F Lights, indicators, ignition, starter motor, fan for heater (of sorts)D Imobiliser !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! (well itsA just a switch really) thats about the sum total of the electronics gizomos.  D Odd, its great fun to drive and thats without SAT NAV, Self levelingB suspension, sequential gears, traction control, power hood and allC the other gubbins that car makers tend to add to their cars to makeL them heavier and slower.  > A friend who bought a Lotus Elise recalled the conversation he had with the Lotus salesman.  3 Friend, can you have electic windows on the Elise ?f. Salesman, do they make the car go any faster ?" Friend, how about power stearing ?1 Salesman, would that make the car go any faster ?o Friend, so no power hood3 Salesman, I think you are beginning to get the ideav  A The nice thing is that one of the UK auto mags recently ran theiri@ annual competition for the best handling car available on the UKC market, competition was stiff, Porsche GTE, Caterham, Lotus, Subarue? Impreza. Guess what won ? a Mazda MX-5 and the panel included a D current F1 driver plus a number of retired F1 drivers and designers.  > I say nice because the Mazda and the Healey come from the same tradition of car design.  < The idea that putting a windows OS in any of the cars in the: competion would have enhanced its drivability is of course
 laughable.   regardsr Andrew Harrisono > bill >    ------------------------------  % Date: Tue, 09 Dec 2003 18:35:00 +0100l From: Dirk Munk <munk@home.nl>> Subject: Re: OT Very scary: Cars running on Microsoft software2 Message-ID: <br51hm$b1f$1@news2.tilbu1.nb.home.nl>   Martin Vorlaender wrote: > Michael Moroney wrote: >  >>William Webb wrote:y >>E >>>All these posts and nobody came up with "Blue Dashboard of Death?"v >>>s >>>I'm disappointed. >>" >>How about Blue Cockpit of Death? >>G >>I saw on someone's jokes page a photo of some modern aircraft cockpittD >>with all kinds of knobs and displays -- including several screens,D >>each displaying a BSOD.  The caption was something like "We assure >>you, flying is very safe." >  > B > Reminds me of a cartoon in c't (famous german computer magazine)A > where the screen in the cockpit in approach to the runway readse9 > "Please update to version 2.0 to use landing gear". ;-)   ? Or the (later) cartoon where a laptop shows the following text: 8 "Windows discovered a new bluetooth device: Airbus A310"+ "Would you like to initialize this device?"s   >  > cu,R
 >   Martin   ------------------------------  # Date: Tue, 09 Dec 2003 12:21:29 GMT " From:   VAXman-  @SendSpamHere.ORG) Subject: Re: porting problems encounterede0 Message-ID: <00A2A1AD.8E78389C@SendSpamHere.ORG>  f In article <bMednZQHMtCspkii4p2dnA@comcast.com>, "Richard B. Gilbert" <rgilbert88@comcast.net> writes:I >A great deal depends on the code being ported.  Porting can be anything mH >from a breeze to a nightmare.  The nightmare end of things is typified   K ...and it can be a trip through legal hell if some satan claims he inventeds' porting.  Been there, sorry I had been.o   --K http://www.legacy-2000.com  for the best OpenVMS system security solutions.e  bK VAXman- A Bored Certified VMS Kernel Mode Hacker   VAXman(at)TMESIS(dot)COMr            g5   "Well my son, life is like a beanstalk, isn't it?" i   ------------------------------   Date: 9 Dec 2003 09:30:28 -0800e$ From: rcbryan@hotmail.com (RC Bryan)) Subject: Re: porting problems encounteredc< Message-ID: <fbcf38dc.0312090930.2851952@posting.google.com>   > > char * sub(char *p)- > > {e > >    int myP=p;0> > >    myP = (myP+15)& 0xfffffff0;/* add and mask back down */ > >    p=(char *)myP; 	 > > . . .a > >  > > G > What exactly happens here ? Why are we doing this & 0xffff... ? If we1H > add my+1 then value would be myP + 1*sizeof(myP). Please correct me if > i am wrong .  E You are wrong.  This example is doing integer arithmetic on a pointer,F value.  We are adding 15 to myP, an integer, not p, the pointer.  If pD had the value 0x40000F44, by adding 15 (decimal) (or 0xF hex) we getA 0x40000f53.  We then mask off the low order bits to get a 16 byte D bounded space.  0x40000f53& 0xfffffff0 == 0x40000f50.  If we had not@ added 15, we would have 0x40000f40 which is a pointer to a spaceF before the start of the p area. (Don't forget C lesson 1, & is bitwiseF and && is logical (the whole variable).)  I actually saw this where weD needed page bounded space which is particularlly not portable.  PageD bounding or 16 byte bounding is the same idea except you can use 511@ and 0xfffffe00 rather than 15 and 0xfffffff0 (on some machines).   > ! > > Using the struct from before:a > > % > >     p=malloc(sizeof(mystruct)*2);  > >     p+=sizeof(int);m; > >     intVal=((struct MYSTRUCT*)p)->val; /* val is long*/e > > G > > This will run with a warning on some platforms, crash on others andl > >o2 > Is the problem due to assigning of long to int ?  D No, it is due to trying to extract a long value from a value that isB not properly bounded.  32 bit machines like to fetch 32 bit valuesF from addresses that end in 0,4,8, and c, 32 bit bounded addresses.  32E bit machines are fine with fetching 64 bit values from 32 bit boundedeC addresses since it requires two operations internally.  On a 64 bitcF machine, they like to fetch 64 bit values from addresses that end in 0D and 8.  On many machines, if you try to fetch values from improperly@ bounded addresses, you will get an exception.  Interestingly, itC depends upon the compiler and the O/S.  I have seen things pass on,s? say Alpha-VMS and fail on Alpha-Tru64, the same hardware with aoB different operating system.  Since this is a VMS newsgroup, I willD note that one of the neat things about the original VAX was that youD could fetch anything from any boundary.  If I recall correctly, thisC included arbitrary bit fields, you were not even restricted to bytem boundaries.r   Here is another good one:e   float a=1.0; printf("%x", a);  A This should give 0x3f800000 (for IEEE floating point) but on many # machines it will give anything but.i  F The reason why has to do with the way the compiler passes arguments toE subroutines.  On many machines, it passes floating point arguments in E floating point registers.  Meanwhile, the printf subroutines does notiF know anything about the type of the argument and prints out what is in( an integer register and you get garbage.  D It took a lot of head scratching before I figured out what was going= on the first time I saw this one.  The solution is to have a:o   union  {     float f;0	    int i;3 }ftoi;  I > Thanks [RC] Bryan for the long post. Most it was useful for me . I have + > some doubts in understanding of the code.d  C I am afraid to understand some of the hardware problems you have todA have studied the machine code to see what is going on.  Sometimes0F understanding THAT requires understanding what the hardware is doing. F When I was in school, some of the guys would complain, "I am not goingF to be coding in assembly, why do I have to learn this?"  The answer isF that like it or not, at some point you will have bugs that can only beF resolved by understanding what the machine is doing.  (This went alongF with the Mechanical Engineering students who complained to me (anotherE ME who just happened to work for Computer Services) "Why do I have toeB learn to program? I am an ME."  I had a lot of sympathy for them.)   Regards,	 /RC Bryanl   ------------------------------  % Date: Tue, 09 Dec 2003 08:42:19 +0100 : From: Karl Rohwedder <extern.karl.rohwedder@volkswagen.de>/ Subject: Re: printing 11x17 landscape documentsi0 Message-ID: <br3ubg$ep515@doiweb4.volkswagen.de>  * Alan Winston - SSRL Admin Cmptg Mgr wrote: > VMSers --u > L > VMS Alpha 7.3-1 (soon to go to 7.3-2), DCPS 2.2.  Currently printing on an
 > HP 5500. > O > I have plain ASCII documents which are (currently) in a 132-column format and-M > have pages of varying length, up to about 130 lines.  (There are form-feeds" > at new pages.) > 0 > For years it's been good enough to print with  > K >  /param=(number_up=2,page_orientation=landscape,input_tray=the-right-try)e > M > which produces a physical sheet with two page images on it, each with a boxhF > around it, with the page image in fairly-small but good-enough type. > N > My users would now like to get more columns onto the physical page, maybe asO > many as 150.  I think the  hardware can obviously do it, or it couldn't printiJ > the box.  Simply extending the lines and printing with the above command > produces ugly line-wrap. > O > Is there a DCPS-style way to get longer line lengths and page lengths?  Can IlQ > direct it to treat the 11x17 page as a single 132-line page (and thus avoid theeO > number_up and the box)?  Can I somehow tell it to give me more columns in thef > page?  > Q > When I had to solve a similar problem a long time ago I ended up embedding LN03pM > escape sequences in the document, which were then handled by the LN03-to-PSTM > translator in DCPS.  This seems like a silly way to have to go about it. IfaQ > there's no secret DCPS incantation for it, can anyone share a setup module thatr > will do this for me?   >  > Summarizing: > I > I want to print 11x17 pages that are 150 columns by 132 lines on a DCPSlM > printer.  What's the best way to do it, and if it takes code (postscript orn4 > escape sequences), does anybody have any to share? > 	 > Thanks,  > 	 > -- AlanSE We have a form to print in landscape format with 185 characters/line.l The form is defined as:0  J DCPS_QUER (stock=DEFAULT)               142   DINA4 Querformat 185-spaltigN      /LENGTH=66 /MARGIN=(BOTTOM=6) /SETUP=(DCPS_QUER) /STOCK=DEFAULT /TRUNCATE      /WIDTH=185l   The setup module is:/ ESC>[!p<ESC>[15m<ESC>[?23 J<ESC>[4w<ESC>[1;256s    Note: works for ASCII text!R --    + mit freundlichen Gren | with best regardsa  3 Karl Rohwedder          | it-ingteam(at)t-online.defA                          | extern.karl.rohwedder(at)volkswagen.des   ------------------------------  # Date: Tue, 09 Dec 2003 10:09:57 GMT 5 From: rdeininger@mindspringdot.com (Robert Deininger).- Subject: Re: REPLY/STATUS in another VMS node L Message-ID: <rdeininger-0912030511510001@user-105n88c.dialup.mindspring.com>  O In article <SJ3Bb.182286$361.13785@news.chello.at>, peter@langstoeger.at wrote:i  > >In article <ddf392ea.0312051508.3ec4155c@posting.google.com>,/ contracer11@uol.com.br (Shiva MahaDeva) writes: A >>I need issue a REPLY/STATUS in another VMS machine (in network) A >>to see if there is any pending request; How could I make this ?_ >n, >Login remotely and then issue the command ? >iI >And for the SYSMAN and RSH suggestors, have you ever tried it yourself ?uI >Neither SYSMAN, nor REXEC has a terminal device on the remote end, whichnJ >REPLY wants (and iff they had one, they wouldn't exist long enough to get >the asynchronous answer).  G This doesn't directly solve the original problem, but here's a method InH use to issue REPLY commands when I don't have a terminal device.  I makeD one.  Here's a snippet from a procedure I use to manage OPERATOR.LOG files...  G $! In order to start a new operator log file, the process's SYS$COMMANDpG $! has to be enabled as an operator terminal.  We don't have a terminallC $! if we are running in a batch or detached process.  So we have to  $! make one. $tJ $! The method used here is fairly hardware independent.  It depends on theJ $! LAT protocol being started, which in turn requires at least one network
 $! device. $t? $! The REPLY command works with whatever device is SYS$COMMAND.w $ A $! Create a LAT port pseudo-terminal to use for REPLY operations.o $ latcp := $latcph> $ latcp create port/application/logical=(name=oplog_oper_port) $s$ $ define sys$command oplog_oper_port $ reply/enable $ reply/log  $ reply/disablet $ deassign sys$command $ # $ latcp delete port oplog_oper_portu $ deassign oplog_oper_port      I The point of this code is to do a REPLY/LOG, but REPLY/STATUS should worki9 as well.  The remaining problem is to capture the output.   H LAT is included with VMS, all you have to do it turn it on, and you will8 be able to create fake terminals whenever you need them.   ------------------------------  % Date: Tue, 09 Dec 2003 04:02:47 -0500t* From: JF Mezei <jfmezei.spamnot@istop.com> Subject: Re: SEARCH enhancement ) Message-ID: <3FD58F8C.C2EBFF78@istop.com>o   Dave Greenwood wrote:aI > So now we'll have the VMS TYPE command which displays the contents of asI > file - unless you want to start in the middle in which case you use theg > SEARCH/SKIP command.  Ugh!    L TYPE/TAIL is one of the best recent improvements to DCL.  The person who put4 that in deserves a large piece of chocolate cake :-)   ------------------------------  % Date: Tue, 09 Dec 2003 04:10:51 -0500e* From: JF Mezei <jfmezei.spamnot@istop.com> Subject: Re: SEARCH enhancemente) Message-ID: <3FD59170.40BA5B36@istop.com>a   Graham Burley wrote:N > > Do you happen to know if GRAB works on CSWS log files?  (TYPE/TAIL doesn't4 > > work because they're not in a supported format.) > >n > F > Haven't tried, but it works fine with Stream_LF files if that's what > they are.r  N TYPE/TAIL has difficulties with *certain* files. I think once it hits a recordF that exceeds a certain record length, then it barfs with a nasty error7 message. It is also unwilling to work on indexed files.t   ------------------------------  % Date: Tue, 09 Dec 2003 10:50:19 +0100r9 From: Jan-Erik =?iso-8859-1?Q?S=F6derholm?= <aaa@aaa.com>l Subject: Re: SEARCH enhancemento' Message-ID: <3FD59ADB.A8E5594C@aaa.com>r   JF Mezei wrote:h > N > TYPE/TAIL is one of the best recent improvements to DCL.  The person who put6 > that in deserves a large piece of chocolate cake :-)    E And "TYPE/TAIL/CONT/INT=n" together with a "SET OUTPUT_RATE=00:00:05"n= is *realy* nice for debugging (or just tracing) batch jobs...   	 Jan-Erik.o   ------------------------------  # Date: Tue, 09 Dec 2003 16:03:45 GMTe0 From: "Guy Peleg" <guy.peleg@hp.com_remove_this> Subject: Re: SEARCH enhancement 3 Message-ID: <BnmBb.10372$2K5.1932@news.cpqcorp.net>l  K My intention was to implement /limit and /skip, only after I implemented ittB Mike Rechtman pointed to me that it can be used for typing a file.  B There is no intention in the utilities group to make this a trend.   Guyv8 "Dave Greenwood" <greenwoodde@ornl.gov> wrote in message+ news:8DEC03.20030555@feda34.fed.ornl.gov...oJ > In a previous article, "Guy Peleg" <guy.peleg@hp.com_remove_this> wrote:I > > Here is a sneak preview of a new enhancement to search that will shipa with
 > > V8.2 : > >  > > BLUSKY> ty test.txt- > > first line > > second line0 > > third line > > forth line > > fifth line > > sixth line > >jD > > Now let's search the file for the word line, limit the number of- > > matches to 2 and skip the first 4 matches , > > BLUSKY> sea test.txt/limit=2/skip=4 line > > fifth line > > sixth line > >rG > > One of the things I like about this is that you can now type a filefH > > from the middle.  Let's say I want to type the file starting the 4th
 > > line : > >wA > > BLUSKY> search test.txt/skip=3/match=nor "nonexistancestring"e > > forth line > > fifth line > > sixth line > > BLUSKY>a > >wE > > Coming soon to a system near you......(actually not so soon.....)  >c > Neat!e >tL > But your last example bothers me a bit because it's so close to one of theK > things I dislike about unix.  That's the frequency that you have to use ahK > different tool than the one you expect to use.  Eg, you use "ls" to get aeL > directory listing.  Unless, of course, you want a listing of files changed2 > in the last 3 days in which case you use "find". >eI > So now we'll have the VMS TYPE command which displays the contents of atI > file - unless you want to start in the middle in which case you use thec > SEARCH/SKIP command.  Ugh! >sH > Unix, where's there's always another way - and you **have** to use it! >1 > Dave > --------------; > Dave Greenwood                Email: Greenwoodde@ORNL.GOV>J > Oak Ridge National Lab        %STD-W-DISCLAIMER, I only speak for myself   ------------------------------  # Date: Tue, 09 Dec 2003 16:06:21 GMTi0 From: "Guy Peleg" <guy.peleg@hp.com_remove_this> Subject: Re: SEARCH enhancementn2 Message-ID: <1qmBb.10373$iL5.253@news.cpqcorp.net>  I On V7.3-2 I fixed TYPE/TAIL to support large records. Give it another tryn and let me know.   Guyt7 "JF Mezei" <jfmezei.spamnot@istop.com> wrote in messageh# news:3FD59170.40BA5B36@istop.com...t > Graham Burley wrote:H > > > Do you happen to know if GRAB works on CSWS log files?  (TYPE/TAIL doesn'td6 > > > work because they're not in a supported format.) > > >i > >rH > > Haven't tried, but it works fine with Stream_LF files if that's what
 > > they are.  >eI > TYPE/TAIL has difficulties with *certain* files. I think once it hits al recordH > that exceeds a certain record length, then it barfs with a nasty error9 > message. It is also unwilling to work on indexed files.    ------------------------------  $ Date: Tue, 9 Dec 2003 12:43:49 -0500+ From: "Martin O'Connor" <moconnor@dvfs.com>2 Subject: Re: SEARCH enhancement.: Message-ID: <br51km$25rj4r$1@ID-118202.news.uni-berlin.de>  ; "Guy Peleg" <guy.peleg@hp.com_remove_this> wrote in messageM- news:BnmBb.10372$2K5.1932@news.cpqcorp.net...:M : My intention was to implement /limit and /skip, only after I implemented itnD : Mike Rechtman pointed to me that it can be used for typing a file. :sD : There is no intention in the utilities group to make this a trend. :d : Guy : : "Dave Greenwood" <greenwoodde@ornl.gov> wrote in message- : news:8DEC03.20030555@feda34.fed.ornl.gov...eL : > In a previous article, "Guy Peleg" <guy.peleg@hp.com_remove_this> wrote:K : > > Here is a sneak preview of a new enhancement to search that will shiph : with : > > V8.2 : : > >u    b I have for a very long time used search to "type" a file by searching for a null string. Among the_ advantages is using the /number to get the record numbers output with the lines (also used with \ explicite strings to see their relative positions of where the string is found in the file).   MartyR   ------------------------------  # Date: Tue, 09 Dec 2003 15:49:56 GMTk9 From: Hein van den Heuvel <hein_netscape@eps.zko.dec.com>a+ Subject: Re: SEARCH enhancement [type/tail] / Message-ID: <3FD5EDD1.1B84461B@eps.zko.dec.com>o  < Type/tail was a good attempt, but IMHO not aggresive enough.O Not enough heuristics built in... or maybe, while late, still too early for itsn time. 1 There are far too many files it will not work on. O Possibly it should learn to look at the FIRST records in the file to understandyN waht they look like to be able to latch onto a good record start after pickingR a random start point at the end of the file. RMS itself has some code build in for	 that now. O [1997-ish. Section 3.16.8 in the OpenVMS V7.1 New Features Manual describes the K  then new SYSGEN parameters RMS_HEURISTIC and RMS_DFLRL SYSGEN parameters.]e    P Type / tail on an indexed file can now possibly be implemented since RMS has the" read-backwards option (RAB$V_REV).L It would be key order based, not time based. It woudl proabbyl need a /INDEX switch to select an index.     JF Mezei wrote:h   > Graham Burley wrote:P > > > Do you happen to know if GRAB works on CSWS log files?  (TYPE/TAIL doesn't6 > > > work because they're not in a supported format.) > > >  > >nH > > Haven't tried, but it works fine with Stream_LF files if that's what
 > > they are.p >aP > TYPE/TAIL has difficulties with *certain* files. I think once it hits a recordH > that exceeds a certain record length, then it barfs with a nasty error9 > message. It is also unwilling to work on indexed files.    ------------------------------  % Date: Tue, 09 Dec 2003 03:43:51 -0500 * From: JF Mezei <jfmezei.spamnot@istop.com>B Subject: Re: Sun to use AMD Opteron - announcement expected Monday) Message-ID: <3FD58B1F.BE20CE61@istop.com>n   Fred Kleinsorge wrote:L > DECwrite was a modified version of a buyout that was intended to be common4 > for Windows, UNIX and VMS.  The marketplace spoke.  J As far as I remember, DECwrite was unleashed before Windows was capable ofN working in a lan. I specifically remember attending a presentation on the "newK workstations", that included the annoucnement of the DDIF architecture withtL coumpound documents etc.  I doubt very much that Word was more powerful thanJ Decwrite at that time. And this was at the time when Digital had made thatM fancy video about the compound document stuff with the then hit song from theyK group "information society" that showed how you could share documents, have L one work on the text while the other worked on the spreadsheet etc etc.  HadL they taken that video and put it on TV, it might have meant a very differentJ life for Digital. But they only showed it to the installed base and we allN know that preaching to the choir when your customer base is decreasing doesn't save the company.u    A DECwrite failed because Digital refused to price its workstations B competitively with PCs andprovide real marketing for its products.  N Imagine if they had make a DECwidows version of FMS and ALLIN1 with integratedM DECwrite. DEC would have had a huge winning product, years before Lotus notesfJ or Office. Of course, pointless to develop all that if you're not going to
 market it.   ------------------------------  % Date: Tue, 09 Dec 2003 12:31:49 +0000nO From: Andrew Harrison SUNUK Consultancy <Andrew_No.Harrison_No@nospamn.sun.com>sB Subject: Re: Sun to use AMD Opteron - announcement expected Monday0 Message-ID: <br4fbl$leq$1@new-usenet.uk.sun.com>   Fred Kleinsorge wrote:M > "Andrew Harrison SUNUK Consultancy" <Andrew_No.Harrison_No@nospamn.sun.com>n= > wrote in message news:br1knm$jrv$1@new-usenet.uk.sun.com...n >  >>Keith Parris wrote:b >>8 >>>JF Mezei <jfmezei.spamnot@istop.com> wrote in message > ' > news:<3FCFB81D.F75699CC@istop.com>...t > J >>>>Based on comments I have heard here, it seems to me that there is very >  > little > J >>>>cross polination between the VMS and Tandem/NSK folks as well as HP-UX >  > with >  >>>>regards to systems design. >>>n >>> C >>>VMS, NSK, and Tru64 share IP stack code.  HP-UX will, of course,xB >>>benefit from VMS Cluster technology by way of TruClusters code. >>>o >>A >>Ummm are you really really sure about TruCluster making it intow8 >>HP-UX, at the moment the project doesn't look to safe. >> >  > M > Yes.  Since I see lots of the traffic on AdvFS and TruClusters - you may be N > listening to sources of NIH - *you* shouldn't put too much faith or FUD into > them.U >  >   B Ohh dear, the fact that this information surfaced in a letter fromG Rich Marcello to your customers is going to rather ruin you point isn'tu it.f  > AdvFS and TruCluster require HP-UX 11i v3 the letter says that' 11i v3 has been delayed by over a year.t  I >>>The designers of Itanium systems have been very helpful and interesteds5 >>>in meeting the needs of OpenVMS in system designs.e >> > K > Unfortunately, because of FUD from people like Andrew - we sometimes feele% > compelled to point out the obvious.= >   B Unfortunately because people like you Fred fail to check the factsA before posting, you end up committing the tried and trusted errork= for any member of the choir of describing uncomfortable factsp as FUD.o   Your mistake again.2   regardsl Andrew HarrisonC   ------------------------------  % Date: Tue, 09 Dec 2003 14:25:01 +0000=O From: Andrew Harrison SUNUK Consultancy <Andrew_No.Harrison_No@nospamn.sun.com>4B Subject: Re: Sun to use AMD Opteron - announcement expected Monday0 Message-ID: <br4lvu$nkp$1@new-usenet.uk.sun.com>   Fred Kleinsorge wrote:K > I restored the quote that Keith responded to, which clearly shows you aree= > incorrect.  Your rebuttal was a non-responsive redirection.. > L > But in answer your rebuttal, yes you can buy a VMS workstation - and thereF > are more than a few people who do.  But the fact is that there is noN > competition for the general "workstation" market any more - Windows has won.F > What remains of the UNIX (and VMS) workstation market is legacy, andN > specialized - or used as SW development stations.  The fact is that there isL > precious little in that market that can't be done with (one or more) PC's.M > It is overwhelmingly Windows.  Even OS's like Solaris or Linux are noise ine > this market. >   G It may suprise you then to know that the UK NHS who are Europes largeste= employer with over 1 million employees are running a trail to[D replace Windows with Sun's JDS which is based on SuSE Linux and thatG the UK's Office of Government Commerce is also trialing JDS for another>! 500,000 desks for civil servants.>  G A Spokesman for the OGC has also said that if JDS trials are sucessfullfD then the civil service roll out could be extended to other state run0 bodies such as the Police and Education sectors.  $ Noise may be but quite a loud noise.   Regardsg Andrew Harrisona   ------------------------------  % Date: Tue, 09 Dec 2003 10:49:52 +0000>O From: Andrew Harrison SUNUK Consultancy <Andrew_No.Harrison_No@nospamn.sun.com> F Subject: Re: VMS clusters prove they are the best - Sun comes in last!0 Message-ID: <br49cg$j7t$3@new-usenet.uk.sun.com>   Bob Ceculski wrote:h > Andrew Harrison SUNUK Consultancy <Andrew_No.Harrison_No@nospamn.sun.com> wrote in message news:<br1l6q$k0m$1@new-usenet.uk.sun.com>...c >  >>Bob Ceculski wrote:n >> >>>Andrew Harrison SUNUK Consultancy <Andrew_No.Harrison_No@nospamn.sun.com> wrote in message news:<bqkvef$aie$1@new-usenet.uk.sun.com>... >>>k >>>v >>> = >>>I am smarter than you Andrew in that I run my platforms onU3 >>>OpenVMS and you don't ... CERT COUNTS DON'T LIE!t >>9 >>CERT counts do lie, you know this and just to reinforceo< >>the point engineers from Compaq have admitted that OpenVMS5 >>CERT advisories have not in the past been accurate.e >>; >>There is also uncontovertable proof that they still arn'te >>accurate.  >>; >>If smart is clinging to a statistical measure of securityh; >>which has been demonstrated to be total tosh then you arel8 >>one very very smart cookie, I congratulate you on your >>smartness. >>	 >>Regardse >>Andrew Harrisonx >  > > > in the real world, I have demonstrated VMS security as being> > valid for over 18 years now ... not one security problem and > not one OS crash ... EVER!    > So why keep refering to CERT counts, concentrate on your "real> world experience". CERT counts at least for OpenVMS are as far, from real world as you can possibly imagine.  ; BTW in the "real world" how is you OpenVMS firewall withoutp a firewall coming on.o   Regardsi Andrew Harrison    ------------------------------   End of INFO-VAX 2003.681 ************************