1 INFO-VAX	Thu, 27 Jul 2006	Volume 2006 : Issue 415       Contents:/ Re: "get" function in HP's sftp fails sometimes  Re: Alpha remembrance day 8 Re: Alphaserver 2100 needs a good home - South Cambs, UK Re: An opportunity for VMS Re: An opportunity for VMS Re: An opportunity for VMS$ Announcing Availability Manager V2.6@ Re: Any Way to Validate Username & Password from an Application?@ Re: Any Way to Validate Username & Password from an Application? Cards supported by XP1000  Re: Cards supported by XP1000  Re: detect user in winxp4 Re: Digital^H^H^H^H^H^H^H hp is a "Software" CompanyP Re: Mercury Interactive (was: Re: Digital^H^H^H^H^H^H^H hp is a "Software" Compa9 Re: PCSI uninstalling products whose location has changed % Re: PIPE redirection as stream file ? % Re: PIPE redirection as stream file ? % Re: PIPE redirection as stream file ? % Re: PIPE redirection as stream file ?  Re: PWS500au manual needed( Re: Tomcat user authentication question.( Re: Tomcat user authentication question. Upgrading disks in Shadowset  Re: Upgrading disks in Shadowset  Re: Upgrading disks in Shadowset  Re: Upgrading disks in Shadowset  Re: Upgrading disks in Shadowset& Re: Using DSF and DMP files with debug& Re: Using DSF and DMP files with debug& Re: Using DSF and DMP files with debug& Re: Using DSF and DMP files with debugE Re: VMS and HPVM (was: Parsec webinar (2006-07-12) OpenVMS Licensing)   F ----------------------------------------------------------------------    Date: 27 Jul 2006 08:23:18 -0700* From: "Palda" <jiri.pallich@logicacmg.com>8 Subject: Re: "get" function in HP's sftp fails sometimesC Message-ID: <1154013798.506373.162320@b28g2000cwb.googlegroups.com>    Hi,   @ I have got the solution already. The error was caused by "ascii"G transfer mode which does not work for HP's sftp client. The binary mode  works OK :)    Regards, Jiri    
 Palda napsal:  > Greetings, > I > I have a problem with HP's sftp client. Sometimes, the "get" and "mget" H > function fails and makes the sftp crashing. Especially when using sftpG > in batch mode (sftp "-B" ...) it happens everytime. Does anybody have H > any idea what is going wrong in here? Thank you for your tips'n'tricks > :) >  > Best regards > Jiri Pallich >  > Here is the output:  > -------------------------- >  > VMS1:SYSTEM> sftp user@node 
 > sftp> ascii ! > File transfer mode is now ascii  > sftp> cd PRIVATE/XYZ/UPLOAD  > /PRIVATE/XYZ/UPLOAD  > sftp> verbose 
 > sftp> ls -l E > SshFCRecurse/SSHFC_RECURSE.C:280: File is "raw", and it needs to be 	 > parsed. E > -rw-r--r--    1 667      503           951 Nov 10  2005 id_rsa_xxxx 9 > -rwxr-xr--    1 667      503           369 Nov 14  2005  > id_rsa_1024_xxxx.pubF > -rwxr-xr-x    1 501      501           139 Jul 24 11:42 ABC00006.LOG= > drwxr-xr-x    2 667      503          4096 Jul 24 13:59 OP1 = > drwxr-xr-x    2 667      503          4096 Jul 21 13:42 OP2 E > SshFCRecurse/SSHFC_RECURSE.C:215: Received error `End of file' (1).  > sftp> get ABC00006.LOGE > SshFCRecurse/SSHFC_RECURSE.C:280: File is "raw", and it needs to be 	 > parsed. = > SshFCTransfer/SSHFC_TRANSFER.C:2051: File list has 2 files. < > SshReadLine/SSHREADLINE.C:3717: Uninitializing ReadLine...E > FATAL: BUILD4$:[TCPIP_V54_BL15.SRC_PAT.SSH2]SSHFC_TRANSFER.C;3:2097 3 > SshFCTransfer (function name unavailable) Asserti 2 > on failed: attrs->source_newline != ((void *) 0) > 8 > %TCPIP-F-SSH_FATAL, non-specific fatal error condition > VMS1:SYSTEM> sftp user@node  > sftp> verbose  > sftp> cd PRIVATE/XYZ/UPLOAD  > /PRIVATE/XYZ/UPLOAD 
 > sftp> ls -l E > SshFCRecurse/SSHFC_RECURSE.C:280: File is "raw", and it needs to be 	 > parsed. F > -rw-r--r--    1 667      503           951 Nov 10  2005 id_rsa_nldlc9 > -rwxr-xr--    1 667      503           369 Nov 14  2005  > id_rsa_1024_nldlc.pub F > -rwxr-xr-x    1 501      501           139 Jul 24 11:42 ABC00006.LOG= > drwxr-xr-x    2 667      503          4096 Jul 24 13:59 OP1 = > drwxr-xr-x    2 667      503          4096 Jul 21 13:42 OP2 E > SshFCRecurse/SSHFC_RECURSE.C:215: Received error `End of file' (1).  > sftp> mget ABC00006.LOG E > SshFCRecurse/SSHFC_RECURSE.C:280: File is "raw", and it needs to be 	 > parsed. = > SshFCTransfer/SSHFC_TRANSFER.C:2051: File list has 2 files. 9 > SshFCTransfer/SSHFC_TRANSFER.C:462: Next source file is $ > PRIVATE/NLLC/UPLOAD/ABC00006.LOG .B > SshFCTransferCore/SSHFC_TRCORE.C:356: Starting transfer for file* > ABC00006.LOG, destination ./ABC00006.LOGI > ABC00006.LOG                    |   139B |   0.1 kB/s | TOC: 00:00:01 |  > 100%8 > SshFCTransferCore/SSHFC_TRCORE.C:814: Writer finished.H > SshFCTransfer/SSHFC_TRANSFER.C:1575: Destination file is not a regular  > file, not changing attributes.I > SshFCTransfer/SSHFC_TRANSFER.C:1641: Finished with file ./ABC00006.LOG.  > sftp> get ABC00006.LOGE > SshFCRecurse/SSHFC_RECURSE.C:280: File is "raw", and it needs to be 	 > parsed. = > SshFCTransfer/SSHFC_TRANSFER.C:2051: File list has 2 files. 9 > SshFCTransfer/SSHFC_TRANSFER.C:462: Next source file is $ > PRIVATE/NLLC/UPLOAD/ABC00006.LOG .B > SshFCTransferCore/SSHFC_TRCORE.C:356: Starting transfer for file* > ABC00006.LOG, destination ./ABC00006.LOGI > ABC00006.LOG                    |   139B |   0.1 kB/s | TOC: 00:00:01 |  > 100%8 > SshFCTransferCore/SSHFC_TRCORE.C:814: Writer finished.H > SshFCTransfer/SSHFC_TRANSFER.C:1575: Destination file is not a regular  > file, not changing attributes.I > SshFCTransfer/SSHFC_TRANSFER.C:1641: Finished with file ./ABC00006.LOG.  > sftp> exit > ssh_pipe_stream_destroy < > SshReadLine/SSHREADLINE.C:3717: Uninitializing ReadLine... >  > VMS1:SYSTEM> ucx sh ver  > = >   HP TCP/IP Services for OpenVMS Alpha Version V5.4 - ECO 5 5 >   on a AlphaServer 800 5/500 running OpenVMS V7.3-2    ------------------------------    Date: 27 Jul 2006 04:20:56 -0700- From: "Andrew" <andrew_harrison@symantec.com> " Subject: Re: Alpha remembrance dayB Message-ID: <1153999256.303693.184110@i3g2000cwc.googlegroups.com>   Dave Froble wrote: > Andrew wrote:  > > Dave Froble wrote: > >> Andrew wrote: > >>> Dave Froble wrote: > >>>> Andrew wrote: > >>>>I > >>>>> So who would Digital have sold FAB capacity to? They had no track N > >>>>> record of working with FABless vendors like Sun and they were in a lotL > >>>>> of instances major competitors for the companies who might have been- > >>>>> target customers to use Digitals FAB. ! > >>>> AMD wanted to use the FAB. K > >>> AMD wanted to use the FAB but didn't. Now unless you know why and you I > >>> don't the options are they couldn't do a deal with Digital, Digital F > >>> wouldn't do a deal with them, they got another better offer etc. > >>> L > >>> But you don't know the answer so describing my post as bullshit FUD is  > >>> painfull but only for you.H > >> You're truly a fucking idiot!  The topic, which your attention spanE > >> couldn't embrace, was that others wanted to use the FAB, but DEC M > >> wouldn't take the business, wanting to keep the FAB ready for more Alpha . > >> business.  Do you possibly remember that? > > H > > Of course I can remember, but as I pointed out and you very stupidlyI > > ignored it is only supposition that DEC turned this business down and J > > even if they did there could be numerous good technical and commercial > > reasons why they did so. > > M > >> So the answer is, DEC wouldn't sell FAB capability to AMD.  Sorry if you H > >> don't consider it possible for me to know why.  It's only the topic > >> under discussion. > >> > > . > > Sadly again only supposition on your part. > >  > >> Idiot!  > >  > > You are getting fainter  > > M > >>>> As for the rest of your bullshit FUD, Alpha didn't need to be accepted R > >>>> by any other vendors.  The VMS and Tru64 business was supporting it nicely. > >>>>J > >>> So why did Digital go out of their way to try to attract design winsJ > >>> onto Alpha. Why did they tout the AlphaPC surely even Digital didn'tM > >>> think that they would clean up and become the only manufacturer of PC's  > >>> worldwide.L > >> Because some people saw it as an opportunity.  Over and above DEC's useI > >> of the CPU.  Not a bad idea, but it not working in no way made Alpha  > >> unsuccessful. > >>I > > Akkk either they did or they didn't. You claim that DEC sabotaged the J > > sucess of the Hudson FAB by turning away FAB business from AMD but nowD > > you claim that they didn't really need this fabrication business > > anyway.  > % > Show me where I claimed 'sabotage'. ? > Show me where I claimed DEC didn't need fabrication business.  > ( I will quote what you said about the FAB  6 "The VMS and Tru64 business was supporting it nicely."  G So if it was supporting it nicely why did DEC need to sell FAB capacity D to AMD and why would it make any difference to you or anyone else if DEC aledgedly blew AMD out ?  B You see the thread you so unwisely got involved in started with anB allegation that Palmer deliberately sabotaged Alpha by not lettingG Hudson be used to FAB third party parts. That is the thread I responded C to and its my response that you attempted to flame. And now you are F apparently claiming that the FAB useage issue is a red-herring because* Tru64 and OpenVMS revenues had it covered.  D So why didn't you say so earlier in response to the origional post ?  G > All I wrote was that AMD wanted to use the FAB, but DEC didn't accept H > the business.  Nothing more.  That was in response to your question of6 > who might have wanted to use the FAB.  Nothing more. >   E As I pointed out earlier your claim that DEC turned the business down B is just supposition on your part and as I also said if they did itG could have been for perfectly sensible technical or commercial reasons.     F > > I hate to say this but each time you accuse me of being stupid youI > > appear to be doing so from a deeper and deeper hole. Soon we won't be  > > able to hear you at all. > I > You are stupid when you try to twist what someone writes into something ? > else.  if you want to tell stories, take the credit yourself.  >   $ Really so where have I told a story?   > > L > >>> The reality is that the whole Alpha plan was a clone of the MIPS/SPARCG > >>> plans, PC's, Workstations, Small, Medium and Large Servers, Super L > >>> Computers lots of vendors in other words a life changing event for the > >>> computer market. > >>> L > >>> And no the Tru64 and OpenVMS businesses were not supporting it nicely,F > >>> at $400-$500 million dollars per new chip design and $2+ billionL > >>> dollars for the FAB plus ongoing running costs itis doubtfull that theG > >>> Tru64 and OpenVMS even supported the development and build of the I > >>> processors let alone support the rest of the ecosystem required for B > >>> Alpha, software, interconnects and platforms, marketing, ISV > >>> development etc. > >>> N > >>>> Note that IBM doesn't have a bunch of vendors using Power, but they areM > >>>> doing well.  The FAB may not be profitable, but a vendor that can do a M > >>>> complete solution has the flexibility that allows for overall profits.  > >>>>F > >>> Ahh the much missused IBM example and not even accurate at that.2 > >> Just because you write it doesn't make it so. > >> > > K > > Well since I said it and then backed it up with facts the ball is sitll  > > in you court.  > > 	 > >>> IBM K > >>> was in fact quite sucessfull at attracting other vendors to use POWER  > >>> or POWER variants.K > >> Variants may be different enough to be considered a different produce.  > >> > > K > > Does it matter, it generates revenue for the POWER business unit and it J > > uses FAB capacity. Exactly what you aledge DEC deliberately refused to > > do.  > I > Agreed.  If DEC had recovered some of the cost by selling FAB services, J > they would have been better off.  How much?  I cannot say.  Enough to be > worthwhile?  I cannot say. >    But as you said   7 "The VMS and Tru64 business was supporting it nicely. "   F So if as you claim Tru64 and OpenVMS were capable of generating enoughG revenues to support the Hudson FAB then what would be the incentive for  DEC?  G I can see that the revenue from FABing AMD parts might have made Hudson G more profitable but not necessarily and the downside would be that they B would not have had as much capacity available had Alpha taken off.  3 > >>> And while Apple has deserted for Intel IBM is J > >>> happily building the POWER based processor for the xbox 360 in their' > >>> Fishkill FAB. Heard of the xbox ? H > >> Apple wasn't enough for IBM to build the types of CPU Apple wanted. > >> Notebooks in particular.  > >>K > >> Yes, I've heard of the xbox, and the cell processor.  I doubt they are 3 > >> interchangable with the Power5+ based systems.  > > E > > Can you show why that would matter to IBM's POWER business unit ? L > >>> But the IBM example would have been a bad one anyway, IBM are and wereM > >>> a much bigger company than Digital, appart from design wins outside IBM C > >>> POWER is a shared platform for AS400/RS6000 (p and i Series). L > >> The IBM example is exactly what Alpha was within DEC.  A differentiator# > >> to allow them to win business.  > > H > > I am at a loss to understand why AS400 using POWER and RS/6000 usingK > > POWER is a differentaitor that allows IBM to win business and I bet you  > > can't explain why either.  > + > Maybe they can show superior performance?  >   F Really and how would AS400 and RS/6000 using POWER help deliver better
 performance ?    Regards  Andrew Harrison    ------------------------------    Date: 27 Jul 2006 03:18:26 -0700 From: etmsreec@yahoo.co.ukA Subject: Re: Alphaserver 2100 needs a good home - South Cambs, UK B Message-ID: <1153995506.159762.176910@i3g2000cwc.googlegroups.com>  ! What do you have to trade Philip?    Steve   / Phillip Helbig---remove CLOTHES to reply wrote: ; > In article <pan.2006.06.29.21.08.59.203505@decadence.it>, - > "gl@decadence.it" <gl@decadence.it> writes:  > = > > On Wed, 14 Jun 2006 05:47:50 +0000, Phillip Helbig wrote: M > > > I also have such a beast looking for a good home.  It's in Germany, but L > > > I will be travelling quite a bit in Europe this summer and could bringE > > > it to most places.  I also have a 2000 looking for a good home.  > >  > > Hello :) > > L > > I have to give public thanks to Phillip who bought to me a perfect shape > > AlphaServer 2100 in Italy.J > > Now the server is sleeping in my lab, waiting to be put at work again. > > That day will be very soon!  > >  > > Thanks Phillip! :) > > See you again! > F > Glad you are enjoying it!  The heat was a disadvantage while drivingH > down (with no air-conditioning), but an advantage during our followingJ > two weeks of holiday across the water in Croatia.  After another week ofH > holiday I've been back for a week and a half now, and am just catchingF > up on email and newsgroups, so apologies to anyone else who has beenG > corresponding with me about hardware and is still awaiting an answer.  > H > While I'm thinking about it, I'll be in England 7--13 August and couldG > collect and/or distribute some hardware.  Hopefully this weekend I'll I > post a list of the few things I want to get rid of.  Below is a list of E > the few things I still need.  Belgium or the Netherlands or Germany C > between Frankfurt am Main and Belgium is also possible or, 16--26 $ > August, Sweden, Denmark or Norway. >  > What I still need: > ) >    o  BA353 ("pizza-box") expansion box  >  >    o  DLT tapes  > , >    o  small ALPHA machines, EV45 or better > ' >    o  SBB  disk drives 1 GB or larger    ------------------------------    Date: 27 Jul 2006 05:45:18 -0700$ From: "AEF" <spamsink2001@yahoo.com># Subject: Re: An opportunity for VMS B Message-ID: <1154004318.393996.30300@m79g2000cwm.googlegroups.com>   david20@alpha2.mdx.ac.uk wrote: k > In article <1153961148.067093.220350@i3g2000cwc.googlegroups.com>, "AEF" <spamsink2001@yahoo.com> writes:  > >  > >Dave Froble wrote:  > >> AEF wrote:  > >> > Dave Froble wrote:  > >> >> AEF wrote: > >> >> K > >> > Then going to Mars? Going to Mars is fraught with huge difficulties, J > >> > not the least of which is the danger of radiation to the crew. This9 > >> > would be incredibly expensive and difficult to do.  > >>& > >> Yeah.  You state facts.  So what. > > G > >Another unpleasant fact is that the distance involved to the moon is J > >about 250,000 miles. Mars at its very closest is 35 million miles away.H > >A reasonable trajectory using a minimum amount of fuel would probablyE > >involve a trip of something like (very rough estimate) 250,000,000 F > >miles! That's a factor of 1000. And maybe its an underestimate. AndG > >launching from Mars would be much harder than from the Moon. How are H > >you going to get enough fuel there to do that? Proposals say that theF > >astronauts should find fuel on Mars itself. Well, I wouldn't bet myI > >life on something like that quite yet. One op-ed I read said to send a E > >few individuals on a one-way trip to Mars. It said that some would 4 > >gladly volunteer even knowing they'd die on Mars. > > P > Mars Direct style missions in which hydrogen is taken to Mars and reacted withO > the martian atmosphere to produce water and methane and the carbon dioxide in E > the atmosphere is broken apart to produce oxygen isn't particularly  > complicated.M > (I was going to say it isn't exactly rocket science - but of course in this  > instance it is :) )  > From figures on L > http://www.geocities.com/marsterraforming/mannedmissions.html with a 6 tonN > hydrogen input from earth you get back 108 tons of propellant after about 10K > months. The estimated cost is about $20-30 billion dollars which compares 6 > reasonably well with the Apollo cost of $20 billion.M > The other advantage of Mars Direct style missions is that infrastructure is N > left behind on Mars to help with future missions and to gradually build into& > what could be a permanent Mars base. >  >  >  > David Webb > Security team leader > CCSS > Middlesex University     Interesting, but I'm skeptical.   < I looked at the reasons to go to Mars. One really stood out:  @ The first manned landing on Mars would serve as an invitation toG adventure for children around the world. There will be some 100 million G kids in the U.S. schools over the next 10 years. If a Mars program were D to inspire just an additional 1 percent of them to pursue scientific@ educations, the net result would be one million more scientists,6 engineers, inventors, medical researchers and doctors.  G So what good are a million unempolyed scientists, engineers, inventors,   medical researchers and doctors?  > OTOH, I'm all for improving the scientific literacy of people.  F If it *really* can be done for $20-30 billion, then maybe we should doC it. But note that it would still take 10 years! I did said it would 
 take a while.    AEF    ------------------------------    Date: 27 Jul 2006 06:03:52 -0700$ From: "AEF" <spamsink2001@yahoo.com># Subject: Re: An opportunity for VMS C Message-ID: <1154005432.383414.138080@m79g2000cwm.googlegroups.com>    david20@alpha2.mdx.ac.uk wrote: k > In article <1153964917.502115.210240@i3g2000cwc.googlegroups.com>, "AEF" <spamsink2001@yahoo.com> writes:  > > " > >david20@alpha2.mdx.ac.uk wrote:o > >> In article <1153917332.057547.201460@i42g2000cwa.googlegroups.com>, "AEF" <spamsink2001@yahoo.com> writes:  > >> > > >> >Dave Froble wrote: > >> >> AEF wrote: > >> >>  > >> > > >>O > >> Premature ? The last 25 years have been pretty well wasted as far as human $ > >> space exploration is concerned. > > E > >Well, yes and no. We have learned a lot about putting people up in H > >space for extended periods. One thing we learned is that zero-gravityH > >can have negative effects on the musculoskeltal system. For a trip toI > >Mars we'd probably have to construct a centrifuge to create artificial C > >gravity. This would cause a significant increase in the payload!  > >  >  > I > >I think trying put humans on Mars now is like trying to go to the Moon H > >in the 1920's. We just don't have the technology yet (well, I have toJ > >admit I haven't read up on this topic, but I think at the very least itJ > >would be very difficult and certainly very, very expensive! Look at howD > >the space program has gone so far: 3 lives lost going to the MoonH > >(Apollo 1). 3 more lives very nearly lost on the same mission (ApolloI > >13). Fourteen lost on the Shuttle to date! I believe the Russians lost D > >at least one cosmonaut, but they didn't go to the moon or build a > >shuttle.  > > E > I'm sure if you had most people a month before Sputnik was launched O > (apart from the Russians involved) about trying to get to the Moon they would # > have expressed a similar opinion.   * OK, but sometimes the naysayers are right.  F I guess my main point is that the difference between going to the Moon and to Mars is HUGE.  E > A Mars Direct style mission should be well within our capabilities.   = I'm not convinced the psychological problems would be modest.   5 > How many early airplane pioneers lost their lives ?   E I don't know but that's not the point. I'm not talking about aversion G to risk. If the astronauts don't survive the trip, the mission would be 9 a failure. You're advocating a successful mission, right?   J > For that matter how many people have lost their lives flying on "proven" > commercial aircraft ?   * A much smaller percentage than astronauts.  N > By their very nature rockets are always going to be at least as dangerous as > commercial aircraft.M > Mars direct does however have one advantage over the Apollo style missions. M > It is possible to build up substantial resources on Mars before anyone goes N > meaning that a crew could survive being marooned on Mars for long enough forQ > a rescue mission to be arranged (indeed it is hard to envisage why anyone would K > plan a Mars mission which stayed on Mars for days rather than staying for  > months or even years).  9 There's still plenty of room for things to go very wrong.    >  >  >  > David Webb > Security team leader > CCSS > Middlesex University [...]    AEF    ------------------------------    Date: 27 Jul 2006 06:40:17 -0700$ From: "AEF" <spamsink2001@yahoo.com># Subject: Re: An opportunity for VMS B Message-ID: <1154007617.322093.48720@s13g2000cwa.googlegroups.com>   david20@alpha2.mdx.ac.uk wrote: k > In article <1154004318.393996.30300@m79g2000cwm.googlegroups.com>, "AEF" <spamsink2001@yahoo.com> writes:  > > " > >david20@alpha2.mdx.ac.uk wrote:n > >> In article <1153961148.067093.220350@i3g2000cwc.googlegroups.com>, "AEF" <spamsink2001@yahoo.com> writes: > >> > > >> >Dave Froble wrote: > >> >> AEF wrote: > >> >> > Dave Froble wrote: > >> >> >> AEF wrote: 
 > >> >> >>N > >> >> > Then going to Mars? Going to Mars is fraught with huge difficulties,M > >> >> > not the least of which is the danger of radiation to the crew. This < > >> >> > would be incredibly expensive and difficult to do. > >> >> ) > >> >> Yeah.  You state facts.  So what.  > >> >J > >> >Another unpleasant fact is that the distance involved to the moon isM > >> >about 250,000 miles. Mars at its very closest is 35 million miles away. K > >> >A reasonable trajectory using a minimum amount of fuel would probably H > >> >involve a trip of something like (very rough estimate) 250,000,000I > >> >miles! That's a factor of 1000. And maybe its an underestimate. And J > >> >launching from Mars would be much harder than from the Moon. How areK > >> >you going to get enough fuel there to do that? Proposals say that the I > >> >astronauts should find fuel on Mars itself. Well, I wouldn't bet my L > >> >life on something like that quite yet. One op-ed I read said to send aH > >> >few individuals on a one-way trip to Mars. It said that some would7 > >> >gladly volunteer even knowing they'd die on Mars.  > >> >S > >> Mars Direct style missions in which hydrogen is taken to Mars and reacted with R > >> the martian atmosphere to produce water and methane and the carbon dioxide inH > >> the atmosphere is broken apart to produce oxygen isn't particularly > >> complicated. P > >> (I was going to say it isn't exactly rocket science - but of course in this > >> instance it is :) ) > >> From figures onO > >> http://www.geocities.com/marsterraforming/mannedmissions.html with a 6 ton Q > >> hydrogen input from earth you get back 108 tons of propellant after about 10 N > >> months. The estimated cost is about $20-30 billion dollars which compares9 > >> reasonably well with the Apollo cost of $20 billion. P > >> The other advantage of Mars Direct style missions is that infrastructure isQ > >> left behind on Mars to help with future missions and to gradually build into ) > >> what could be a permanent Mars base.  > >> > >> > >> > >> David Webb  > >> Security team leader 	 > >> CCSS  > >> Middlesex University  > >  > > " > >Interesting, but I'm skeptical. > > ? > >I looked at the reasons to go to Mars. One really stood out:  > > C > >The first manned landing on Mars would serve as an invitation to J > >adventure for children around the world. There will be some 100 millionJ > >kids in the U.S. schools over the next 10 years. If a Mars program wereG > >to inspire just an additional 1 percent of them to pursue scientific C > >educations, the net result would be one million more scientists, 9 > >engineers, inventors, medical researchers and doctors.  > > J > >So what good are a million unempolyed scientists, engineers, inventors,# > >medical researchers and doctors?  > > A > >OTOH, I'm all for improving the scientific literacy of people.  > > I > >If it *really* can be done for $20-30 billion, then maybe we should do F > >it. But note that it would still take 10 years! I did said it would > >take a while. > > J > Kennedy set the goal of going to the moon "before this decade is out" on6 > 25th May 1961 and it was achieved on 20th July 1969.  5 Kennedy didn't say anything about going to Mars. (:-)   G > Hence a 10 year timescale for going to Mars is reasonably consistent.   G For 1000 times the distance?! Twice the gravity?! Much more radiation?!    > M > A Mars base is probably one of the major steps necessary for opening up the & > resources of the whole solar system.  9 What resources in particular are you thinking about here?       >---o---<  # Have you seen 2001 a Space Odyssey?    >  > David Webb > Security team leader > CCSS > Middlesex University >  > >AEF > >    ------------------------------  # Date: Thu, 27 Jul 2006 15:34:20 GMT 0 From: "Barry Kierstein" <Barry.Kierstein@HP.Com>- Subject: Announcing Availability Manager V2.6 1 Message-ID: <025yg.1188$eU2.794@news.cpqcorp.net>   E The Availability Manager team is pleased to announce the Availability I Manager Version 2.6.  This version includes updates to both Data Analyzer  and Data Collector kits.  K New features in the Availability Manager Version 2.6 include the following:   :     -  Support for OpenVMS V8.3, including 64 CPU support.  :     -  System Overview window redone to look like DECamds.  I     -  Additional Group Overview window and Single Group window for sites         with many groups.  @     -  New Single Process views to view all information at once.  D     -  New events added to detect high CPU usage on the primary CPU.    C Problems corrected in Availability Manager Version 2.6 include the  
 following:  @     -  Disk Status and Disk Volume filtering problems corrected.  I     -  Disk free space collection corrected, occasionally reported stale   values.   E     -  Single process windows fixes - fixed hangs in data collection.     J The Availability Manager kits can be obtained from the following web site:  6        www.hp.com/products/openvms/availabilitymanager    I Thank you for using our products.  Let us know if you have any questions.    The Availability Manager Team  OpenVMS Engineering    ------------------------------  # Date: Thu, 27 Jul 2006 10:15:54 GMT " From:   VAXman-  @SendSpamHere.ORGI Subject: Re: Any Way to Validate Username & Password from an Application? 0 Message-ID: <00A594CC.945436C9@SendSpamHere.ORG>  p In article <44c824a5.524144494f47414741@radiogaga.harz.de>, martin@radiogaga.harz.de (Martin Vorlaender) writes: >  >  >Phillip Helbig wrote:. >> "Craig Dedo" <Dedo.Craig@mayo.edu> writes:  >>  
 >>> Everyone: D >>>     Is there any way to validate a username and password from anJ >>> application?  We need to have this functionality in one of our OpenVMS >>> applications.  >>  K >> The source code for the OSU HTTP server uses system services to do this. % >> You could take that as an example.  > " >I have a C source code snippet atF >http://www.pdv-systeme.de/users/martinv/VMS_Programming_FAQ.html#3.8. >  >cu,	 >  Martin   K They're not going to like that Martin, you used $GETUAI and $HASS_PASSWORD.    --  K VAXman- A Bored Certified VMS Kernel Mode Hacker   VAXman(at)TMESIS(dot)COM              5   "Well my son, life is like a beanstalk, isn't it?"     ------------------------------  # Date: Thu, 27 Jul 2006 14:05:25 GMT , From: Hoff Hoffman <hoff-remove-this@hp.com>I Subject: Re: Any Way to Validate Username & Password from an Application? 1 Message-ID: <FK3yg.1180$NR2.422@news.cpqcorp.net>   H There's a tool over at Hunter Goatly's pages at process.com that claims D to include breakin evasion and other operations as part of password H verification.  Off the top, I don't remember the name (and a quick look E doesn't indicate which of various of the SYSUAF/PASSWORD/LOGIN tools  F might be the one that I recall), but I do recall seeing references to I the NSA calls and the password dictionary in the detailed description of  % the package as I was uploading stuff.   "    http://www.process.com/openvms/  F Or you can call sys$acm, which should "get it right" for both current  and for new schemes.  I I don't think we document the whole evasion process anywhere (like we do  G document the privilege access checks), but that material might well be  B fodder for a reasonable discussion, as application programmers do H regularly implement this stuff (and for any of various reasons) without ; benefit of or without access to mechanisms such as sys$acm.    ------------------------------   Date: 27 Jul 2006 15:29:18 GMT From: healyzh@aracnet.com " Subject: Cards supported by XP1000, Message-ID: <eaam4e022d9@enews1.newsguy.com>  3 I currently have the following in my "new" XP1000.     KZPSA (Wide Diff SCSI) PowerStorm 300! KZPCA (U2/LVD SCSI, 53C895 based)   H I need Narrow SCSI for the TLZ06 I use to transfer data to my PDP-11.  IH have the following additional SCSI Cards.  I've heard there is a problemK with the KZPAA in an XP1000, and I see David's site lists it as not working L with 53C895 controllers.  As a result I'm guessing the KZPAA is out, but canB I use the KZPBA-CA for Narrow SCSI (at one point I was in my PWS). KZPAA (Narrow SCSI)  KZPBA-CA (SE UW-SCSI QLA1040B)K I suppose I could also try dropping the onboard QLA1040B controller down to  Narrow SCSI.  G If I can find it, I'll replace the KZPSA with a KZPBA-CB (Diff UW-SCSI  H QLZ1041B).  In the long run I hope to add a GigE card (don't have that).  D The SCSI-Diff adapter will be used to drive a single DLT tape drive.   		Zane   ------------------------------  % Date: Thu, 27 Jul 2006 08:53:16 -0700 # From: "Tom Linden" <tom@kednos.com> & Subject: Re: Cards supported by XP1000) Message-ID: <op.tdcue2fxzgicya@hyrrokkin>   @ On Thu, 27 Jul 2006 08:29:18 -0700, <healyzh@aracnet.com> wrote:  4 > I currently have the following in my "new" XP1000. >  > KZPSA (Wide Diff SCSI) > PowerStorm 300# > KZPCA (U2/LVD SCSI, 53C895 based)  > I > I need Narrow SCSI for the TLZ06 I use to transfer data to my PDP-11. =   II > have the following additional SCSI Cards.  I've heard there is a probl=  emH > with the KZPAA in an XP1000, and I see David's site lists it as not  =  	 > working I > with 53C895 controllers.  As a result I'm guessing the KZPAA is out, b=  ut  =    > can D > I use the KZPBA-CA for Narrow SCSI (at one point I was in my PWS).  6 Yes, just get a 68 -> 50 pin adapter ( < $10 at Fry's)   > KZPAA (Narrow SCSI)   > KZPBA-CA (SE UW-SCSI QLA1040B)I > I suppose I could also try dropping the onboard QLA1040B controller do=  wn  =    > to > Narrow SCSI. > I > If I can find it, I'll replace the KZPSA with a KZPBA-CB (Diff UW-SCSI=   I > QLZ1041B).  In the long run I hope to add a GigE card (don't have that=  ).  7 I have 3 DEGPA-SA  that I would be willing to part with  > F > The SCSI-Diff adapter will be used to drive a single DLT tape drive. >  > 		Zane >    ------------------------------    Date: 27 Jul 2006 04:30:15 -0700 From: dooleys@snowy.net.au! Subject: Re: detect user in winxp C Message-ID: <1153999815.831055.112120@m73g2000cwd.googlegroups.com>    lashnjo wrote:/ > i'm sorry i posted in the wrong group, my bad 
 > kill me -.-  > lashnjo wrote:G > > how can i detect the current user in winxp, is there a dos command? $ USERNAME is an Enivironment variable, use %USERNAME% or  "set username" to show it Phil   ------------------------------    Date: 27 Jul 2006 07:20:26 -0700- From: "Andrew" <andrew_harrison@symantec.com> = Subject: Re: Digital^H^H^H^H^H^H^H hp is a "Software" Company C Message-ID: <1154010026.228839.183990@p79g2000cwp.googlegroups.com>    Robert Deininger wrote: E > In article <4ipl4kF4ttq2U1@individual.net>, bill@cs.uofs.edu wrote:  > 6 > >In article <tExJPKUSE8YC@eisner.encompasserve.org>,: > >        Kilgallen@SpamCop.net (Larry Kilgallen) writes:F > >> In article <ea7hgd$o5e$1@news-02.connect.com.au>, "Richard Maher"' > <maher_rj@hotspamnotmail.com> writes: 	 > >>> Hi,  > >>> @ > >>> So what if anything does it mean for VMS and hp direction? > >>> P > >>> A quick flick suggests Mercury is better known for fiddling the books withI > >>> dodgy share options than software - As much sense as buying Compaq?  > >>> 3 > >>> http://uk.biz.yahoo.com/060726/323/ghtqm.html  > >>>  > >>> Regards Richard Maher  > >>I > >> What Richard was unwilling to summarize and thus insisted on wasting  > >> our lookup time with is:  > >>H > >>       Hewlett Packard is planning on purchasing Mercury Interactive# > >>       for 4.5 billion dollars.  > > 6 > >Which, of course, still didn't answer the question:? > >   "what if anything does it mean for VMS and hp direction?"  > > K > >I didn't see any mention of VMS on their website, so is this yet another 4 > >HP product intended to lead people away from VMS? >  > J > Absolutely.  HP is spending $4.5 Billion for the sole purpose of getting > customers away from VMS.   Hardly  = Mercury provide systems managment, performance management and % measurement products. IT finance etc.    LoadRunner, WebRunner etc.  # http://www.mercury.com/uk/products/   F The financial and business products listed in the page are not SAP etc> equivalents they are for manageing IT costs charging back etc.  G I should know Symantec compete with Mercury, we certainly don't compete  against SAP or Oracle.   Regards  Andrew   ------------------------------  # Date: Thu, 27 Jul 2006 13:53:00 GMT , From: Hoff Hoffman <hoff-remove-this@hp.com>Y Subject: Re: Mercury Interactive (was: Re: Digital^H^H^H^H^H^H^H hp is a "Software" Compa 1 Message-ID: <0z3yg.1178$KR2.676@news.cpqcorp.net>   D    The Mercury Interactive acquisition has about as much to do with I OpenVMS as your average new printer cartridge -- other than its entirely  I unintended but and entirely unsurprising consequence of providing fodder  & for here in the newsgroups, of course.  ?    What follows is a direct quote from the HP press release at:   >    <http://www.hp.com/hpinfo/newsroom/press/2006/060725a.html>  I    "The transaction brings together the strength of HP OpenView systems,  F network and IT service management software with Mercury's strength in @ application management, application delivery, IT governance and D service-oriented architecture governance. This combination provides @ customers with the industry's most robust suite for optimizing, 9 automating and aligning IT services with business needs."   E    As for the implications for OpenVMS, OpenVMS already has OpenView  I clients for various operations of interest to OpenView users.  I know of  E no current clients for network application management within OpenVMS  H that are being maintained within OpenVMS Engineering; AFAIK, we have no D tie-ins with the existing Mercury Interactive software environment, ) unless there's some add-on client around.    	--   G     Quantum Physics does theorizes that the flapping of the wings of a  I butterfly in the Amazon will eventually provide fodder for postings here  ( in the comp.os.vms newsgroup, after all.   ------------------------------  # Date: Thu, 27 Jul 2006 13:35:59 GMT , From: Hoff Hoffman <hoff-remove-this@hp.com>B Subject: Re: PCSI uninstalling products whose location has changed1 Message-ID: <3j3yg.1177$2R2.187@news.cpqcorp.net>   , Alan Winston - SSRL Central Computing wrote:"   for), and everything seems cool. > L > But PRODUCT REMOVE insists on removing from DKA2:[WEB], which doesn't even > exist any more.    > E > I don't seem to be able to specify a /DESTINATION for uninstalling.   E    HELP PRODUCT RECONFIGURE and (particularly for this case) PRODUCT  A REGISTER VOLUME.  The latter is how you register volume shuffles.    ------------------------------   Date: 27 Jul 2006 12:56:20 GMT* From: js@cs.tu-berlin.de (Joerg Schilling). Subject: Re: PIPE redirection as stream file ?+ Message-ID: <4irrfkF56efsU1@news.dfncis.de>   2 In article <06072622554083_20246DB2@antinode.org>,+ Steven M. Schweda <sms@antinode.org> wrote:   O >> If VMS ignores standards and implements standard functions incorreclty, then  >> VMS is unusable :-( > G >  I'd say that "unusable" is a little harsh.  (I use it all the time.)   G Well, this is a verz central function in libc. If we cannot rely on it,  many programs will fail.  @ >   Is it a bug?  I'd say that that's a matter of opinion.  It'sB >certainly different from UNIX, but the file system on VMS is alsoG >different from a UNIX file system, and that leads to such differences.   K I don't see how the filesystem should have an influence on the way fwrite()  works.  $ >> The correct call in libschily is: >>    >> cnt = fwrite(buf, 1, len, f); >>  I >> It is the only way that works as unfortunately, the fwrite return code L >> is "nitems", we need to set the "itemsize" parameter to 1 to get a useful
 >> result. > D >   Oh.  I was careless with the byte count.  So, what's wrong with: > & >   cnt = len* fwrite(buf, len, 1, f);   It does not work correctly.   @ write(2) may return less than the expected amount in some cases.B In this case, fwrite(buf, len, 1, f); would return 0 although some bytes have been written.    E >So far as I can see, it does the same thing in this case.  (I assume F >that ferror() will still return a non-zero value if not all the bytesI >were written.)  It might be more fun to complain about the C RTL in VMS, < >but is that more productive than making so simple a change?  D ferror() will not return != 0 in case write(2) returns less than the expected amount.    4 In this case, fwrite(buf, len, 1, f); will return 0.  I >   For the normal reader, the code in question is in the latest cdrtools & >kit, "[.libschily.stdio]filewrite.c": >  >[...]# >        if (my_flag(f) & _IOUNBUF) 5 >                return (write(fileno(f), buf, len)); C >        cnt = fwrite(buf, 1, len, f);                   /* <--- */  >  >        if (!ferror(f)) >                return (cnt);( >        if (!(my_flag(f) & _IONORAISE))" >        raisecond(_writeerr, 0L); >        return (-1);  >[...] > G >   If this would be too awful for non-VMS systems, then I suppose that ' >an "ifdef __VMS" might be needed here.   P An #ifdef would not help either as fwrite(buf, len, 1, f); is wrong in any case.  F >   In any case, I'm glad that you're looking at the proposed changes,E >and I would also be glad to hear any other complaints you might have  >about them.  N Well, I tried to discuss the problem from the current patch set with Eberhard ' Heuser-Hofmann but did not get a reply.   K I am willing to support VMS but it needs to be done in a way that does not   affect other OS.  J I thought that VMS has a POSIX compat layer these days. If it does, there ) should be no problem supporting cdrecord.    --  L EMail:joerg@schily.isdn.cs.tu-berlin.de (home) Jrg Schilling D-13353 Berlin!       js@cs.tu-berlin.de		(uni)   L       schilling@fokus.fraunhofer.de	(work) Blog: http://schily.blogspot.com/M URL:  http://cdrecord.berlios.de/old/private/ ftp://ftp.berlios.de/pub/schily    ------------------------------  # Date: Thu, 27 Jul 2006 13:17:46 GMT , From: Hoff Hoffman <hoff-remove-this@hp.com>. Subject: Re: PIPE redirection as stream file ?1 Message-ID: <_13yg.1175$Jp2.649@news.cpqcorp.net>   
 Pierre wrote:   G > this port http://www.pi-net.dyndns.org/anonymous/kits/ (src are there H > http://www.pi-net.dyndns.org/anonymous/kits/sources/) in (from the top/ > of my head, I'm at home right now...) xmlIO.c   I    libxml2 2.6.24 is posted at the OpenVMS Freeware website (in the V8.0  D staging area), and is the libxml2 port that is shipping out via the G upcoming Freeware V8.0 distro.  (I don't know how far back .20 is, nor  ( if there are any changes relevant here.)  E    I'll take a look at xmlio.c (in 2.6.24) when I next get a chance,  F though it would be useful to know more about the particular sequence. 3 (Can you send me some code to help replicate this?)   I    As for that null call, there is/was a requirement around one of the C  D I/O calls, that a null I/O was required as part of constructing the I output file correctly.  (At one time eons ago, failing to have this null  I I/O even caused RMS to bugcheck; that's been fixed.)  I'll have to check   my notes on this.    ------------------------------    Date: 27 Jul 2006 07:02:38 -0700% From: "Pierre" <pierre.bru@gmail.com> . Subject: Re: PIPE redirection as stream file ?B Message-ID: <1154008958.887701.24530@m73g2000cwd.googlegroups.com>   Hoff Hoffman wrote:  > Pierre wrote:  > I > > this port http://www.pi-net.dyndns.org/anonymous/kits/ (src are there J > > http://www.pi-net.dyndns.org/anonymous/kits/sources/) in (from the top1 > > of my head, I'm at home right now...) xmlIO.c  > J >    libxml2 2.6.24 is posted at the OpenVMS Freeware website (in the V8.0E > staging area), and is the libxml2 port that is shipping out via the H > upcoming Freeware V8.0 distro.  (I don't know how far back .20 is, nor* > if there are any changes relevant here.) > F >    I'll take a look at xmlio.c (in 2.6.24) when I next get a chance,G > though it would be useful to know more about the particular sequence. 5 > (Can you send me some code to help replicate this?)  > J >    As for that null call, there is/was a requirement around one of the CE > I/O calls, that a null I/O was required as part of constructing the J > output file correctly.  (At one time eons ago, failing to have this nullJ > I/O even caused RMS to bugcheck; that's been fixed.)  I'll have to check > my notes on this.    I'll take a look at v2.6.24 D in the meantime, I did this patch as a workaround in xmlFileWrite in xmlIO.c    ---cut here---
 #ifdef VMS     if (len < 1) {         items = 0;     } else {=         items = fwrite(&buffer[0], len, 1, (FILE *) context); 9         if ((items == 0) && (ferror((FILE *) context))) { $             xmlIOErr(0, "fwrite()");             return(-1); 	         }      }  #else 9     items = fwrite(&buffer[0], len, 1, (FILE *) context); 5     if ((items == 0) && (ferror((FILE *) context))) {           xmlIOErr(0, "fwrite()");         return(-1);      }  #endif ---cut here---  D as I also need libxslt for xmlstarlet, will it be in freeware v8.0 ?D BTW, do you think xmlstarlet can be interresting for freeware v8.0 ?   Pierre.    ------------------------------  # Date: Thu, 27 Jul 2006 15:48:59 GMT 7 From: John Malmberg <malmberg@dskwld.zko.hp.compaq.dec> . Subject: Re: PIPE redirection as stream file ?1 Message-ID: <Lf5yg.1190$IT2.466@news.cpqcorp.net>    Joerg Schilling wrote:4 > In article <06072622554083_20246DB2@antinode.org>,- > Steven M. Schweda <sms@antinode.org> wrote:  > O >>>If VMS ignores standards and implements standard functions incorreclty, then  >>>VMS is unusable :-(  H Many of the functions in the OpenVMS C Library have multiple modes.  By H default, for better or for worse, it usually falls back to the behavior 9 of the original implementation of the routine on OpenVMS.   H The X/Open and POSIX "standards" that are now in effect were not always 	 the case.   G As the standards have evolved, in order to be compatible with existing  F source code on OpenVMS that expects the pre-standard behavior, it has % required some compromises to be made.   G >> I'd say that "unusable" is a little harsh.  (I use it all the time.)  >   I > Well, this is a verz central function in libc. If we cannot rely on it,  > many programs will fail.  @ >>  Is it a bug?  I'd say that that's a matter of opinion.  It'sC >>certainly different from UNIX, but the file system on VMS is also H >>different from a UNIX file system, and that leads to such differences. > M > I don't see how the filesystem should have an influence on the way fwrite()  > works.  I A performance optimization, because writing a few big chunks is far more  H efficient that writing a lot of little chunks.  This takes advantage of D CPU instructions and DMA hardware that can do block moves.  OpenVMS 6 traditionally does less caching of I/O than UNIX does.   > $ >>>The correct call in libschily is: >>>   >>>cnt = fwrite(buf, 1, len, f); >>> I >>>It is the only way that works as unfortunately, the fwrite return code L >>>is "nitems", we need to set the "itemsize" parameter to 1 to get a useful
 >>>result. >>D >>  Oh.  I was careless with the byte count.  So, what's wrong with: >>& >>  cnt = len* fwrite(buf, len, 1, f); >  >  > It does not work correctly.   F Correct because the standard does not provide a method of determining A how may bytes were written.  But this is only important if it is  I possible to continue writing that file from that point with out starting   over.   G So what really matters is what use is "cnt" to handling the error case.   B > write(2) may return less than the expected amount in some cases.D > In this case, fwrite(buf, len, 1, f); would return 0 although some > bytes have been written.  . The behavior of write(2) is not an issue here.  G In the case of OpenVMS in the default "record" mode, either a complete  C item will have been written or not at all.  There is a performance  H penalty for doing single byte writes when the I/O subsystem is designed  for handling blocks.  I There may or may not be a performance penalty on different UNIX or other  B platforms for doing single byte writes over writing larger items, C depending on if they have coded their library to take advantage of  ! hardware that can do block moves.   I What has happened here to cause a porting problem on OpenVMS is that the  I default behavior of fwrite() from the early days was to assume that each  D item was a separate record to be written out to a device, for files  opened in record mode.  5 This behavior on files is actually controlled by the  D fopen()/create()/open() calls, and if not explicitly set by OpenVMS G specific extra arguments to that call are inherited by either existing   files or the device type.   E Because of the "dynamic" default behavior, new disk files tend to be  F created in the "stream" format that UNIX programs expect, unless they H are replacing a record format file, or are being sent to a special file 3 like stdout, stderr.  So many never see this issue.   I Unfortunately, the stream behavior expected by UNIX can not be forced by  G a run-time or compile time option transparent to the source code.  And  E there are quite a variety of options that may be needed to produce a  = file that matches what a program ported from UNIX may expect.   F Using a macro to override the fopen() and such calls is typically the  easiest way to do this.   I By using the /FIRST_INCLUDE qualifier on current OpenVMS C/C++ compilers  G either in an OpenVMS specific build procedure, or included on an alias  D to the CC command, it is possible in many cases to make such macros B override the C library routines with out changing the source code.   /* File vms_1st_inc.h */ #include <stdio.h>0 #define fopen(a, b, c) fopen(a, b, c, "ctx=stm") > F >>So far as I can see, it does the same thing in this case.  (I assumeG >>that ferror() will still return a non-zero value if not all the bytes J >>were written.)  It might be more fun to complain about the C RTL in VMS,= >>but is that more productive than making so simple a change?  >   F > ferror() will not return != 0 in case write(2) returns less than the > expected amount.    G But that has no bearing on the behavior of fwrite() which must set the  G error bit for ferror() to report if any error occurred in the routine.  D That is explicitly stated on the standard page you referenced in an 
 earlier post.   I So if you can not rely on ferror() on a UNIX platform for fwrite doing a  E block write instead of a number of byte writes, then it is that UNIX  1 platform that is not compliant with the standard.   6 > In this case, fwrite(buf, len, 1, f); will return 0.  G The issue is that on some platforms, unless you do single byte writes,  G you can not tell how much actually got written, in the case that it is  7 possible to do error recovery by retrying the fwrite().   " How important is this information?   > I >>  For the normal reader, the code in question is in the latest cdrtools ' >>kit, "[.libschily.stdio]filewrite.c":  >> >>[...] # >>       if (my_flag(f) & _IOUNBUF) 5 >>               return (write(fileno(f), buf, len)); C >>       cnt = fwrite(buf, 1, len, f);                   /* <--- */  >> >>       if (!ferror(f)) >>               return (cnt);( >>       if (!(my_flag(f) & _IONORAISE))" >>       raisecond(_writeerr, 0L); >>       return (-1);  >>[...]  >>G >>  If this would be too awful for non-VMS systems, then I suppose that ( >>an "ifdef __VMS" might be needed here. >   R > An #ifdef would not help either as fwrite(buf, len, 1, f); is wrong in any case.  E On some applications on OpenVMS, a noticeable performance penalty is  E seen  from doing single byte writes when a block write could be done.   5 Has anyone done such testing on various *NIX systems?   D Of course this fragment does not indicate of what "cnt" is used for.  F >>  In any case, I'm glad that you're looking at the proposed changes,F >>and I would also be glad to hear any other complaints you might have
 >>about them.  > P > Well, I tried to discuss the problem from the current patch set with Eberhard ) > Heuser-Hofmann but did not get a reply.  > M > I am willing to support VMS but it needs to be done in a way that does not   > affect other OS. > L > I thought that VMS has a POSIX compat layer these days. If it does, there + > should be no problem supporting cdrecord.   B Because of requiring backward source and binary compatibility, it G sometimes requires special compile options, or run time features to be  A enabled.  The run-time feature settings can be put in an OpenVMS  G specific source module with a special LIB$INITIALIZE psect, which only  - needs to be linked with the resulting binary.   E With OpenVMS 8.3, the bash and cc/gcc programs in the associated GNV  F layered product have been improved to the point that they can now run = almost all of the configure scripts that I have tried so far.   C Of course because configure is testing the "default" mode of the C  E library, and not the ANSI or POSIX modes of the C library, I usually  F have to set up some OpenVMS specific backdoor files so that configure G gets the correct answers.  Configure is also sometimes expecting a GCC  H specific behavior insures that unless special GNV backdoors are used it ' will always produce the wrong behavior.   F These configure issues impact other platforms besides OpenVMS, but in > general it just results in system libraries or routines being I incorrectly marked broken or missing and the code either replaces it, or  G assumes that a non-X/Open equivalent function is there instead, and it  D usually is.  So in the end, unless someone familiar with a platform C manually inspects what configure did on it, no one notices that the  result may not be optimal.  I On OpenVMS this is more noticeable, because where there is an X/Open API  C that is available as an option, the fallback non X/Open API is not  I there.  Plus because I know configure can not be trusted, and why it can  4 not be trusted, I am manually verifying the results.  I Those that attended the last OpenVMS bootcamp have a presentation on how  E to use those new GNV features, and they even have patches for GNV on  * OpenVMS 8.2/8.2-1 to start using them now.  C By the way, the last time I looked at MKISOFS, it had a hard coded  G assumption that the files "." and ".." would exist in a directory, and  2 considers it a fatal error if they do not show up.  E X/Open, the standard that you referred to earlier does not have that  G requirement, and as I recall, specifically indicates their presence is  0 optional and should not be expected by programs.   -John ! malmberg@dskwld.zko.hp.compaq.dec  Personal Opinion Only    ------------------------------    Date: 27 Jul 2006 01:45:47 -0700  From: "Ian Miller" <ijm@uk2.net># Subject: Re: PWS500au manual needed A Message-ID: <1153989947.901593.16810@i3g2000cwc.googlegroups.com>   C 2032 battery - flat round shiny 1/2" in diameter. Widely available. 2 Probably on the main board, held down with a clip.   ------------------------------  % Date: Thu, 27 Jul 2006 12:07:22 +0200 2 From: =?ISO-8859-15?Q?Jean-Fran=E7ois_Pi=E9ronne?=1 Subject: Re: Tomcat user authentication question. 4 Message-ID: <44c89061$0$895$ba4acef3@news.orange.fr>   Arne Vajhj wrote: > Tom Linden wrote:  > H >> On Wed, 19 Jul 2006 14:04:46 -0700, Rick Barry <richard.barry@hp.com>	 >> wrote:  >>E >>> CSWS_JAVA (Tomcat) does not yet contain a built-in authentication 
 >>> mechanism D >>> for OpenVMS. You could architect such a mechanism using JAAS and >>> Tomcat'sK >>> JAAS Realm plug-in, but you'll need to do a bit of coding using the JNI G >>> interface. We may add this to a future version of CSWS_JAVA. In the ( >>> meantime, you need to roll your own. >> >>= >> Alternatively, you use a web server designed for VMS, WASD  >  > = > Considering that Tomcat is *not* a web server but a servlet 7 > container, then I doubt that WASD would do much good.  >     E but Webware + WASD can do this and more (running the server under the  authenticate user account).    JF   ------------------------------  % Date: Thu, 27 Jul 2006 06:54:02 -0400 ) From: "Neil Rieck" <n.rieck@sympatico.ca> 1 Subject: Re: Tomcat user authentication question. < Message-ID: <44c8a055$0$24196$9a6e19ea@news.newshosting.com>  D "Jean-Franois Pironne" <jf.pieronne@laposte.net> wrote in message . news:44c89061$0$895$ba4acef3@news.orange.fr... > Arne Vajhj wrote: [...snip...] >  > G > but Webware + WASD can do this and more (running the server under the  > authenticate user account).  >  > JF  $ Under the authenticate user account?   Could you expand on this?   
 Neil Rieck Kitchener/Waterloo/Cambridge,  Ontario, Canada.! http://www3.sympatico.ca/n.rieck/    ------------------------------   Date: 27 Jul 2006 15:40:13 GMT From: healyzh@aracnet.com % Subject: Upgrading disks in Shadowset , Message-ID: <eaamot122d9@enews1.newsguy.com>   I am running OpenVMS 7.3-2.   F With my previous disk upgrade (4GB -> 18GB), I was able to start doingL volume shadowing on two of my disks.  Now it's time to replace at least 5 ofG the 6 18GB drives with 36GB drives (the driving factor is moving to LVD L Drives).  Always before when I've upgraded a disk, I've simply booted off of CD and done a BACKUP/IMAGE.   K What is the best way to move an 18GB shadowset to a 36GB shadowset?  As I'm L also in the process of hooking up a DLT drive, tape is once again an option.   		Zane   ------------------------------    Date: 27 Jul 2006 09:30:00 -0700  From: "Ian Miller" <ijm@uk2.net>) Subject: Re: Upgrading disks in Shadowset C Message-ID: <1154017800.508560.198820@b28g2000cwb.googlegroups.com>   B I mean look at the docs re DDS - shadowing of dissimilar devices - introduced with 4 V7.3-2. You could do this and keep the disks online.   ------------------------------    Date: 27 Jul 2006 09:28:45 -0700  From: "Ian Miller" <ijm@uk2.net>) Subject: Re: Upgrading disks in Shadowset C Message-ID: <1154017725.550419.127530@h48g2000cwc.googlegroups.com>    look at the docs re DVE    ------------------------------    Date: 27 Jul 2006 09:41:29 -0700 From: etmsreec@yahoo.co.uk) Subject: Re: Upgrading disks in Shadowset C Message-ID: <1154018489.569958.317610@i42g2000cwa.googlegroups.com>   B With moving forwards to bigger disks you may also want to considerD doing an init on the volumes to increase the maximum number of filesC and potentially modify the cluster size to something smaller.  This B will, of course, necessitate an image copy of the data rather than shadowing the bigger disks in.  F Depending on hardware available, tape can be quicker than disk to disk6 backup these days.  you mileage, of course, will vary.   Steve   # dave.baxter@bannerhealth.com wrote:  > healyzh@aracnet.com wrote: > > I am running OpenVMS 7.3-2.  > > J > > With my previous disk upgrade (4GB -> 18GB), I was able to start doingP > > volume shadowing on two of my disks.  Now it's time to replace at least 5 ofK > > the 6 18GB drives with 36GB drives (the driving factor is moving to LVD P > > Drives).  Always before when I've upgraded a disk, I've simply booted off of > > CD and done a BACKUP/IMAGE.  > > O > > What is the best way to move an 18GB shadowset to a 36GB shadowset?  As I'm P > > also in the process of hooking up a DLT drive, tape is once again an option. > > 
 > > 		Zane > D > Since you are on 7.3-2, this is a relatively simple task using DDSH > (Dissimilar Device Shadowing) and DVE (Dynamic Volume Expansion), both: > of which are enhancements which came with OpenVMS 7.3-2. > I > Basically, DDS allows you to add a 36GB drive to an 18GB shadowset, and F > then when the shadow is complete, DVE allows you to Dynamically grow8 > the volume from the original 18GB, up to the new 36GB. > G > There is one "gotcha", which probably requires downtime, at least for E > your apps.   (I havent had to apply this to a system disk yet, so I  > dont have a pat procedure.)  > F > Each shadow volume (DSA volume) must have DVE enabled, and a maximumG > "potential" size set up.     This requires that the volume is mounted H > privately, i.e. only to your process.      So for each volume you must >  > $ dismount /cluster DSAnnn: G > $ Mount/NoAssist DSAnnn:/Shad=(<orig_unit_1:>,<orig_unit_2:>) <label>  > <logical>  > $ Set Volume DSAnnn: /Limit  > $ Dismount DSAnnn:B > $ Mount/Clust/NoAss DSAnnn:/Shad=(<orig_unit_1:>,<orig_unit_2:>) > <label> <logical>  > D > Note you should read the Help on "Set Vol /Limit",  however if the? > Cluster Size is > 8, and no value is supplied with the /Limit 3 > Qualifier, then the default max size will be 1TB. E > A show dev /full on the DSA Vol will tell you if this has been set.  > B > Note:: you dont need to do this to be able to use DDS, (the onlyI > requirement for DDS is that the new disk be equal to or larger than the F > existing volume),  however if this is not done, then you will not be* > able to grow the volume to its new size. > . > Once it is done, follow the following steps. > 2 > 1.   Add the new (larger) disk to the shadowset.) > 2.   Allow the shadow copy to complete. H > 3.   Reduce the shadow set by removing the smaller, old units (leaving > only the new) ? > 4.   Execute (on all nodes)   " $  Set Volume /Size  DSAnnn:" : > 5.   Add second (larger) unit to complete the shadowset.( > 6.   "Bob's your Uncle!!"   Your done. > A > The best way to tell if you have been successful is to note the   > "Freeblocks" before and after. >  > Dave.    ------------------------------    Date: 27 Jul 2006 09:34:01 -0700" From: dave.baxter@bannerhealth.com) Subject: Re: Upgrading disks in Shadowset C Message-ID: <1154018041.458376.157280@h48g2000cwc.googlegroups.com>    healyzh@aracnet.com wrote: > I am running OpenVMS 7.3-2.  > H > With my previous disk upgrade (4GB -> 18GB), I was able to start doingN > volume shadowing on two of my disks.  Now it's time to replace at least 5 ofI > the 6 18GB drives with 36GB drives (the driving factor is moving to LVD N > Drives).  Always before when I've upgraded a disk, I've simply booted off of > CD and done a BACKUP/IMAGE.  > M > What is the best way to move an 18GB shadowset to a 36GB shadowset?  As I'm N > also in the process of hooking up a DLT drive, tape is once again an option. >  > 		Zane  B Since you are on 7.3-2, this is a relatively simple task using DDSF (Dissimilar Device Shadowing) and DVE (Dynamic Volume Expansion), both8 of which are enhancements which came with OpenVMS 7.3-2.  G Basically, DDS allows you to add a 36GB drive to an 18GB shadowset, and D then when the shadow is complete, DVE allows you to Dynamically grow6 the volume from the original 18GB, up to the new 36GB.  E There is one "gotcha", which probably requires downtime, at least for C your apps.   (I havent had to apply this to a system disk yet, so I  dont have a pat procedure.)   D Each shadow volume (DSA volume) must have DVE enabled, and a maximumE "potential" size set up.     This requires that the volume is mounted F privately, i.e. only to your process.      So for each volume you must   $ dismount /cluster DSAnnn: E $ Mount/NoAssist DSAnnn:/Shad=(<orig_unit_1:>,<orig_unit_2:>) <label> 	 <logical>  $ Set Volume DSAnnn: /Limit  $ Dismount DSAnnn:@ $ Mount/Clust/NoAss DSAnnn:/Shad=(<orig_unit_1:>,<orig_unit_2:>) <label> <logical>   B Note you should read the Help on "Set Vol /Limit",  however if the= Cluster Size is > 8, and no value is supplied with the /Limit 1 Qualifier, then the default max size will be 1TB. C A show dev /full on the DSA Vol will tell you if this has been set.   @ Note:: you dont need to do this to be able to use DDS, (the onlyG requirement for DDS is that the new disk be equal to or larger than the D existing volume),  however if this is not done, then you will not be( able to grow the volume to its new size.  , Once it is done, follow the following steps.  0 1.   Add the new (larger) disk to the shadowset.' 2.   Allow the shadow copy to complete. F 3.   Reduce the shadow set by removing the smaller, old units (leaving
 only the new) = 4.   Execute (on all nodes)   " $  Set Volume /Size  DSAnnn:" 8 5.   Add second (larger) unit to complete the shadowset.& 6.   "Bob's your Uncle!!"   Your done.  ? The best way to tell if you have been successful is to note the  "Freeblocks" before and after.   Dave.    ------------------------------    Date: 26 Jul 2006 22:57:44 -0700/ From: "David B Sneddon" <dbsneddon@bigpond.com> / Subject: Re: Using DSF and DMP files with debug C Message-ID: <1153979864.046884.183070@m79g2000cwm.googlegroups.com>    ababeel wrote: > Thanks for your comment,I > I am not actually debugging code, but finding a mechanism to do so. Let 
 > me explain, ' > contents of omartest.c are as follows  > 
 > void main()  > {  > 	char * p = malloc(32);  > 	int x = 10;	 > 	if(!p) ( > 		printf("Error Allocating Memory\n"); > 	else ! > 		printf("Allocated memory\n");  >  > 	strncpy(p,"Hello World",11);  >  > 	p=0; $ > 	printf(p); // will access violate >  >   > 	printf("\nFreeing p now..."); > 	free(p);  > 	printf("Done"); > }  > D > I know where the problem is in the code, I am trying to find a wayE > using a debugger to get to that offending line so that in future if D > there are any access violations I can debug them using DSF and DMP& > files. Hope I explained it properly. > Thanks again....  @ Using the above code, compiling and linking, I get the DMP file.  " tardis_FTA7> analyze/proc omartest  /          OpenVMS Alpha Debug64 Version V7.3-200     C %SDA-I-SINGLEMEM, single member shadow set; accessing dump file via 	 _DSA6000: ; %SYSTEM-F-ACCVIO, access violation, reason mask=00, virtual  address=000000000000& 0000, PC=FFFFFFFF80EA9F78, PS=0000001BA break on unhandled exception at SHARE$DECC$SHR_EV56_CODE0+1056632 : %DEBUG-I-SOURCESCOPE, Source lines not available for 0\%PC#         Displaying source for 2\%PC  DBG> show callG  module name    routine name     line           rel PC           abs PC <  SHARE$DECC$SHR_EV56_DATA1                  0000000000101F78 FFFFFFFF80EA9F78<  SHARE$DECC$SHR_EV56_DATA1                  0000000000103BC0 FFFFFFFF80EABBC0< *OMARTEST       main               13       0000000000000148 0000000000020148< *OMARTEST       __main              1       0000000000000064 0000000000020064<                                             FFFFFFFF80269ED4 FFFFFFFF80269ED4 DBG>  6 The above indicates line 13 in the code... which it is   Dave   ------------------------------    Date: 26 Jul 2006 23:43:03 -0700' From: "ababeel" <farooq.omar@gmail.com> / Subject: Re: Using DSF and DMP files with debug C Message-ID: <1153982583.906109.166340@m73g2000cwd.googlegroups.com>   A Thanks again...I had tried that but as u see I didnt get the line 2 number u reckon it could be a version problem ????  . Also I dunno why I get the following statement  E "target image file OMARTEST does not match with the image recorded in 
 the dumpfile"   % farooq.develop> analyze/proc omartest   /          OpenVMS Alpha Debug64 Version V8.2-017     E %DEBUG-I-AIMGMISMATCH, target image file OMARTEST does not match with " the image recorded in the dumpfile; %SYSTEM-F-ACCVIO, access violation, reason mask=00, virtual : address=0000000000000000, PC=FFFFFFFF81377E78, PS=0000001BA break on unhandled exception at SHARE$DECC$SHR_EV56_CODE0+1080952 
 DBG> sho call G  module name    routine name     line           rel PC           abs PC <  SHARE$DECC$SHR_EV56_DATA1                  0000000000107E78 FFFFFFFF81377E78<  SHARE$DECC$SHR_EV56_DATA1                  0000000000109C30 FFFFFFFF81379C30<                 MAIN                        0000000000010148 0000000000020148<                 __MAIN                      0000000000010064 0000000000020064<                                             FFFFFFFF8032FF94 FFFFFFFF8032FF94 DBG>  
 Thanks....   David B Sneddon wrote: > ababeel wrote: > > Thanks for your comment,K > > I am not actually debugging code, but finding a mechanism to do so. Let  > > me explain, ) > > contents of omartest.c are as follows  > >  > > void main()  > > {  > > 	char * p = malloc(32);  > > 	int x = 10; > > 	if(!p) * > > 		printf("Error Allocating Memory\n");	 > > 	else # > > 		printf("Allocated memory\n");  > > ! > > 	strncpy(p,"Hello World",11);  > > 	 > > 	p=0; & > > 	printf(p); // will access violate > >  > > " > > 	printf("\nFreeing p now...");
 > > 	free(p);  > > 	printf("Done"); > > }  > > F > > I know where the problem is in the code, I am trying to find a wayG > > using a debugger to get to that offending line so that in future if F > > there are any access violations I can debug them using DSF and DMP( > > files. Hope I explained it properly. > > Thanks again.... > B > Using the above code, compiling and linking, I get the DMP file. > $ > tardis_FTA7> analyze/proc omartest > 1 >          OpenVMS Alpha Debug64 Version V7.3-200  >  > E > %SDA-I-SINGLEMEM, single member shadow set; accessing dump file via  > _DSA6000: = > %SYSTEM-F-ACCVIO, access violation, reason mask=00, virtual  > address=000000000000( > 0000, PC=FFFFFFFF80EA9F78, PS=0000001BC > break on unhandled exception at SHARE$DECC$SHR_EV56_CODE0+1056632 < > %DEBUG-I-SOURCESCOPE, Source lines not available for 0\%PC% >         Displaying source for 2\%PC  > DBG> show callI >  module name    routine name     line           rel PC           abs PC > >  SHARE$DECC$SHR_EV56_DATA1                  0000000000101F78 > FFFFFFFF80EA9F78> >  SHARE$DECC$SHR_EV56_DATA1                  0000000000103BC0 > FFFFFFFF80EABBC0> > *OMARTEST       main               13       0000000000000148 > 0000000000020148> > *OMARTEST       __main              1       0000000000000064 > 0000000000020064> >                                             FFFFFFFF80269ED4 > FFFFFFFF80269ED4 > DBG> > 8 > The above indicates line 13 in the code... which it is >  > Dave   ------------------------------    Date: 26 Jul 2006 23:53:59 -0700/ From: "David B Sneddon" <dbsneddon@bigpond.com> / Subject: Re: Using DSF and DMP files with debug C Message-ID: <1153983239.409113.141300@m79g2000cwm.googlegroups.com>   5 May I suggest that you start with an empty directory.  Create the source file. 
 $ cc omartest  $ link/notraceback/dsf omartest  $ set process/dump $ run omartest $ set process/nodump $ analyze/process omartest   Dave   ------------------------------  % Date: Thu, 27 Jul 2006 18:21:37 +0200 2 From: martin@radiogaga.harz.de (Martin Vorlaender)/ Subject: Re: Using DSF and DMP files with debug ; Message-ID: <44c8e811.524144494f47414741@radiogaga.harz.de>   . David B Sneddon <dbsneddon@bigpond.com> wrote: > ababeel wrote:( >> contents of omartest.c are as follows ... E >> I know where the problem is in the code, I am trying to find a way F >> using a debugger to get to that offending line so that in future ifE >> there are any access violations I can debug them using DSF and DMP ' >> files. Hope I explained it properly.  >> Thanks again....  > B > Using the above code, compiling and linking, I get the DMP file. > $ > tardis_FTA7> analyze/proc omartest > 0 >         OpenVMS Alpha Debug64 Version V7.3-200 >  > = > %SYSTEM-F-ACCVIO, access violation, reason mask=00, virtual  > address=000000000000( > 0000, PC=FFFFFFFF80EA9F78, PS=0000001BC > break on unhandled exception at SHARE$DECC$SHR_EV56_CODE0+1056632 < > %DEBUG-I-SOURCESCOPE, Source lines not available for 0\%PC$ >        Displaying source for 2\%PC > DBG> show callH > module name    routine name     line           rel PC           abs PC= > SHARE$DECC$SHR_EV56_DATA1                  0000000000101F78  > FFFFFFFF80EA9F78= > SHARE$DECC$SHR_EV56_DATA1                  0000000000103BC0  > FFFFFFFF80EABBC0> > *OMARTEST       main               13       0000000000000148 > 0000000000020148> > *OMARTEST       __main              1       0000000000000064 > 0000000000020064= >                                            FFFFFFFF80269ED4  > FFFFFFFF80269ED4 > DBG> > 8 > The above indicates line 13 in the code... which it is  F Beware that when you #include files (like e.g. stdio, stdlib, string -D so that you don't get screenfuls of IMPLICITFUNC...) you come out atG line 2786 for the above. You need a listing (CC /LIST/SHOW=ALL) to find  the line number in that case.    cu,    Martin --  <                         | Martin Vorlaender | OpenVMS rules!1  OpenVMS: Where do you  | work: mv@pdv-systeme.de D  want to BE today?      |   http://www.pdv-systeme.de/users/martinv/8                         | home: martin@radiogaga.harz.de   ------------------------------  % Date: Thu, 27 Jul 2006 20:05:47 +0800 3 From: "Richard Maher" <maher_rj@hotspamnotmail.com> N Subject: Re: VMS and HPVM (was: Parsec webinar (2006-07-12) OpenVMS Licensing)1 Message-ID: <eaaa2a$qp9$1@news-02.connect.com.au>    Hi Rob,   > > No, the virtual VMS instances know nothing about each other.  & Oh well, thanks for the answer anyway.  @ Why doesn't VMS Engineering make up for it by introducing nativeH Cache-Fusion support? The DLM appears to be the most likely place for itJ (apparently that's how Oracle does it) A new flag that says "When the LockJ Value Block changes ownership to another node then send this many bytes atJ this address along with it". If only we could get some of that Bridgeworks budget. . .    Regards Richard Maher   ; "Rob Brooks" <brooks@cuebid.zko.hp.nospam> wrote in message & news:Yp+divgOizNw@cuebid.zko.hp.com...7 > "Richard Maher" <maher_rj@hotspamnotmail.com> writes:  > 2 > >> Does anyone here actually use OpenVMS Galaxy? > > H > > With these VMs be capable of sharing a Global Section between nodes? > > > No, the virtual VMS instances know nothing about each other. >  > --   > J > Rob Brooks    VMS Engineering -- Exec Group     brooks!cuebid.zko.hp.com   ------------------------------   End of INFO-VAX 2006.415 ************************