1 INFO-VAX	Mon, 24 Oct 2005	Volume 2005 : Issue 594       Contents:2 Re: "Jumbo Packets" and TCPIP Services for OpenVMS/ Re: HP : Massive strike and protest march today / Re: HP : Massive strike and protest march today , Is anyone porting openoffice to OpenVMS yet?0 Re: Is anyone porting openoffice to OpenVMS yet?0 Re: Is anyone porting openoffice to OpenVMS yet?* Re: Lack of response from HP field service LG Printer Advice  Re: LG Printer Advice   Re: LIB$WAIT & COBOL/VMS problem  Re: LIB$WAIT & COBOL/VMS problem  Re: LIB$WAIT & COBOL/VMS problem  Re: LIB$WAIT & COBOL/VMS problem  Re: LIB$WAIT & COBOL/VMS problem@ Re: Porting VAX/VMS to 8086 (Was: Re: Porting VMS back to VAX ?)@ Re: Porting VAX/VMS to 8086 (Was: Re: Porting VMS back to VAX ?) Re: SEARCH HTML fileM Re: Technical Update Days - Trip report Germany, Italy, Switzerland  (part 1) P Re: Technical Update Days - Trip report Germany, Italy, Switzerland  (part 1)  (P Re: Technical Update Days - Trip report Germany, Italy, Switzerland  (part 1)  (L Re: Technical Update Days - Trip report Germany, Italy, Switzerland (part 1)L Re: Technical Update Days - Trip report Germany, Italy, Switzerland (part 1)P Re: Technical Update Days - Trip report Germany, Italy, Switzerland (part 1) (pa2 Re: Trip Report part 2 France, Netherlands, SwedenA Re: UCX performance on VMS 6.2 - Unexpected rise in CPU usage....   F ----------------------------------------------------------------------  % Date: Mon, 24 Oct 2005 14:22:21 +0100 * From: "Richard Brodie" <R.Brodie@rl.ac.uk>; Subject: Re: "Jumbo Packets" and TCPIP Services for OpenVMS 2 Message-ID: <djin6d$bq9$1@blackmamba.itd.rl.ac.uk>  : "Jeff Goodwin" <jgoodwin@maine.rrr-r.com> wrote in message2 news:JIa6f.90387$7b6.62213@twister.nyroc.rr.com...  H > >> Q4. If I turn on jumbo frames, is OpenVMS smart enough to negotiateD > >> a smaller frame size when communicating to a peer that does not > >> support jumbo frames?  Q Yes. That's been true for a long while (ever since FDDI clusters were supported).    ------------------------------  % Date: Mon, 24 Oct 2005 18:50:49 +1300 $ From: "Lurker" <nowhere@nothing.com>8 Subject: Re: HP : Massive strike and protest march today3 Message-ID: <dA_6f.2176$S24.150786@news.xtra.co.nz>   8 "leslie" <LESLIE@JRLVAX.HOUSTON.RR.COM> wrote in message2 news:JiV6f.13939$5e4.12051@tornado.texas.rr.com...4 > Bill Gunshannon (bill@triangle.cs.uofs.edu) wrote: > : J > : Whagt does any of this have to do with unions other than to prove thatJ > : they offer no preotection at all (hint, the teachers around here stillF > : have very strong unions which they use to strangle the educational	 process).  > :  > I > Workers have no protection from corporations that now "own and operate"  > the government.   B Wrong. When you accept a job you sign a legally binding agreement.F It includes things like your salary, annual leave entitlement, minimumK time for termination notice (and conditions under which it can be ignored).   B Once that's done, both you and the company are bound by the terms.C But then - a union walks in and wants to change those terms. Hullo? B I have not signed up for the change and I fully appreciate that my employer hasn't either.    ------------------------------   Date: 24 Oct 2005 11:07:54 GMT( From: bill@cs.uofs.edu (Bill Gunshannon)8 Subject: Re: HP : Massive strike and protest march today+ Message-ID: <3s3tk9Fmd74fU1@individual.net>   3 In article <3L_6f.2178$S24.153352@news.xtra.co.nz>, ' 	"Lurker" <nowhere@nothing.com> writes: 0 > "Doc" <doc@openvms-rocks.com> wrote in message< > news:Xns96F8EE24087CBdocopenvmsrockscom@212.100.160.126... > K >> Which side you choose to strongly subscribe to may well be based on your  >> experiences.  >  > I guess that's true. > 3 >> Me?  I remember one or two colleagues suggesting I >> unionization and being ridiculed by the majority who thought they were + >> doing "quite well, thank you very much".  > C > They were I presume. At least, I'll take their word for that over  > anyone else's. > ! >>  No help when the company made K >> them all redundant.  Admittedly unlikely a union could have done much in  >> the situation > G > What could it do? I've been in a situation like that myself - a small E > software company hit financial difficulties and simply couldn't pay A > all of it's employees. A number of people were laid off. So how D > would a union help? Force all other developers to go on strike and, > sink the company? Much good that would do.   I And the ne time I was forced to belong to a union I was layed off so that G the company could hire a replacement at a lower hourly rate.  The union H did nothing because, "either one of them will be paying union dues so it4 doesn't matter to us which one the company employs."   bill      --  J Bill Gunshannon          |  de-moc-ra-cy (di mok' ra see) n.  Three wolvesD bill@cs.scranton.edu     |  and a sheep voting on what's for dinner. University of Scranton   |A Scranton, Pennsylvania   |         #include <std.disclaimer.h>       ------------------------------    Date: 24 Oct 2005 07:02:02 -0700 From: bob@instantwhip.com 5 Subject: Is anyone porting openoffice to OpenVMS yet? C Message-ID: <1130160898.844466.148070@g49g2000cwa.googlegroups.com>   8 I just emailed the group who has this and they said they: get requests for this but noone has really stepped forward- to work on the port for vms ... is this true?    ------------------------------  % Date: Mon, 24 Oct 2005 16:10:01 +0200 / From: "Johnny Willemsen" <jwillemsen@remedy.nl> 9 Subject: Re: Is anyone porting openoffice to OpenVMS yet? 6 Message-ID: <435ceb3c$0$11071$e4fe514c@news.xs4all.nl>   Hi,   " See http://www.oooovms.dyndns.org/   Johnny  ' <bob@instantwhip.com> wrote in message  = news:1130160898.844466.148070@g49g2000cwa.googlegroups.com... 9 >I just emailed the group who has this and they said they < > get requests for this but noone has really stepped forward/ > to work on the port for vms ... is this true?  >    ------------------------------  % Date: Mon, 24 Oct 2005 14:55:20 +0000 - From: David B Sneddon <dbsneddon@bigpond.com> 9 Subject: Re: Is anyone porting openoffice to OpenVMS yet? * Message-ID: <435CF5D8.2010603@bigpond.com>  ) bob@instantwhip.com was overheard to say: : > I just emailed the group who has this and they said they< > get requests for this but noone has really stepped forward/ > to work on the port for vms ... is this true?     Go for it Boob -- do it in Dibol   Regards, Dave.  --  D David B Sneddon (dbs)  VMS Systems Programmer  dbsneddon@bigpond.comD Sneddo's quick guide ...     http://www.users.bigpond.com/dbsneddon/D DBS freeware     http://www.users.bigpond.com/dbsneddon/software.htm   ------------------------------    Date: 24 Oct 2005 07:40:59 -0700$ From: "Ed Wilts" <ewilts@ewilts.org>3 Subject: Re: Lack of response from HP field service C Message-ID: <1130164859.619799.178190@o13g2000cwo.googlegroups.com>    Bill Gunshannon wrote:  E > Is the power supply hot swappable?  If so, wouldn't it be easier to 0 > just buy one from Island and swap it yourself?  E For an Alphaserver 4000, I don't think so.  I've still got a bunch of C 4100s here and they're definitely not.  They're not that hard to do ; yourself, but they're certainly not as easy as say an ES40.   	    .../Ed    mailto:ewilts@ewilts.org   ------------------------------  % Date: Mon, 24 Oct 2005 10:57:07 -0500 # From: "Dan Moore" <dmoore@sosu.edu>  Subject: LG Printer Advice+ Message-ID: <nv77f.12243$xk2.8551@fe06.lga>   
 Greetings,  J      I recently received a notice from HP that our remaining LGXX (serial H impact) printers will be end-of-life soon. We have already replaced our H smaller printers (LAXX, e.g.) with laser printers, but I'm wondering if I anyone has advice for dealing with the higher capacity printers yet.  We  L have three left (LG02, LG05, LG15 plus), and I'm considering high end laser I printers as a replacement. Another option is to find a different support  % channel for our existing LG printers.   7    Any thoughts, ideas, or experiences are appreciated.   
    Thanks,  	 Dan Moore    ------------------------------  % Date: Mon, 24 Oct 2005 12:19:35 -0400 , From: Carl Friedberg <frida.fried@gmail.com> Subject: Re: LG Printer AdviceI Message-ID: <890539d90510240919g11671e91l702731696e0d7efc@mail.gmail.com>    Hi Dan,   D We've had excellent results using HP printers for more than 10 years< now. With DCPS now bundled with VMS, you get a comprehensiveD printing solution for reasonable prices. The cost of service vs. theC cost of a replacement printer is something else you might consider.   = Near the top end, the HP LJ 9000tn as been superb, as has the @ older 4100dn and even holder 4000. All offer postscript support,G duplexing if you need it, and great reliability and print quality. DCPS F support is seamless, as is TCPIP support via Multinet (I'm reasonsably@ sure that applies to TCPware, and possibly UCX/TCPIP services as. well, but I only know for sure with Multinet.)   Best,    Carl Friedberg www.comets.com  / On 10/24/05, Dan Moore <dmoore@sosu.edu> wrote:  > Greetings, > K >      I recently received a notice from HP that our remaining LGXX (serial I > impact) printers will be end-of-life soon. We have already replaced our I > smaller printers (LAXX, e.g.) with laser printers, but I'm wondering if J > anyone has advice for dealing with the higher capacity printers yet.  WeL > have three left (LG02, LG05, LG15 plus), and I'm considering high end las= erJ > printers as a replacement. Another option is to find a different support' > channel for our existing LG printers.  > 9 >    Any thoughts, ideas, or experiences are appreciated.  >  >    Thanks, >  > Dan Moore  >  >  >  >  >  >  >  >    ------------------------------    Date: 24 Oct 2005 08:18:34 -0500 From: briggs@encompasserve.org) Subject: Re: LIB$WAIT & COBOL/VMS problem 3 Message-ID: <Xeexgtk9kYXy@eisner.encompasserve.org>   g In article <djhhj7$7ap$1@news-02.connect.com.au>, "Richard Maher" <maher_rj@hotspamnotmail.com> writes:  > Hi Larry,  > H >> So this was fundamentally something that would be corrected by a more >> strongly typed language.  > M > The flight-simulator usually picks up a lot of 'em as well :-) As someone's J > posting tag at DEC once read "Week-typing is for Strong stomachs!". Make > mine a Prawn Phal! >  >>( >> Ada package LIB defines LIB$WAIT with >> >> SECONDS : in FLOAT  >>: >> Pascal module PASCAL$LIB_ROUTINES defines LIB$WAIT with >> >> seconds : SINGLE  > N > Good to see they're kept up to date with new VMS functionality then. (or did- > you just not include the other parameters?)  > H > BTW. What happens in ADA, Pascal etc if a routine wants an argument BYG > DESCRIPTOR but you'd like to build your own descriptor and pass it by K > reference? (You have a User Defined Datatype of  "Descriptor" I suppose?) E > And if you dynamiccaly allocate memory it's how you MAP it I guess? 2 > Procedure calling standard Arg-Lists on the fly?  @ One would typically do it statically.  Define two interfaces to,A for instance, LIB$PUT_OUTPUT, one using a string parameter passed = by descriptor and one using a descriptor passed by reference.   D Building a descriptor is straightforward.  You could build it out ofE ordinary storage on the stack or you could allocate it from the heap.   + 	status = my.LIB$PUT_OUTPUT ( "hi there" );  	...>        	status = my.LIB$PUT_OUTPUT ( some_string_descriptor );  = The compiler can disambiguate this "overloaded" definition of > LIB$PUT_OUTPUT by looking at the data type of the parameter(s)B and of the function result and matching it up with a corresponding= interface definition.  So if you've defined LIB$PUT_OUTPUT in > package "my" as taking a string and returning a condition codeA then the first reference will use that definition.  And if you've > also defined LIB$PUT_OUTPUT in package "my" as taking a stringC descriptor and returning a condition code then the second reference ! will use that definition instead.   D If the overloading resolution cannot be made statically the compilerA will complain and you may have to use explicit type casts to nail C things down so that exactly one overloaded subroutine profile fits.   A Typically this becomes an issue with literals and nested function E calls where the data type of the literal or of the function result is  not always statically knowable.    ------------------------------  # Date: Mon, 24 Oct 2005 14:12:07 GMT & From: John Reagan <john.reagan@hp.com>) Subject: Re: LIB$WAIT & COBOL/VMS problem 4 Message-ID: <XY57f.15184$6v1.12629@news.cpqcorp.net>   Richard Maher wrote:  H > BTW. What happens in ADA, Pascal etc if a routine wants an argument BYG > DESCRIPTOR but you'd like to build your own descriptor and pass it by K > reference? (You have a User Defined Datatype of  "Descriptor" I suppose?) E > And if you dynamiccaly allocate memory it's how you MAP it I guess? 2 > Procedure calling standard Arg-Lists on the fly? >   I The Ada and Pascal compilers know how to automagically build descriptors  I for string arguments.  Pascal also has foreign mechanism specifiers like  H %DESCR and %STDESCR that you can place in front of any actual parameter H and the compiler will try to build the descriptor for you.  In 25 years F of programming with VAX/DEC/Compaq/HP Pascal, I've built a descriptor ? "by hand" once (and that for a test in the Pascal test system).      --   John Reagan / HP Pascal/{A|I}MACRO for OpenVMS Project Leader  Hewlett-Packard Company    ------------------------------  % Date: Mon, 24 Oct 2005 15:39:46 +0200 & From: Paul Sture <paul.sture@decus.ch>) Subject: Re: LIB$WAIT & COBOL/VMS problem + Message-ID: <3s46h3Fm9kp5U1@individual.net>    AlexB wrote:7 > On Sat, 22 Oct 2005 21:52:53 +0200, Paul Sture wrote:  >  > G >>Maybe you transcribed the above incorrectly, but libnowake should be   >>declared as a longword:  >>( >>01 libnowake     pic 9(5)comp value 1. >  > ? > WOOO HOOO!!!!! That was it!!  You are awesome!!  Thanks Paul.  >    You are welcome. ;.)   ------------------------------    Date: 24 Oct 2005 09:16:42 -0500- From: Kilgallen@SpamCop.net (Larry Kilgallen) ) Subject: Re: LIB$WAIT & COBOL/VMS problem 3 Message-ID: <CUgxM7U87Oio@eisner.encompasserve.org>   c In article <pan.2005.10.24.02.41.19.392888@yahoo.com>, AlexB <BartonekDragRacing@yahoo.com> writes:   F > it wasnt the seconds in question...it was the way I had parameter #2A > declared..the libnowake parameter..   As for using a 'stronger' " > language..do you mean 'C' ?  ;)   7 Of course not :-)   I guess you are new around here :-)    Larry Kilgallen    ------------------------------    Date: 24 Oct 2005 09:26:06 -0500- From: Kilgallen@SpamCop.net (Larry Kilgallen) ) Subject: Re: LIB$WAIT & COBOL/VMS problem 3 Message-ID: <Pbz4qZphhsjB@eisner.encompasserve.org>   g In article <djhhj7$7ap$1@news-02.connect.com.au>, "Richard Maher" <maher_rj@hotspamnotmail.com> writes:   ( >> Ada package LIB defines LIB$WAIT with >> >> SECONDS : in FLOAT  >>: >> Pascal module PASCAL$LIB_ROUTINES defines LIB$WAIT with >> >> seconds : SINGLE  > N > Good to see they're kept up to date with new VMS functionality then. (or did- > you just not include the other parameters?)   K I use older compilers so LJK/Security can run on VMS versions back to V4.2. J It would not do to depend on a parameter that has no effect on some of the0 VMS versions on which the software will be used.  H > BTW. What happens in ADA, Pascal etc if a routine wants an argument BYG > DESCRIPTOR but you'd like to build your own descriptor and pass it by K > reference? (You have a User Defined Datatype of  "Descriptor" I suppose?)   D There are escapes for that, but there is typically no reason to ever+ construct a descriptor by hand.  One codes:   # 	XT.XtMenuPopup ( "my Menu name" );   7 Construction of descriptors is handled by the compiler.   E > And if you dynamiccaly allocate memory it's how you MAP it I guess?    I don't understand that part.   2 > Procedure calling standard Arg-Lists on the fly?  A The standard argument lists for each routine are specified in the 0 definition of the package (for LIB$, STR$, etc.)   ------------------------------  # Date: Mon, 24 Oct 2005 14:19:50 GMT & From: John Reagan <john.reagan@hp.com>I Subject: Re: Porting VAX/VMS to 8086 (Was: Re: Porting VMS back to VAX ?) 3 Message-ID: <a467f.15185$FE1.5928@news.cpqcorp.net>    JF Mezei wrote:   J > And no, he was not pulling my leg. Consider how vehemently VMS engineersE > have been denying the port to the 8086. They know it is going on in F > ZKO's basement but have to kill any rumours of it happening until HPH > decides they are ready to annnounce it. There is even a topic about it > in  "Ask the Wizzard".  I OK, I'll spill the beans.  I'll tell you exactly what is going on in the   basement of ZKO.   Ssssshhhh, be very quiet...   ? There is a very large room full of ... CHAIRS!  One very large   conference room.  D Now, down the hall is where the real work is taking place.  We have I another large room full of ... EXERCISE EQUIPMENT!  I spend 45mins a day  I down there on either the eliptical trainer or recumbant bicycle.  Others  H use the weight machines and tread mills.  There is also a room used for ! yoga, step aerobics, and pilates.   F OK, I feel better now that I've gotten that huge secret off my chest. 3 Please don't hate me for giving in to the pressure.    --   John Reagan D HP Pascal/{A|I}MACRO for OpenVMS Project Leader, oh yeah, COBOL also Hewlett-Packard Company    ------------------------------    Date: 24 Oct 2005 09:30:08 -0500- From: Kilgallen@SpamCop.net (Larry Kilgallen) I Subject: Re: Porting VAX/VMS to 8086 (Was: Re: Porting VMS back to VAX ?) 3 Message-ID: <EQU9WgtwBT10@eisner.encompasserve.org>   \ In article <a467f.15185$FE1.5928@news.cpqcorp.net>, John Reagan <john.reagan@hp.com> writes:  K > OK, I'll spill the beans.  I'll tell you exactly what is going on in the   > basement of ZKO.  F > Now, down the hall is where the real work is taking place.  We have K > another large room full of ... EXERCISE EQUIPMENT!  I spend 45mins a day  K > down there on either the eliptical trainer or recumbant bicycle.  Others  J > use the weight machines and tread mills.  There is also a room used for # > yoga, step aerobics, and pilates.   B Ok, now that the secret is out, I will tell the world that John isA spending his time there when he should be getting his exercise by A going out and contra-dancing, at which time he could get involved ? in the politics of all the suburban outside agitators trying to A keep the City of Cambridge from turning the sprung dance floor at  the VFW post into office space.    ------------------------------    Date: 24 Oct 2005 07:55:12 -0500; From: koehler@eisner.nospam.encompasserve.org (Bob Koehler)  Subject: Re: SEARCH HTML file 3 Message-ID: <tFFTWWP+6U+N@eisner.encompasserve.org>   N In article <opsy1vdpgczgicya@hyrrokkin>, "Tom Linden" <tom@kednos.com> writes:L > Many HTML files have no carriage returns so they appear to VMS as a file   > withJ > a record length greater than 32768, which causes SEARCH to give up, e.g.  H    Actually, they probably do have some sort of end-of-line marker.  YouB    just need to tell RMS what, since it has been told incorrectly.         "set file/attributes"       is your friend.   ------------------------------  % Date: Mon, 24 Oct 2005 09:45:53 +0200 7 From: "Guy Peleg" <guy.peleg@remove_this_header@hp.com> V Subject: Re: Technical Update Days - Trip report Germany, Italy, Switzerland  (part 1)* Message-ID: <435c9136@usenet01.boi.hp.com>  : "JF Mezei" <jfmezei.spamnot@teksavvy.com> wrote in message& news:435B171E.29D1D26F@teksavvy.com...$ > susan_skonetski@hotmail.com wrote: > > D > > I think the Guido was pretty much between us until this message. > H > Now Susan, look at what you've done... you've given the very respectedF > Mr Peleg a new nickname that will haunt him for the remainder of his > life :-) ;-)  A I enjoyed Italy so much that I do not mind to be called Guido.... I it has it's charm....maybe I'll get lucky...;-)....infact I'm considering ' applying for an Italian citizenship....   6  Perhaps Mr Peleg could chime in here and divulge what > your nickname is :-) :-)  9 A wise woman once told me...what's goes on the road stays 7 on the road....so I won't be violating this rule....;-)    >  > D > > The customer visits are set up via the local Ambassadors who are& > > involved with the local customers. >  > J > Are these generally just "sightseeing" trips, or do the customers have aJ > specific agenda they wish to discuss with VMS management ? Or do the VMSG > ambassadors tell you that the customer is thinking of leaving VMS and I > that your visit would ve very appreciate to help maintain the account ?  > H > (since you didn't mention the names of the customers you visited, your< > answer couldn't be associated with any specific customer). >  >  > > hope this helps  >  > E > Sue, every message you post is appreciated, even if some of us will  > argue with what you post.    ------------------------------  # Date: Mon, 24 Oct 2005 14:26:25 GMT & From: John Reagan <john.reagan@hp.com>Y Subject: Re: Technical Update Days - Trip report Germany, Italy, Switzerland  (part 1)  ( 3 Message-ID: <la67f.15186$FE1.8491@news.cpqcorp.net>    JF Mezei wrote: $ > susan_skonetski@hotmail.com wrote: > > >>Subject: Technical Update Days - Trip report Germany, Italy, >  > H > And you expect us to NOT be jealous of your HP paid tourist adventures > ?????? :-) :-) :-) >   I If your jealous of this trip, wait until she writes the trip reports for  7 our next adventure to Australia, Singapore, and Taiwan.      --   John Reagan E HP Pascal/{A|I}MACRO for OpenVMS Project Leader, now with COBOL, yum!  Hewlett-Packard Company    ------------------------------  % Date: Mon, 24 Oct 2005 08:37:34 -0700 4 From: Alan Frisbie <Usenet02_REMOVE@Flying-Disk.com>Y Subject: Re: Technical Update Days - Trip report Germany, Italy, Switzerland  (part 1)  ( % Message-ID: <1130168093.810945@smirk>    John Reagan wrote:  M > If you're jealous of this trip, wait until she writes the trip reports for  9 > our next adventure to Australia, Singapore, and Taiwan.   9 I can't wait to read Sue's reaction to meeting kangaroos, ; wombats and koalas.   Hopefully she won't have an encounter 7 with a drop bear.   That would be a shame.   Be sure to ! warn her about how to avoid them.    Alan   ------------------------------    Date: 24 Oct 2005 09:03:56 -0700! From: susan_skonetski@hotmail.com U Subject: Re: Technical Update Days - Trip report Germany, Italy, Switzerland (part 1) C Message-ID: <1130169836.204892.211600@g14g2000cwa.googlegroups.com>   A and India you forgot India, which I am really looking forward to.    Whats a drop bear?   Sue    John Reagan wrote: > JF Mezei wrote: & > > susan_skonetski@hotmail.com wrote: > > @ > >>Subject: Technical Update Days - Trip report Germany, Italy, > >  > > J > > And you expect us to NOT be jealous of your HP paid tourist adventures > > ?????? :-) :-) :-) > >  > J > If your jealous of this trip, wait until she writes the trip reports for9 > our next adventure to Australia, Singapore, and Taiwan.  >  >  > --
 > John Reagan G > HP Pascal/{A|I}MACRO for OpenVMS Project Leader, now with COBOL, yum!  > Hewlett-Packard Company    ------------------------------  + Date: Mon, 24 Oct 2005 17:07:08 +0000 (UTC) < From: gartmann@non.immunbio.mpg.de.sens (Christoph Gartmann)U Subject: Re: Technical Update Days - Trip report Germany, Italy, Switzerland (part 1) ) Message-ID: <djj4bs$kbi$2@news.BelWue.DE>   g In article <1130169836.204892.211600@g14g2000cwa.googlegroups.com>, susan_skonetski@hotmail.com writes: B >and India you forgot India, which I am really looking forward to. >  >Whats a drop bear?   # Have a look at en.wikipedia.org ;-)    Regards,    Christoph Gartmann    --  E  Max-Planck-Institut fuer      Phone   : +49-761-5108-464   Fax: -452   ImmunbiologieI  Postfach 1169                 Internet: gartmann@immunbio dot mpg dot de   D-79011  Freiburg, Germany 9                http://www.immunbio.mpg.de/home/menue.html    ------------------------------  % Date: Mon, 24 Oct 2005 05:29:24 -0400 - From: JF Mezei <jfmezei.spamnot@teksavvy.com> Y Subject: Re: Technical Update Days - Trip report Germany, Italy, Switzerland (part 1) (pa , Message-ID: <435CA962.A83F819D@teksavvy.com>   Guy Peleg wrote:C > I enjoyed Italy so much that I do not mind to be called Guido.... K > it has it's charm....maybe I'll get lucky...;-)....infact I'm considering ) > applying for an Italian citizenship....   F BTW "Guy" is a french name. It is pronounced like Geek without the k.     E I take it the next qualifiers to be added to DCL will have an italian  flair to them ???? ;-) ;-) :-)  H I want: SET CPU/EXPRESSO  to set my all mighty microvax II into its 3ghz clock speed mode :-)G         SET CPU/NOCALIENTE to set IA64 machines into energy saving mode  to generate less heat  	    		  ; > A wise woman once told me...what's goes on the road stays 9 > on the road....so I won't be violating this rule....;-)   G Humm... isn't that like those Las Vegas commercials that insinuate that F any sinful things you do in Las Vegas stay in las vegas ?????? :-) :-)    H So, you are implicitely, by lack of denial, admitting that Sue does have a nickname ?   ------------------------------  % Date: Mon, 24 Oct 2005 07:49:33 +0200 & From: Paul Sture <paul.sture@decus.ch>; Subject: Re: Trip Report part 2 France, Netherlands, Sweden + Message-ID: <3s3aveFm5ksnU1@individual.net>   " susan_skonetski@hotmail.com wrote: > -----Original Message----- > From: Skonetski, Susan( > Sent: Sunday, October 23, 2005 7:16 PM > To: Skonetski, Susan9 > Subject: Trip Report part 2 France, Netherlands, Sweden  >   . Thanks Sue. I thoroughly enjoyed both reports.   ------------------------------    Date: 24 Oct 2005 02:02:18 -0700( From: "Lawrie" <stroker_ace@hotmail.com>J Subject: Re: UCX performance on VMS 6.2 - Unexpected rise in CPU usage....C Message-ID: <1130144538.611189.292320@g49g2000cwa.googlegroups.com>    Dave Weatherall wrote:F > On Fri, 21 Oct 2005 08:01:40 UTC, "Lawrie" <stroker_ace@hotmail.com> > wrote: > K > > I forgot to add that my TCP/IP implementation uses non blocking sockets  > > as well. > >  > G > Does this mean you're polling to do your synchronisation now, whereas C > previously you might have been using AST's and/or event flags? We D > certainly achieved an improvement by switching our (simple) socketB > implimentation from non-blocking sockets with 'poll' to blocking > socket access. >  > -- > Cheers - Dave W.  D Yes, I am using the select funtion to check readiness to write to my sockets.  A The original implementation using QIOs and Serial comms was fully 2 blocking. This caused problems at high data rates.  C The process in question does a blocking read on an input PPL queue. E Once a message is received the Serial version then did a blocking QIO  write to the Serial line.   D At high data rates the 19.2K serial line could not keep up with dataA rates and would cause data to back up in the PPL queue eventually   causing the application to fail.  D I needed a simple method of decoupling the read and write operations: and I was stuck with the blocking read from the PPL queue,  E My solution was to use non blocking TCP/IP sockets and 10Mb ethernet.   E When the blocking read on the PPL queue returns the message is placed A into a ring buffer (one is allocated per TCP/IP socket instance).   E The select function is used (with a timeout of 1/1000 of a second) to E check what sockets are free to write data. An attempt is then made to ! write the message to each socket.   D If the number of sent bytes is as expected (i.e the message has beenD completely sent) the message is removed from the ring buffer and the@ ringbuffer is checked to see if it contains any oustanding data.  D If there is then an attempt is made to send it. This process repeatsD until the ring buffer is empty or a write attempt to the socket does& not send the expected number of bytes.  C Using this approach even if (for some reason) I cannot write to the E TCP/IP sockets all I do is overwrite data in the ring buffer and data " does not back up in the PPL queue.   Cheers   Laurie   ------------------------------   End of INFO-VAX 2005.594 ************************