1 INFO-VAX	Sat, 13 Sep 2003	Volume 2003 : Issue 507       Contents:+ Re: "HP breakup on the way" - Merrill Lynch + Re: "HP breakup on the way" - Merrill Lynch 1 Re: Advanced Server - copying large file problem? % Re: Choosing between ASTs and Threads   Re: Cleaning up license database0 Re: First VMS IPF difference... not very nice...0 Re: First VMS IPF difference... not very nice... HSD50 Cache Memory( Re: IDL on VMS. Was: Image tools for VMS( Re: IDL on VMS. Was: Image tools for VMS Re: Image tools for VMS & Re: Mac OS X -> Pathworks/Mac problems RE: MESSAGE problem  Re: MESSAGE problem  Re: Newbie question about NTP / Porting from UNIX to OpenVMS for a real newbie. 3 Re: Porting from UNIX to OpenVMS for a real newbie. # Re: Question on running GnuPG 1.2.0 # Re: Question on running GnuPG 1.2.0 # Re: Question on running GnuPG 1.2.0 8 Strange DCL Problem:  DCL Can't Find My GOSUB Labels....< Re: Strange DCL Problem:  DCL Can't Find My GOSUB Labels.... VMS and MP3s2 Re: Why does MAIL DIR go slow during mail receive?  F ----------------------------------------------------------------------    Date: 12 Sep 2003 12:46:37 -0500; From: koehler@eisner.nospam.encompasserve.org (Bob Koehler) 4 Subject: Re: "HP breakup on the way" - Merrill Lynch3 Message-ID: <pgCTU86H1tCw@eisner.encompasserve.org>    In article <bjsv17$9ke$1@new-usenet.uk.sun.com>, Andrew Harrison SUNUK Consultancy <andrew_nospam.harrison_remove_this@sun#.com> writes: > 6 > Ohh goody notebooks I guess the board can go back to2 > the shareholders and say that it has in fact all2 > turned out OK because although HP's market share8 > in every other market has declined they have increased% > their share of the notebook market.   C    Since notebooks are outselling desktops that's an important step A    for any company that thinks it will someday make money on PCs.    ------------------------------  # Date: Fri, 12 Sep 2003 18:42:08 GMT & From: Rick Jones <foo@bar.baz.invalid>4 Subject: Re: "HP breakup on the way" - Merrill Lynch2 Message-ID: <4so8b.4642$mr2.2208@news.cpqcorp.net>  P Andrew Harrison SUNUK Consultancy <Andrew_No.Harrison_No@nospamn.sun.com> wrote:A > 0.3% is the theoretical maximum. Sun shipped a resonable number D > of V210/V240's before the NIC manufacturing defect was discovered.  C I am not sure if that is positive or not. I think you mean to imply > that Sun shipped a non-trivial number of systems before anyone noticed.  C > We didn't to my knowledge have any reports of network performance E > issues, the issue with the defect was triggered by a customer using F > a non IP protocol. I am working with an ex OpenVMS exchange customerB > who ran a series of network performance tests on V240's with theE > defect for their WAN connected gateway servers they didn't find any D > anomilies and were rather disgruntled when we refused to ship themD > any more systems after we discovered the defect because it clearly > didn't impact them.   C Indeed, on a WAN, where one is likely to already have a non-trivial E packet loss rate measured in single digits or more to the left of the ? decimal an additional 0.3% is likely in the noise.  I thought I D already agreed with that :) And those systems tested were _known_ toC have the defect, or simply from the batch that _could_ have had the  defect?   C > When the issue was discovered Sun stopped all V210/V240 deliverys A > and we are replacing all the affected V210/240 motherboards out  > in the field.    That's nice.  F > Ironically your estimates of the theoretical effects 0.3% error rateD > in Sun's NIC are rather similar to the actual effect of the remedyF > offered by Intel to Itanium 2 customers with the most recent Itanium( > bug, clock back from 1 Ghz to 800 Mhz.  1 > http://www.computerweekly.com/Article121736.htm   < I see a vague description of the problem, but not any of theD performance effect of dropping frequency by 20%.  One might assume aD 20% loss in perf, but that is less than what I was calculating for aD 0.3% packet loss rate when the rtx's had to be timeout based (65% or more).  @ And, if you want to discuss the perf effet of the workaround, weF should then also discuss the effect of using just a quad fast-ethernetE interface rather than the four GbE interfaces in the V210/V240, which A was Sun's suggested workaround to their GbE problem... unless the F V210/V240 do not have the CPU horseposer to do more than 400 Mbit/s itC was more than 20% for those bulk-transfers that may or may not have C been affected by retransmissions over the GbE links.  The treatment 3 then may have actually been worse than the disease.   : IIRC you have been fond to point-out the number of ItaniumE processors/systems shipped over time, and since you've said here that F the number of V210's and V240's Sun shipped before discovering the bugE in the component of their system was "reasonable," was that number of E systems greater, equal to, or less than the number of Itanium systems D shipped before the bug you reference in the URL above was discoveredB and screened-out?  (the URL you reference seems to be from May, soD before Madison and Deerfield) If greater or equal than it would seemC the number of Itanium systems that made it out into the world would D also be reasonable if the number of broken V2[1]40's Sun shipped was reasonable.   
 rick jones  	 > Regards  > Andrew Harrison F >> Some old data showing the effect on bulk transfer of having to wait9 >> for RTX timeouts at various RTX rates can be found at:  >>  @ >> ftp://ftp.cup.hp.com/dist/networking/briefs/SE_372_2_99.ps.gz >>  ? >> Which showed that for a 0.01% loss rate the timeout-only TCP H >> connection lost 20% performance, going to a loss of about 65% at 0.1% >> and 95% by 1%.  >>  C >> When the connection was able to fast RTX, a 0.1% throughput held = >> steady (with a 32KB window), and was down about 55% at 1%.  >>  G >> Also, while 0.3% might be background noise for an Internet connected G >> server, for something in the department or intranet that is going to ' >> be quite large comapred with normal.  >>  F >> Lets assume we have a synchronous request/reply application going -G >> say very much like single instance of netperf TCP_RR :).  A UP J5000 H >> (440 MHz PA-8500) can do roughly 13100 single-byte TCP_RR's through aI >> J6825A GbE NIC (Broadcom like the V2[14]0, but 5701 rather than 5704). 9 >> That suggests an RTT of 1/13000 or ~0.08 milliseconds.  >>  H >> If there were a 0.3% retransmission rate, then that RTT would become: >>  " >>   (.997 * 0.08) + (.003 * 500)  >>  F >> or ~1.57 milliseconds per transaction, which would be a transaction! >> rate of ~635 or a loss of 95%.  >>  I >> (If one is using the typical tcp_rexmit_interval_min. I'm ass-u-me-ing D >> that Solaris doesn't use the 1 second default the RFC's suggest -$ >> HP-UX uses 500 ms as the default) >>  F >> Now, it is perhaps unreasonable to expect real applications to haveF >> 0.08 ms response time, and I've no idea if a V240 can achieve 13000H >> TCP_RR transactions per second through one of its GbE interfaces withE >> a single instance of netperf, packet losses or no, so lets instead C >> assume that the application RTT is 1 millisecond when the NIC is ? >> functioning correctly.  That suggests it would be doing 1000 E >> transactions per second without loss.  With that 0.3% loss rate we  >> instead have  >>   >>   (.997 * 1) + (.003 * 500) >>  H >> or 2.497 ms response time, which means a transaction rate of ~400 per' >> second or a 60% loss of performance.    --  H Wisdom Teeth are impacted, people are affected by the effects of events.F these opinions are mine, all mine; HP might not want them anyway... :)A feel free to post, OR email to raj in cup.hp.com  but NOT BOTH...    ------------------------------  % Date: Fri, 12 Sep 2003 14:42:31 -0400 $ From: "PEN" <paul.nuneznosp@mhp.com>: Subject: Re: Advanced Server - copying large file problem?, Message-ID: <bjt42t$hb1$1@hplms2.hpl.hp.com>  C "Tom Simpson" <thomas.simpson1@nospam.comcast.net> wrote in message ) news:p678b.418865$o%2.188682@sccrnsc02... J > Are you comparing an MS copy operation to an FTP file transfer?  I don'tK > think that will give you a good comparison.  I would think (and from what 0 > I've seen) FTP will be significantly faster... > 
 > Regards, > Tom  >  Hi,   E With the v7.3a release (ECO1 being the most current at the moment) of @ Advanced Server, FTP shouldn't be any more than 10-15% faster...  7 Engineering has worked a lot on performance recently...      Paul   ------------------------------  % Date: Sat, 13 Sep 2003 04:24:17 +0800 , From: Paul Repacholi <prep@prep.synonet.com>. Subject: Re: Choosing between ASTs and Threads- Message-ID: <87oexpg3gu.fsf@prep.synonet.com>   3 usenet_vms@lehrerfamily.com (Joshua Lehrer) writes:   - > John Santos <JOHN@egh.com> wrote in message 3 > news:<1030910172703.3498A-100000@Ives.egh.com>...   E >> I haven't done any thread programming, but AFAIK, a single process 0 >> can have multiple locks on the same resource.  ? > You need to be careful.  The lock manager knows nothing about E > threads.  If thread A has an exclusive lock, and thread B queues an C > exclusive lock, it will return immediately with a deadlock error. B > The queue manager doesn't know that you are going to release theC > lock in thread A at some point, at which point, thread B can have  > the lock.   A No, there are mearly 2 requests for the one lock, one of which is E alreadt granted. There is not YET a deadlock. However, the conversion B timer on the second request starts and if it times out, a deadlock search will be done.  D > There is a flag that you can set on the queueing of the request to > ignore deadlocks.   E No, it inhibits the timer on the conversion from trigering a deadlock F search. Some lock are designed to wait forever, IE a fail over system.C You DON'T want VMS doing useless deadlock searchs all the time that D nothing has failed :) Been on a system that had that problem. It was ugly!    --  < Paul Repacholi                               1 Crescent Rd.,7 +61 (08) 9257-1001                           Kalamunda. @                                              West Australia 6076* comp.os.vms,- The Older, Grumpier Slashdot. Raw, Cooked or Well-done, it's all half baked.F EPIC, The Architecture of the future, always has been, always will be.   ------------------------------  # Date: Fri, 12 Sep 2003 20:02:02 GMT 4 From: "Mark Buda" <buda@tabasco.zko.dec.no.spam.com>) Subject: Re: Cleaning up license database 2 Message-ID: <_Cp8b.4653$CB2.1692@news.cpqcorp.net>  C "Peter 'EPLAN' LANGSTOEGER" <peter@langstoeger.at> wrote in message - news:0Zn7b.12004$ef4.142165@news.chello.at... 1 > >I will add this to a WISH LIST for the future.  > G > Super. And while you are there, please add LICENSE UNLOAD * (or /ALL) 
 as well...   OK.   C Please share with me how you plan on using it and what is gained by  implementing it.     --    
 Sincerely,	 Mark Buda  Hewlett-Packard Company  VMS Engineering  110 Spitbrook Road
 MS: ZK3-4/X57  Nashua, NH 03062 Voice: (603) 884-1969  FAX: (603) 884-3451   * VMS Home Page http://www.hp.com/go/openvms   ------------------------------    Date: 12 Sep 2003 12:48:22 -0500; From: koehler@eisner.nospam.encompasserve.org (Bob Koehler) 9 Subject: Re: First VMS IPF difference... not very nice... 3 Message-ID: <ESYJK1xOtrxn@eisner.encompasserve.org>   U In article <00A25CB8.3D049EC1@SendSpamHere.ORG>,   VAXman-  @SendSpamHere.ORG writes:  > ' > Doesn't or didn't prior to V7.3(-1?).  > L > Just tried to define LIB$DEBUG to DELTA and run a simple program and I get > a stack dump.  > A > Really hell bent on keeping me out of elevated access modes eh?   H    Is there a real need to use DELTA for user mode code?  Or do you just    miss ODT et. al.?    8-)    ------------------------------  # Date: Fri, 12 Sep 2003 19:17:54 GMT " From:   VAXman-  @SendSpamHere.ORG9 Subject: Re: First VMS IPF difference... not very nice... 0 Message-ID: <00A25CC9.827D289A@SendSpamHere.ORG>  q In article <ESYJK1xOtrxn@eisner.encompasserve.org>, koehler@eisner.nospam.encompasserve.org (Bob Koehler) writes: V >In article <00A25CB8.3D049EC1@SendSpamHere.ORG>,   VAXman-  @SendSpamHere.ORG writes: >>  ( >> Doesn't or didn't prior to V7.3(-1?). >>  M >> Just tried to define LIB$DEBUG to DELTA and run a simple program and I get  >> a stack dump. >>  B >> Really hell bent on keeping me out of elevated access modes eh? > I >   Is there a real need to use DELTA for user mode code?  Or do you just  >   miss ODT et. al.?    8-) >   L Things might start out in user mode but are quickly promoted to inner modes.    --L 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: Fri, 12 Sep 2003 21:39:20 +0000 (UTC) ) From: "Jeff" <dirkdiggler@totalise.co.uk>  Subject: HSD50 Cache Memory / Message-ID: <bjtee7$4ds$1@titan.btinternet.com>   I Does anybody know what kind of SIMM memory chips are compatible to act as  cache in an HSD50 Controller ?  L The manual doesn't state the specification (speed, parity e.t.c) of the SIMM( modules that the controller can utilize.  J We currently have 32Mb of V.3 cache memory, and am looking to increase it.  L Have found some DEC part numbers for the memory, DS-HSSIM-AB seems to be the7 part number for the 64Mb kit.... but is very expensive.   J Does anybody know if there are 3rd party solutions ??  For example if it'sA simply 72pin 60ns SIMM's will some from www.crucial.com suffice ?   K Also does anybody have any good strategies for devising the characteristics @ of storage units i.e cache policies for units depending on use ?  & Thanks in advance for any information,   Jeff   ------------------------------  % Date: Fri, 12 Sep 2003 15:13:30 -0400 < From: "Peter Weaver" <WeaverConsultingServices@sympatico.ca>1 Subject: Re: IDL on VMS. Was: Image tools for VMS 9 Message-ID: <bjt5ss$mrl74$1@ID-141708.news.uni-berlin.de>    Jonathan Boswell wrote:  >... > RSI  > http://www.rsinc.com/  >...> > The following note was penned late last night by Jim Kelley, the RSI ; > product manager for IDL and OK'd by him for posting here. 
 Please sit > down before reading. >  >  - JB 9 >                 _______________________________________  > 9 > RSI dropped support for VMS several years after DEC had  dropped = > support for that operating system.  The decision by RSI was  driven  3 Did you (or anybody else) write back to correct his  misunderstanding?    >...8 > engineering challenge that would certainly not be cost effective,  we< > really could not replicate IDL 6.0 today on VMS because of the many> > 3rd party components and libraries that also no longer offer support  > for the VMS operating system.  >...  ; Which 3rd party components and libraries? Maybe he is wrong  about this also.   --   Peter Weaver Weaver Consulting Services Inc.  Canadian VAR for CHARON-VAX  www.weaverconsulting.ca    ------------------------------  # Date: Fri, 12 Sep 2003 19:48:07 GMT # From: "John Smith" <a@nonymous.com> 1 Subject: Re: IDL on VMS. Was: Image tools for VMS J Message-ID: <Xpp8b.441472$4UE.396498@news01.bloor.is.net.cable.rogers.com>   Peter Weaver wrote:  > Jonathan Boswell wrote:  >> ... >> RSI >> http://www.rsinc.com/ >> ...G >> The following note was penned late last night by Jim Kelley, the RSI G >> product manager for IDL and OK'd by him for posting here. Please sit  >> down before reading.  >> >>  - JB: >>                 _______________________________________ >>B >> RSI dropped support for VMS several years after DEC had droppedE >> support for that operating system.  The decision by RSI was driven  > 5 > Did you (or anybody else) write back to correct his  > misunderstanding?  >  >> ...H >> engineering challenge that would certainly not be cost effective,  weF >> really could not replicate IDL 6.0 today on VMS because of the manyG >> 3rd party components and libraries that also no longer offer support   >> for the VMS operating system. >> ... > = > Which 3rd party components and libraries? Maybe he is wrong  > about this also.   Peter,  L He's wrong about VMS and he may be wrong about which 3rd-party components heI needs are still available for VMS, but there is one thing he's absolutely L right about - that he can't push on a rope and get customers to buy VMS just because he has a toolset.   K HP has to create the product (VMS) awareness and base level demand for VMS. K If HP did that and it was quite visible in market share once again, lots of - ISV's would get involved with VMS once again.   K Say you were an ISV, but not just on VMS - in fact most of your business in K not on VMS. Your wouldn't have a lot of traction on your own marketing your I solution on VMS. You'd be looking for help from HP. But HP doesn't market F VMS to anyone except to 'large' current customers and other deals thatJ windup in their laps despite their best efforts to ignore them. HP doesn'tF advertise VMS, period. What impression would you have about the futureL prospects for your VMS application in terms of market acceptance and uptake?  J Of course there are exceptions but you have a wife, mortgage, a drawn downJ line of credit at the bank for your business, two kids still in braces andK summer camp, and two kids at college.What course of action would you likely G follow with respect to your VMS business - contine to invest in the VMS K version and hope that business walks in the door? - cut your losses? launch I a marketing campaign extolling the virtues of your product on VMS without   co-op advertising money from HP?   ------------------------------  # Date: Fri, 12 Sep 2003 19:37:46 GMT - From: "John E. Malmberg" <wb8tyw@qsl.network>   Subject: Re: Image tools for VMS: Message-ID: <egp8b.572$Pd2.540346@news1.news.adelphia.net>   VAXman- wrote: > J > 'Tis a shame too...  My HP SCSI scanner does a real nice job and quicklyD > too.  I've seen new USB scanners in operation and they are slow.   > K > If Jack can fine a SCSI scanner, he should contact me via a private email  > for the scanner software.   I This may be of interest to those of you that want to do your own hacking  K with HP all in one scanners with a JetDiret card that understands scanners.   A http://www.hp.com/cposupport/networking/support_doc/bpj01014.html   % Particularly ports 9280 through 9292.   G I do not know how hard it would be to adapt either VAXman's or a LINUX  # scanner driver to these on OpenVMS.    -John  wb8tyw@qsl.network Personal Opinion Only    ------------------------------   Date: 12 Sep 2003 23:08:06 GMT From: healyzh@aracnet.com / Subject: Re: Mac OS X -> Pathworks/Mac problems , Message-ID: <bjtjkm0167v@enews1.newsguy.com>  < Bob Koehler <koehler@eisner.nospam.encompasserve.org> wrote:O > In article <bjr0i70ofp@enews1.newsguy.com>, healyzh@NOaracnetSPAM.com writes:  > > L > > I actually happened to be playing with just this feat this morning afterP > > reading a Macworld article.  It worked OK to a Unix system, however, I endedO > > up with the same results with OpenVMS.  I'm using OpenVMS 7.2-1H1 and TCPIP  > > 5.3 with Mac OS X 10.2.6.   E >    Does TCP/IP 5.3 have an emulate-UNIX setting, like Multinet has?    No idea.   	Zane    ------------------------------    Date: 12 Sep 2003 15:40:10 -0500- From: Kilgallen@SpamCop.net (Larry Kilgallen)  Subject: RE: MESSAGE problem3 Message-ID: <U1YcrDyxCqbR@eisner.encompasserve.org>   _ In article <CIEJLCMNHNNDLLOOGNJIOEHAHPAA.tom@kednos.com>, "Tom Linden" <tom@kednos.com> writes:   ( >>> It is not in the PL/I documentation. >>E >>It looks to me like Section 12.2.2 of the PL/I User Manual explains F >>how to do it.  The Message value usage might be added to the list in% >>Section 12.2.3 to be more complete.  > I > Possibly, but in order to be complete then, I suppose one ought to list C > all global symbols and where they are used.  Is that what you are 
 > suggesting?   I No, but naming a compiler (MESSAGE) that does everything by symbol values  would help the PL/I programmer.    ------------------------------  % Date: Fri, 12 Sep 2003 22:29:08 +0100 & From: Edward Brocklesby <ejb@goth.net> Subject: Re: MESSAGE problem< Message-ID: <JSq8b.468$QO6.211@news-binary.blueyonder.co.uk>   briggs@encompasserve.org wrote: E > It's not an issue with the MESSAGE utility.  It's an issue with the G > language you use to write the code that is linked against the message  > file.  > E > When you resolve external references at link time, you need to tell A > your compiler whether you're importing addresses or values.  In ? > general and by default, your compiler will assume that you're  > importing addresses.  G Right -- I understand that specifying the mechanism used to access the  G symbol is a language issue; but how am I meant to know (without asking  I comp.os.vms) that the message id symbols are passed by value rather than  E address?  The only relevant text I can find from the message utility  
 manual is:  I "You can refer to the message code in your programs by means of a global  G symbol called the message symbol, which also is defined by information   from the message source file."  G A quick look through the PL/I user and reference manuals also fails to  I turn up anything which seems relevant (it describes the VALUE attribute,  I of course, but gives no indication that the MESSAGE utility requires it).   I Now, it seems to me there are two ways that such things could be listed;   either:   D 	a) Every language supported by VMS would list a description of how C MESSAGE globals should be declared, and then updating every manual  8 whenever a new utility is added or something changes; or  D 	b) Every language supported by VMS would list a description of how I global symbols should be passed by value (eg. section 12.2.2 in the PL/I  D user manual); then simply stating in the MESSAGE documentation that ' these symbols should be used as values.   F At least in my opinion, choice b) seems like the most obvious.  Maybe A I'm just strange, but I certainly would have expected to find an  A explanation of how MESSAGE symbols should be used in the MESSAGE  I manual--even if it's just a single sentence to the effect of "..., these  @ symbols should be accessed by value (for details, refer to your  language's documentation)".   G I'm not trying to lay blame at anyone's feet here, but would it really  B be /too/ much effort for this to be documented at least somewhere?   Regards, Edward.    ------------------------------  % Date: Sat, 13 Sep 2003 04:58:29 +0800 , From: Paul Repacholi <prep@prep.synonet.com>& Subject: Re: Newbie question about NTP- Message-ID: <87d6e5g1vu.fsf@prep.synonet.com>     david20@alpha2.mdx.ac.uk writes:  D > The implementation of ntpdate on VMS TCPIP Services does not allowF > it to be used when the server is an NTP server. I don't know whetherA > this limitation is restricted to VMS and TCPIP Services or is a E > general feature of ntpdate.  The usual use of ntpdate is to set the ? > time from a remote NTP server when you are not running an NTP D > server. Hence if you could run it on an NTP server system it would7 > only make sense to run it in query mode ie ntpdate -q   @ Don't use ntpdate if you can possibly avoid it. It is going awayD `soon' from ntp, and it can and will tell you lies. Use ntpd instead? to jump your clock. See www.ntp.org or twiki.ntp.org or the ntp : newsgroup. You will need to understand Millspeak though :)   --  < Paul Repacholi                               1 Crescent Rd.,7 +61 (08) 9257-1001                           Kalamunda. @                                              West Australia 6076* comp.os.vms,- The Older, Grumpier Slashdot. Raw, Cooked or Well-done, it's all half baked.F EPIC, The Architecture of the future, always has been, always will be.   ------------------------------    Date: 12 Sep 2003 18:05:52 -0700. From: drkirkby@ntlworld.com (Dr. David Kirkby)8 Subject: Porting from UNIX to OpenVMS for a real newbie.= Message-ID: <bbdc5d83.0309121705.4282adf3@posting.google.com>    Hi, @ 	Many many years ago I used VMS on a VAX - even writing some VAXE assembler at one point! But it's a long time and I can't rememver any ; commands. It is dir to list files ? I really have forgotten  everything.   E I have here a Dec Alpha 600a Personal Workstation which I think would B run VMS and/or OpenVMS, but I don't think I can be bothered buying- OpenVMS media, so I probably wont install it.   D However, I am aware of the OpenVMS systems at the HP testdrive site:  http://www.testdrive.compaq.com/  = I'm considering porting a UNIX command-line GPL'ed scientific  application of mine:   http://atlc.sourceforge.net/   to OpenVMS.   @ The code is I believe very portable on UNIX systems, having been> tested on a all the main Linux distributions, IBM's AIX, Sun's? Solaris, SCO's UNIXWARE, HP's Tru64, HP's HP-UX, Cray's UNICOS, - NetBSD, OpenBSD, FreeBSD ... and any I missed   F I've put a **lot** of effort into making it portable on UNIX, but haveA never given other operating systems too much thought. A couple of C other people have created command-line Windoze versions without too  much hassle.  @ The problem is the application uses the normal (in UNIX circles)@ autoconf/automake. What effort is likely to be needed to port to= OpenVMS? There are probably about 12000 lines of C and has 15 C binaries. Each C file uses a number of header files, which only get @ included if they exist on the UNIX system. So a test is made forC /usr/include/foo.h and foo.h is only included if it exists. On UNIX 6 systems, the following (typically) have to be defined:  > -DPACKAGE_VERSION=\"4.4.1\" -DPACKAGE_STRING=\"atlc\ 4.4.1\" \B -DPACKAGE_BUGREPORT=\"drkirkby@ntlworld.com\" -DPACKAGE=\"atlc\" \2 -DVERSION=\"4.4.1\" -DHAVE_LIBM=1 -DSTDC_HEADERS=1 -DHAVE_SYS_TYPES_H=1 \9 -DHAVE_SYS_STAT_H=1 -DHAVE_STDLIB_H=1 -DHAVE_STRING_H=1 \ : -DHAVE_MEMORY_H=1 -DHAVE_STRINGS_H=1 -DHAVE_INTTYPES_H=1 \F -DHAVE_UNISTD_H=1 -DHAVE_STDIO_H=1 -DHAVE_MATH_H=1 -DHAVE_MALLOC_H=1 \E -DHAVE_STDLIB_H=1 -DHAVE_STRING_H=1 -DHAVE_TIME_H=1 -DHAVE_LIMITS_H=1  \ @ -DHAVE_STRINGS_H=1 -DHAVE_SYS_TYPES_H=1 -DHAVE_SYS_SYSINFO_H=1 \@ -DHAVE_SYS_PROCESSOR_H=1 -DHAVE_SCHED_H=1 -DHAVE_SYS_PARAM_H=1 \F -DHAVE_SYS_UNISTD_H=1 -DHAVE_SYS_UTSNAME_H=1 -DHAVE_SYS_SYSTEMINFO_H=1 \ 6 -DHAVE_UNISTD_H=1 -DHAVE_ERRNO_H=1 -DWORDS_BIGENDIAN=1 -DSIZEOF_SHORT=2 \B -DSIZEOF_INT=4 -DSIZEOF_LONG=4 -DSIZEOF_SIZE_T=4 -DHAVE_MEMSET=1 \F -DHAVE_STRCHR=1 -DHAVE_STRTOL=1 -DHAVE_PROCESSOR_INFO=1 -DHAVE_UNAME=1 \ 3 -DHAVE_SYSCONF=1 -DHAVE_SYSINFO=1 -DHAVE_LIBGSL=1 \ 1 -DTRY_TO_GET_HARDWARE_INFO=1 -I/usr/local/include  -I/usr/local/include/gsl \  F in order that it will built. Are there any quick ways of porting this?@ Can I assume (unlike on UNIX systems) that VMS and OpenVMS has a> common set of header files? The code does try to find out someF hardware information, but I'm happy to remove that bit to make portingB easier. The code is multi-threaded to exploit multiple CPUs, but I? guess that is something that can be disabled - it is by default  anyway.   F Any comments welcome. I don't want to spend a lot of time over it, but2 if the effort is not too great, I might do a port.   Dr. David Kirkby, PhD.   ------------------------------   Date: 13 Sep 2003 05:15:55 GMT1 From: JONESD@er6.eng.ohio-state.edu (David Jones) < Subject: Re: Porting from UNIX to OpenVMS for a real newbie.: Message-ID: <bju96b$hhq$1@charm.magnus.acs.ohio-state.edu>  = In message <bbdc5d83.0309121705.4282adf3@posting.google.com>, 3    drkirkby@ntlworld.com (Dr. David Kirkby) writes: A >The problem is the application uses the normal (in UNIX circles) A >autoconf/automake. What effort is likely to be needed to port to > >OpenVMS? There are probably about 12000 lines of C and has 15
 >binaries.    * >Are there any quick ways of porting this?  H No.  Have you ever counted the number lines in the configure script that autoconf generates?     A >Can I assume (unlike on UNIX systems) that VMS and OpenVMS has a  >common set of header files?  K There is a common set of header files, but they have differences based upon L the compiler version (as they've gradually made the run-time more compatible with UNIX).   K I used the following command procedure for converting the samba config.h_in 6 to config.h, perhaps you could do something similar...  O -------------------------------------------------------------------------------  $!; $! Create config.h for VMS using config.h_in as a template.  $!F $! Save current default directory and search for parent of source treeJ $! [.include] sub-directory.  Begin at directory containing this procedure $! and look in all ancestors.  $! $ h_out_open = 0# $ defdir = f$environment("DEFAULT") " $ proc = f$environmen("PROCEDURE")( $ proc_dir = f$parse(proc,,,"DIRECTORY") $ set default 'proc_dir' $ find_inc_dir: + $     if f$search("include.DIR;1") .eqs. "" 
 $     then# $	curdir = f$environment("DEFAULT") E $	if f$parse(f$environment("DEFAULT"),,,"DIRECTORY") .eqs. "[000000]"  $	then4 $	    write sys$output "Include directory not found" $	    goto abort $	else $            set default [-] $	    goto find_inc_dir  $	endif  $     endif  $!2 $ open/read h_in [.include]config.h_in/error=abortB $ create [.include]config.h		! gives file normal text record attr. #ifndef CONFIGURE_VMS_H  #define CONFIGURE_VMS_H 1  /*D  * Make some global settings based on the OS and site confiugration.  */ C #include "vms_stat_wrapper.h"		/* Re-jigger the stat() functions */ : #include "vms.h"			/* Redirect rtns that need emulation */5 #include "conf.h"			/* site-specific configuration */   I #include "external_name_renames.h"	/* Shortens extern names > 31 chars */   4 $ open/append h_out [.include]config.h/error=cleanup $ h_out_open = 1 $ on error then goto cleanup  $ on control_y then goto cleanupK $ write h_out "/* This file created by configure_vms.com, ", f$time(), " */  $!F $! Load the macro definitions we want replaced in config.h_in into DCLL $! symbols.  Each symbol can hold multiple definitions, separated by commas.O $! The numeric suffixes at the end of the symbols must form a contiguous range.  $!D $ symlist1 = "_ALL_SOURCE,HAVE_MMAP,HAVE_SYS_WAIT_H,inline=__inline"L $ symlist2 = "RETSIGTYPE=void,STDC_HEADERS,TIME_WITH_SYS_TIME,HAVE_VOLATILE"L $ symlist3 = "HAVE_ERRNO_DECL,HAVE_LONGLONG,HAVE_UNSIGNED_CHAR,HAVE_UTIMBUF"? $ symlist4 = "HAVE_SIG_ATOMIC_T_TYPE,loff_t=off_t,HAVE_CONNECT" L $ symlist5 = "HAVE_FTRUNCATE_EXTEND,HAVE_GETTIMEOFDAY_TZ,STAT_STATFS2_BSIZE"E $ symlist6 = "HAVE_CRYPT,HAVE_FILE_MACRO,HAVE_CRYPT_DECL,HAVE_MEMSET" G $ symlist7 = "HAVE_STRCASECMP,SIZEOF_INT=4,SIZEOF_SHORT=2,HAVE___CHDIR" P $ symlist8 = "HAVE___CLOSE,HAVE___CLOSEDIR,HAVE___DUP,HAVE___DUP2,HAVE___FCHDIR"P $ symlist9 = "HAVE___FCNTL,HAVE___FSTAT,HAVE___GETCWD,HAVE___LSEEK,HAVE___LSTAT"D $ symlist10= "HAVE___OPEN,HAVE___OPENDIR,HAVE___READ,HAVE___READDIR"E $ symlist11= "HAVE___SEEKDIR,HAVE___STAT,HAVE___TELLDIR,HAVE___WRITE" P $ symlist12= "HAVE_ATEXIT,HAVE_BZERO,HAVE_CHMOD,HAVE_CHOWN,HAVE_DUP2,HAVE_EXECL"M $ symlist13= "HAVE_FSTAT,HAVE_FSYNC,HAVE_FTRUNCATE,HAVE_GETCWD,HAVE_GETGRNAM" O $ symlist14= "HAVE_INITGROUPS,HAVE_MEMMOVE,HAVE_MKTIME,HAVE_PATHCONF,HAVE_PIPE" H $ symlist15= "HAVE_RAND,HAVE_RANDOM,HAVE_RENAME,HAVE_SELECT,HAVE_SETENV"G $ symlist16= "HAVE_SIGACTION,HAVE_SIGBLOCK,HAVE_SIGPROCMASK,HAVE_SRAND" O $ symlist17= "HAVE_SRANDOM,HAVE_STRCHR,HAVE_STRDUP,HAVE_STRERROR,HAVE_STRFTIME" Q $ symlist18= "HAVE_STRPBRK,HAVE_STRTOUL,HAVE_UTIME,HAVE_WAITPID,HAVE_ARPA_INET_H" O $ symlist19= "HAVE_CTYPE_H,HAVE_DIRENT_H,HAVE_FCNTL_H,HAVE_GRP_H,HAVE_LIMITS_H" K $ symlist20= "HAVE_MEMORY_H,HAVE_NET_IF_H,HAVE_NETINET_TCP_H,HAVE_STDARG_H" H $ symlist21= "HAVE_STDLIB_H,HAVE_STRING_H,HAVE_STRINGS_H,HAVE_STROPTS_H"@ $ symlist22= "HAVE_SYS_IOCTL_H,HAVE_SYS_MMAN_H,HAVE_SYS_PARAM_H"B $ symlist23= "HAVE_SYS_SOCKET_H,HAVE_SYS_STATFS_H,HAVE_SYS_TIME_H"C $ symlist24= "HAVE_UNISTD_H,HAVE_UTIME_H,HAVE_STRNLEN,HAVE_INO64_T" @ $ symlist25= "HAVE_SNPRINTF_DECL,WITH_SMBPASSWD_SAM,USE_SETEUID" $ symlist26= "HAVE_SETGROUPS"  $!K $! Make an index of the macros we've defined by defining DCL symbols of the F $! form sidx_'symbol'="n,m" where the macro definition is obtained via $! f$element(m,",",symlist'n').  $! $ symnum = 0 $ bld_next_list: $    symnum = symnum + 1; $    if f$type(symlist'symnum') .eqs. "" then goto bld_done  $    symlist = symlist'symnum'
 $    i = 0 $   bld_next_element:X6 $        def = f$edit(f$element(i,",",symlist),"TRIM")+ $	 if def .eqs. "," then goto bld_next_listc# $        key = f$element(0,"=",def)r! $	 if f$type(sidx_'key') .nes. ""o $	 thengD $	    write sys$output "Warning, symbol ""''key'"" multiply defined" $	 endif& $        sidx_'key' = "''symnum',''i'" $	 i = i + 1 $	goto bld_next_element  $ bld_done:  $!A $! scan records of config.h_in and copy to output file (config.h)f $! $ line_count = 0, $ write sys$output "Scanning config.h_in..." $ next_line:$ $    read h_in line/end=nomore_lines  $    line_count = line_count + 1 $    out_line = line, $    tok_line = f$edit(line,"COMPRESS,TRIM")0 $    if f$element(0," ",tok_line) .eqs. "#undef"	 $    thena $!B $!	Substitue #undef line with definition if found or comment line. $!! $	key = f$element(1," ",tok_line)   $	if f$type(sidx_'key') .nes. "" $       then $	  idx = sidx_'key'2 $	   sidx_'key' = idx + "," + f$string(line_count)+ $	  symlist = symlist'f$element(0,",",idx)'I5 $	  def = f$element(f$element(1,",",idx),",",symlist)-  $	  value = f$element(1,"=",def)0 $	  out_line = "#define " + f$element(0,"=",def) $	  if value .nes. "=" $	  then5 $		out_line = out_line + " " + value	! explicit valueo $	  else- $		out_line = out_line + " 1"	! default valuen	 $	  endifs $	else$ $	   out_line = "/* " + line + " */"
 $       endif,
 $    endif  $    write/symbol h_out out_line $    goto next_lineh $! $ nomore_lines:  $!8 $ write sys$output "Total lines processed: ", line_count $! $! Check for null references.s $! $ symnum = 0 $ chk_next_list: $    symnum = symnum + 1; $    if f$type(symlist'symnum') .eqs. "" then goto chk_done  $    symlist = symlist'symnum'
 $    i = 0 $   chk_next_element:t6 $        def = f$edit(f$element(i,",",symlist),"TRIM")+ $	 if def .eqs. "," then goto chk_next_list # $        key = f$element(0,"=",def)-( $	 refline = f$element(2,",",sidx_'key') $	 if refline .eqs. ","@ $	 thennU $	    write sys$output "Warning, symbol ""''key'"" (symlist''symnum') not referenced"r" $	    value = f$element(1,"=",def)! $	    out_line = "#define " + keyo $	    if value .nes. "="
 $	    then5 $		out_line = out_line + " " + value	! explicit valueb
 $	    else- $		out_line = out_line + " 1"	! default valuee $	    endifo $	    write h_out out_line $	 endif $	 i = i + 1 $	goto chk_next_element  $ chk_done:i+ $ write h_out "#endif /* CONFIGURE_VMS_H */n $! $!
 $ cleanup: $ close h_in  $ if h_out_open then close h_out $ abort: $ set default 'defdir' $ exit  O -------------------------------------------------------------------------------   < David L. Jones               |      Phone:    (614) 292-6929- Ohio State University        |      Internet:eL 140 W. 19th St. Rm. 231a     |               jonesd@er6s1.eng.ohio-state.edu: Columbus, OH 43210           |               vman+@osu.edu  1 Disclaimer: I'm looking for marbles all day long.s   ------------------------------  % Date: Fri, 12 Sep 2003 16:03:38 -0500l* From: Patrick Spinler <pspinler@yahoo.com>, Subject: Re: Question on running GnuPG 1.2.0( Message-ID: <3F6234AA.6010408@yahoo.com>   $ nopriv $ gpgo  gpg: can't lock memory: error -1( %NONAME-E-NOMSG, Message number 00000002 $ set proc/priv= PSWAPMf $ gpg ' gpg: Go ahead and type your message ...a   -- Pat   Keith A. Lewis wrote:en > healyzh@NOaracnetSPAM.com writes in article <bjrgpc02agj@enews4.newsguy.com> dated 12 Sep 2003 04:07:08 GMT: > L >>I got the GnuPG 1.2.0 kit from HP and have finally gotten it running afterO >>rebuilding it from the source.  That is it will run from the SYSTEM account. aE >>What privileges are needed for a user account to be able to run it?t >  > N > The PGP I built back in the day did not require any special privileges.  AllK > it does is read and write files and some heavy number crunching, right?  u > I > Are you getting an error when you run it under a normal user account?  s > If so, post it.  > - > --Keith Lewis              klewis$mitre.orgf@ > The above may not (yet) represent the opinions of my employer.   ------------------------------  % Date: Fri, 12 Sep 2003 19:03:10 -0400r  From: John Santos <JOHN@egh.com>, Subject: Re: Question on running GnuPG 1.2.05 Message-ID: <1030912185539.3498B-100000@Ives.egh.com>V  + On Fri, 12 Sep 2003, Patrick Spinler wrote:d  
 > $ nopriv > $ gpg." > gpg: can't lock memory: error -1* > %NONAME-E-NOMSG, Message number 00000002 > $ set proc/priv= PSWAPM  > $ gpgk) > gpg: Go ahead and type your message ...a >  > -- Pat  G I don't know why it needs PSWAPM, but it does.  Other weirdnesses is iteH only works on STREAM_LF files, so you need to convert text files to/fromG STREAM_LF before/after using it.  I imagine for binary files, you couldiG just lie about the attributes with "$ set file/attribute=RFM:STMLF" andiE convert back to sequential fixed 512 after, but I haven't tried this.   @ The first time you run it, it creates a subdirectory (under yourC default login directory) to hold key rings, etc. and then complains 4 if the subdirectory has world or group read access.   +   $ set file/prot=(w,g) sys$login:gnupg.dird makes it shut up.t   >  > Keith A. Lewis wrote: p > > healyzh@NOaracnetSPAM.com writes in article <bjrgpc02agj@enews4.newsguy.com> dated 12 Sep 2003 04:07:08 GMT: > > N > >>I got the GnuPG 1.2.0 kit from HP and have finally gotten it running afterQ > >>rebuilding it from the source.  That is it will run from the SYSTEM account. XG > >>What privileges are needed for a user account to be able to run it?h > >  > > P > > The PGP I built back in the day did not require any special privileges.  AllM > > it does is read and write files and some heavy number crunching, right?  f > > K > > Are you getting an error when you run it under a normal user account?  s > > If so, post it.p > > / > > --Keith Lewis              klewis$mitre.orgaB > > The above may not (yet) represent the opinions of my employer. >  >  >    -- o John Santoss Evans Griffiths & Hart, Inc. 781-861-0670 ext 539   ------------------------------   Date: 12 Sep 2003 23:14:47 GMT From: healyzh@aracnet.comh, Subject: Re: Question on running GnuPG 1.2.0, Message-ID: <bjtk171167v@enews1.newsguy.com>  + Patrick Spinler <pspinler@yahoo.com> wrote:n
 > $ nopriv > $ gpgg" > gpg: can't lock memory: error -1* > %NONAME-E-NOMSG, Message number 00000002 > $ set proc/priv= PSWAPMT > $ gpgs) > gpg: Go ahead and type your message ...    That does the trick.  P > > The PGP I built back in the day did not require any special privileges.  AllM > > it does is read and write files and some heavy number crunching, right?  a  J Overall, PGP 2.6.3ia-multi06 seems preferable to GnuPG 1.2.0.  What reallyJ stinks about GnuPG (besides the above privilige problem) is the following  from the README:K "For encryption/decryption & making/verifying signatures, only 'Stream LF' N  format is supported on OpenVMS."J This means that you need to convert files to "Stream LF" before encrypting them.2  J OTOH, I suspect that GnuPG is more compatible with the rest of the world. @ I'm thinking that a combination of the two is the best solution.   			Zanet   ------------------------------    Date: 12 Sep 2003 14:53:29 -07001 From: byer@mail.ourservers.net (Robert Alan Byer)sA Subject: Strange DCL Problem:  DCL Can't Find My GOSUB Labels.... = Message-ID: <c4bfc78e.0309121353.722a0531@posting.google.com>7  F I have a rather large DCL script I've been working on for the past few monthsF (it's 175 blocks and growing).  It's been working for testing purposes untileF today when I added a big subroutine and now I get the following error.  > %DCL-W-USGOSUB, target of GOSUB not found - check spelling and presence of label   D It's the VERY first GOSUB in my DCL script.  I've checked everything and-A everything is spelled correctly and everything appears to be justnD fine, never had a problem with it during testing.  I removed that to go to the next one,2 and I get the same error.   E Is my DCL script too large?  If that's the case I've got some seriouse( reworking to do in order to split it up.   Any help would be appreciated.   ------------------------------  # Date: Fri, 12 Sep 2003 23:35:47 GMTo> From: Michael Austin <maustin@no-more-spam.firstdbasource.com>E Subject: Re: Strange DCL Problem:  DCL Can't Find My GOSUB Labels....h9 Message-ID: <nLs8b.260$9A2.22@newssvr23.news.prodigy.com>n   Robert Alan Byer wrote:P  H > I have a rather large DCL script I've been working on for the past few > monthsH > (it's 175 blocks and growing).  It's been working for testing purposes > until H > today when I added a big subroutine and now I get the following error. > @ > %DCL-W-USGOSUB, target of GOSUB not found - check spelling and > presence of label  > F > It's the VERY first GOSUB in my DCL script.  I've checked everything > and>C > everything is spelled correctly and everything appears to be just.F > fine, never had a problem with it during testing.  I removed that to > go to the next one,p > and I get the same error.. > G > Is my DCL script too large?  If that's the case I've got some seriousd* > reworking to do in order to split it up. >   > Any help would be appreciated.  ) did you remember to put colon on the end:r     $ gosub test1   !note NO colon $exit  $test1:   !Note the colone $ write sys$output "got here"i $ return  '$status'h     Michael Austin   ------------------------------    Date: 12 Sep 2003 13:54:20 -05004 From: kuhrt@nospammy.encompasserve.org (Marty Kuhrt) Subject: VMS and MP3ss3 Message-ID: <wf5yHbu5AkrR@eisner.encompasserve.org>m  ` In article <bjrc9p$avo0$1@news3.infoave.net>, "Jeff Morgan" <vmswiz@geonospamcities.com> writes:N > And then, of course I've ripped all my music CDs to MP3s (YES, legally!) andM > every pc in the house can play them off a pathworks fileshare using WindowsE > Media Player.e  4 Did you rip your CD's on the VMS box?  If so, how?    = Does anyone know of a handy mp3 streamer for VMS web servers?    ------------------------------  % Date: Sat, 13 Sep 2003 05:11:26 +0800 , From: Paul Repacholi <prep@prep.synonet.com>; Subject: Re: Why does MAIL DIR go slow during mail receive? - Message-ID: <878yotg1a9.fsf@prep.synonet.com>.  , bleau@umtof.umd.edu (Lawrence Bleau) writes:  F > I have a question about an observed slowness in VMS MAIL.  I think IE > know why, but want confirmation of my theory, or how to investigatea: > it further, as well as suggestions for a fix.  First, myD > configuration: I'm running OpenVMS AXP V7.1-2, TCPIP V5.1 ECO 4 on > my Alpha workstation.5 ....E > By noting the time carefully and checking the accounting file, I'verD > noticed that an email message came in at that approximate time.  IA > checked the TCPIP$SMTP_RECV_RUN.LOG files, and there's one thatA? > corresponds to that time.  The incoming email message was notsB > destined for this user, however, but some other user.  Still, it# > might be related, and here's how..   <chomp>   F The MAIL.MAI file is a standard RMS multi-key file. But, once the mailE delivery starts, it will be opened shared write by both processes and / they will be pinging locks backward and foward.w  B Now that should not slow you down to 1 minutes worth, but I'll betC that the MAIL.MAIs have never been re-built or purged. They do needbD standard care and feeding or they go into RMS sulk mode. If you haveE users who do LOTS of mailing, a suitable FDL for their mail files may0 be in order.  A As a first step, get them to try doing a PURGE/RECLAIM and see ifs) that gets things back to near reasonable.u  > (btw, on unix this would result in a mail delivery failure...)   --  < Paul Repacholi                               1 Crescent Rd.,7 +61 (08) 9257-1001                           Kalamunda.u@                                              West Australia 6076* comp.os.vms,- The Older, Grumpier Slashdot. Raw, Cooked or Well-done, it's all half baked.F EPIC, The Architecture of the future, always has been, always will be.   ------------------------------   End of INFO-VAX 2003.507 ************************