1 INFO-VAX	Sun, 05 Nov 2006	Volume 2006 : Issue 608       Contents:$ Re: DECW Session Manager and MOZILLA Re: Flight 3.1 on VMS 8.3 ? 3 Re: Itanium model numbers {re F$getsyi("HW_MODEL")} * Re: Java Applet pulled from VMS Web Server* Re: Java Applet pulled from VMS Web Server- Re: Process Software OpenVMS Security Webinar  Re: TCPIP Services SPAM config Re: Time change questions !  Re: Time change questions !  Re: Time change questions !   F ----------------------------------------------------------------------  # Date: Sun, 05 Nov 2006 04:28:49 GMT   From: John Santos <john@egh.com>- Subject: Re: DECW Session Manager and MOZILLA ' Message-ID: <5Md3h.279$244.99@trnddc01>    JF Mezei wrote: K > I've experienced a problem documented in some release notes for Mozilla.  K > They discuss a problem where when sending a message, it fails because of  K > a Temporary Directory is screwed up. The notes mention that one needs to  * > undefine the logicals TMP and/or TMPDIR. >  > Here is the catch: > G > I was able before to send messages when Mozilla was started from the  / > session manager, and then it stopped working.  > H > While I cannot send message from session manager initiated Mozilla, I M > can still get an instance of Mozilla that works when invoked interactively.  > D > The session manager's "Session->Logical Names"  says there are no # > definitions for "TMP" or "TMPDIR"  > H > Is it possible that some application created a logical for TMP and/or H > TMPDIR and that process survived and is being used to start Mozilla ? E > (aka: one application polluting the process logical table and thus  4 > affecting the next app that re-uses that process ? > E > Short of updating the MOZILLA.COM to do a SHOW LOG TMP* , is there  B > another way to find out if those logicals would be defined when L > decw$session starts a process with the SYS$COMMON:[CSWB]MOZILLA.COM file ? > @ > So far, I have to say that i am not overly impressed with the F > cleanliness of the move from my old rock solid reliable VAX to this ! > newphangled thing called alpha.   E I've encountered this.  My company's applications have used a logical E name TMP (in the group table or in application-specific tables on the D development systems) since the late 1970's.  It's extremely annoyingE that some newcomer like Mozilla should disrupt existing applications. E Fortunately, Mozilla seems to only look at the value of TMP (not sure C about TMPDIR) when it starts up, so I can remove TMP from the group , table, start Mozilla, and then redefine TMP.  B When the problem occurs, the symptom is pretty obvious.  Trying toC send mail or create a news post causes an error message to display, A which includes a fairly bogus looking file name that includes the A translation of TMP in the middle of the directory path.  From the E looks of things, it doesn't appear that any reasonable value of "TMP" D would cause anything useful to happen in Mozilla, so the whole thingE is pretty stupid.  (For example, defining TMP to point to a different E disk or a scratch directory on a different disk doesn't cause Mozilla C work files to be created on the other disk, it just breaks posting. 	 Useless.)   E The C RTL seems pretty confused about the distinction between logical = names and DCL symbols, so it is possible having TMP or TMPDIR  symbols may also break Mozilla.    --   John Santos  Evans Griffiths & Hart, Inc. 781-861-0670 ext 539   ------------------------------  % Date: Sat, 04 Nov 2006 14:47:26 -0500 - From: JF Mezei <jfmezei.spamnot@teksavvy.com> $ Subject: Re: Flight 3.1 on VMS 8.3 ?8 Message-ID: <96a57$454cee08$cef8887a$29144@TEKSAVVY.COM>  = Patrick MOREAU, DTI Athis ex CENA, Tel: 01.69.57.68.40 wrote: Q > JF, is your display 8bit Pseudocolor ? Flight doesn't like a Truecolor display.     K Ahh, this may be it. I am in Truecolor.  And looking at xdpuinfo, it seems  L it suppports TrueColor and DirectColor only. I recall there is something to L setup in the private server setup to make it support variour colour schemes.  I If the server support Pseudocolors (but is on trucolor by default), will  
 Flight work ?   E I was under the impression that TrueColor was automatically going to   support pseudocolor.  H Or does the fact that I run the display as 24 bits preclude Flight from . ever working again ? That would be a sad loss.   ------------------------------  % Date: Sat, 04 Nov 2006 15:58:19 -0500 ' From: Dave Froble <davef@tsoft-inc.com> < Subject: Re: Itanium model numbers {re F$getsyi("HW_MODEL")}9 Message-ID: <tYidnbz1g8C6YtHYnZ2dnUVZ_qudnZ2d@libcom.com>     VAXman- @SendSpamHere.ORG wrote: > In article <rdeininger-0411060934450001@dialup-4.233.149.4.dial1.manchester1.level3.net>, rdeininger@mindspringdot.com (Robert Deininger) writes:  >>< >> In article <eig69c$2blo$1@pyrite.mv.net>, Stephen Hoffman, >> <Hoff@HoffmanLabs-RemoveThis-.Org> wrote: >>$ >>> VAXman- @SendSpamHere.ORG wrote: >>> G >>>> Are you saying that my "HP rx2600  (1.50GHz/6.0MB)" returned from   >>>>1 >>>> $ WRITE SYS$OUTPUT "''F$getsyi("HW_NAME")'"   >>>>G >>>> could change from release to release of VMS?  I know the amount of = >>>> memory show could but that's not dependent upon OS rev.  K >>>   The model string is embedded in the platform itself -- in the box --  M >>> by whomever built the box.  OpenVMS I64 fetches the string, and displays  O >>> it.  (This much like the changes made back around V6.2 for SCSI devices...)  >>> F >>>   What the particular platform vendor might decide to do with the H >>> string is, well, another discussion -- that string arises from what J >>> amounts to the system vendor platform firmware, and it's analogous to K >>> how SRM stored and returned that value to OpenVMS Alpha as part of the  M >>> dynamic system recognition (DSR) stuff.  Though unlike SRM, the platform  E >>> doesn't have a model number value, and (AFAIK) nothing precludes  J >>> multiple vendors from selecting the same string.  Whether or not that M >>> collision is remotely possible or would ever occur is another discussion. K >> HP's Integrity systems do have a binary identifier known as the "product J >> ID", which the firmware makes available and which VMS doesn't currentlyF >> make visible to the user.  I don't really know why VMS doesn't takeK >> advantage of the product ID; it is a close analog of the Alpha SMM.  The M >> Alpha SMM has much finer granularity, and I think the registration process ' >> was perhaps more tightly controlled.  >>G >> The PA-RISC analog of the product ID has a bit of a bad reputation.  M >> Apparently there were some registration screw-ups and numbers were re-used J >> inappropriately.  HP-UX folks are shy of trusting the Integrity ProductL >> ID, instead relying on the model string to guide their platform-dependentF >> code. (They have LOTS compared to VMS.)  Relying on an ASCII stringL >> controlled by marketeers, who capriciously rename products at the drop of" >> a hat, causes periodic trouble. >  > Hi Robert....  > I > Both yours and Hoff's information are great but really haven't answered K > by question; that bein, will these HW Model names remain consistent?  If, I > for example, I update the Itanium's F/W, what are the chances of seeing & > a different product naming scheme?   >   @ Actually Brian, I think they did answer your question, but in a H backhanded manner.  If the data 'could' be changed, and there is no one I to control such, then you really cannot count on it not changing.  About  G the best they could have said was that it's highly unlikely to change.  E For me at least, that's not good enough.  I learned long ago, when a  D user says "that never happens", you're assured to see 'that' happen.   --  4 David Froble                       Tel: 724-529-0450> Dave Froble Enterprises, Inc.      E-Mail: davef@tsoft-inc.com DFE Ultralights, Inc.  170 Grimplin Road  Vanderbilt, PA  15486    ------------------------------  # Date: Sat, 04 Nov 2006 22:42:00 GMT + From: =?ISO-8859-1?Q?Jan-Erik_S=F6derholm?= 3 Subject: Re: Java Applet pulled from VMS Web Server 2 Message-ID: <YG83h.21602$E02.8886@newsb.telia.net>   Richard Maher wrote :   9 > Having said that and given the very basic nature of the 6  > HTTP protocol, I will look at writing a slim-lined,?  > high-performance applet uploader that talks Restaurant-http.    No need to re-invent anything.* The OSU server is just that (if you want).  5 Now, the Apache might be another issue, I don't know.   	 Jan-Erik.    ------------------------------  $ Date: Sun, 5 Nov 2006 10:26:36 +08003 From: "Richard Maher" <maher_rj@hotspamnotmail.com> 3 Subject: Re: Java Applet pulled from VMS Web Server 1 Message-ID: <eijhta$g7b$1@news-02.connect.com.au>    Hi Jan-Erik,  H Before I respond directly to your post can you just tell me if I've doneG enough for you to give this Java Sockets stuff a go? Maybe not for your G current project, but at least a proof-of-concept for something down the F track? Did you look at Bojan's example? Can it really get much easier?    > No need to re-invent anything., > The OSU server is just that (if you want). > 7 > Now, the Apache might be another issue, I don't know.   = "Listen svery carefully; I shall zay ziss only 100 times :-)"   I I don't really care what you *choose* to upload your applets with! If you L are *already* running a full-blown, bloated and flatulent  web-server on theI *same* VMS server that will be hosting your Java Applets then please feel H free to use a HTTP URL to locate your JAR files, and everyone's a winnerJ chickens for dinner! But if you, and Arne, could open your minds up just aJ little bit to the following scenario, I think you'll find that it's not as uncommon as you may imagine.  G I submit it as conjecture that the percentage of companies in the world L today that are using VMS to host their web-sites is minimal to non-existant.I Even with only those companies that have VMS in their server stables, the K percentage will not reach double figures. Now these companies still want to L access their wealth of resources on their VMS servers and are sick and tiredJ of having to run data-mining bullshit to replicate (usually ftp) their VMSI riches to a UNIX box just to make it available to the UNIX/NT web-servers  and other applications.   I What I have dared to do is suggest to these mainstream sites that, rather K than configure and run a resource hungry Albatros on VMS, when all you want D to do is upload some small JAR files, they consider FTP as an optionC instead. Bojan Nemec has gone further and provided a template for a G lightweight  (22 DCL lines) http server. Now, Bojan's demonstration was H obviously not intended for production use but I'm willing to wager that,I even with the overhead of INETd having to fire up a VMS process for every H file request, Bojan's code will perform competitively when compared withK some of the web servers available on VMS today. I don't know what motivates I other people, but me and Bojan are just trying to crack-a-nut and get our E applets upto the PC browser as quickly, directly, and effortlessly as 	 possible.   I I've had a quick look at the http protocol RFC and can imagine a HTTP/1.0 D AST threaded server listining on port 80 (and only responding to GETC methods) to be very easy to implement and, more importantly, highly H efficient. Even if the Class and Archives have to be in seperate files IH think what we'll lose on the 2-connections-reqd swings we'll gain on theK code-path round-about. (BTW. Are all the web-servers available on VMS today J HTTP/1.1 compatible? If not then switch to one that is! As with Tier3, youE *want* connection-oriented! But let your mind ponder for a minute the G additional code reqd to conform to HTTP/1.1 reather than just 1.0) With L Applets it's a one-off hit, up front, per browser activation. That's a hit IL think is worth taking. Good luck with your Jack-of-all-trades Master-of-noneL Web-server. But I'm sure that, after it's peeled the potatoes and toned your+ belly, it folds up neatly under the bed :-)   D I am actually interested in the design and peformance of the variousK Web-Servers on VMS today but I just don't have the time or inclination. Me? K I'm all for pluralism and maximising your options. Life's all about choices C don't you think? Whatever floats a customer's boat is fine with me! J Unfortunately, there are many unscrupulous consultancy companies out thereB that have a vested interest in convincing the VMS client base thatH Web-Servers are not just for the web but they are in fact the middlewareF backbone of choice for all inter-application communication :-( They'reJ usually seen building their CVs and scurrying down the anchor-chain beforeJ the good-ship VMS goes under. But then all you guys still get your jolliesL out of looking at the world thru XML glasses! So why am I wasting my breath?
 I give up :-(   : I'm off to get my weather SOAP service to talk to my GreekH language-transation service and tie it all in with the News from Bogota.G That's *REAL* computing! All this debits/credits business stuff is just  soooo passe!   Cheers Richard Maher  D "Jan-Erik Sderholm" <jan-erik.soderholm@telia.com> wrote in message, news:YG83h.21602$E02.8886@newsb.telia.net... > Richard Maher wrote :  > ; > > Having said that and given the very basic nature of the 8 >  > HTTP protocol, I will look at writing a slim-lined,A >  > high-performance applet uploader that talks Restaurant-http.  >   > No need to re-invent anything., > The OSU server is just that (if you want). > 7 > Now, the Apache might be another issue, I don't know.  >  > Jan-Erik.    ------------------------------  $ Date: Sat, 4 Nov 2006 17:02:13 -0500) From: "Neil Rieck" <n.rieck@sympatico.ca> 6 Subject: Re: Process Software OpenVMS Security Webinar< Message-ID: <454d0d64$0$14803$9a6e19ea@news.newshosting.com>  = "Paul Sture" <paul.sture.nospam@hispeed.ch> wrote in message  D news:paul.sture.nospam-9AA426.11105004112006@mac.sture.homeip.net...E > In article <1162471238.574615.170490@h48g2000cwc.googlegroups.com>, # > "Ian Miller" <ijm@uk2.net> wrote:  > H >> Unfourtunally 8pm GMT on a Friday I have other things to do. I expectH >> the webinar presentation will appear on their web site sometime after' >> the event and I can look at it then.  > G > If I heard correctly, they announced at the end that it will still be  > available online for a month.  > H > I don't know if it was an internet glitch, but the presentation slides? > were out of of sync with the speech, moving to the next slide H > prematurely enough to be irritating. Just in case that happens to you,H > I'd recommend downloading the PDF so that you can scroll through it in > sync with the speech.  > @ > For the benefit of OS X users, it worked fine with Firefox and
 > RealPlayer.  >  > --   > Paul Sture  D I attended that webinar and it is nice to see that someone is still I developing new products for OpenVMS. (One of my colleagues is working on  L something with ACME on OpenVMS but I'm wondering if we shouldn't first look L into PSC's new VMS Authentication Module (VAM). No sense in reinventing the  wheel)  
 Neil Rieck Kitchener/Waterloo/Cambridge,  Ontario, Canada." http://www3.sympatico.ca/n.rieck/    ------------------------------  % Date: Sat, 04 Nov 2006 15:13:37 -0500 - From: JF Mezei <jfmezei.spamnot@teksavvy.com> ' Subject: Re: TCPIP Services SPAM config 7 Message-ID: <429fe$454cf46f$cef8887a$6580@TEKSAVVY.COM>    Michael Austin wrote: Q > Why then does my system constantly try to process usernames that do not exist?  Z >And why would anyone make up a username like 3ab8dd.5003023@my-dom-main.??? (rhetorical)     J Remember what Mr SMS has said many times: 5.4's checks are somewhat brain I dead because any username that is longer than 12 bytes gets an automatic  4 entry pass into the system whether it exists or not.  H What is not clear to me is whether the SMTP Receiver checks against the 3 VMSMAIL PROFILE file or against SYSUAF.DAT or both.    ------------------------------  % Date: Sat, 04 Nov 2006 14:39:14 -0500 - From: JF Mezei <jfmezei.spamnot@teksavvy.com> $ Subject: Re: Time change questions !8 Message-ID: <737c1$454cec1c$cef8887a$28724@TEKSAVVY.COM>   Neil Rieck wrote:   L > Right now OpenVMS doesn't have any near-term date limitations so only the N > time (zone) thing needs to be fixed in order to make it as good as UNIX. So  > what are our options?   H As more and more of the software on VMS  has origins in Unix, I wouldn'tF brag about VMS being immune from the Unix time format limitation in a  couple of decades.  K I think a fair amount of software now converts time (either implicitely by  7 using C rtl routines or explicitely) to/from Unix time.    ------------------------------  $ Date: Sat, 4 Nov 2006 16:55:53 -0500) From: "Neil Rieck" <n.rieck@sympatico.ca> $ Subject: Re: Time change questions !< Message-ID: <454d0be8$0$14793$9a6e19ea@news.newshosting.com>  ; "JF Mezei" <jfmezei.spamnot@teksavvy.com> wrote in message  2 news:737c1$454cec1c$cef8887a$28724@TEKSAVVY.COM... > Neil Rieck wrote:  > M >> Right now OpenVMS doesn't have any near-term date limitations so only the  L >> time (zone) thing needs to be fixed in order to make it as good as UNIX.  >> So what are our options?  > J > As more and more of the software on VMS  has origins in Unix, I wouldn'tH > brag about VMS being immune from the Unix time format limitation in a  > couple of decades. > M > I think a fair amount of software now converts time (either implicitely by  9 > using C rtl routines or explicitely) to/from Unix time.  >   I You've brought up a very interesting point. OpenVMS engineering has done  F lots of work to allow UNIX code to run on OpenVMS will little, or no, / modifications. According to the following link: 0 http://h71000.www7.hp.com/portability/index.htmlL lots of this magic happens in the C-RTL. So it is possible that a Y2038 and K Y2106 compatibility bugs may exist in these modules. Some time before 2038  L someone in the UNIX/C world will need to either redeclare time_t and time_b H structures as unsigned. Hopefully long before 2106 those people will be 3 smart enough to create a new 64-bit clock variable.   
 Neil Rieck Kitchener/Waterloo/Cambridge,  Ontario, Canada." http://www3.sympatico.ca/n.rieck/    ------------------------------  * Date: Sat, 4 Nov 2006 20:24:51 -0600 (CST)* From: sms@antinode.org (Steven M. Schweda)$ Subject: Re: Time change questions !2 Message-ID: <06110420245123_2022581F@antinode.org>  ) From: "Neil Rieck" <n.rieck@sympatico.ca>    > [...]  Some time before 2038  N > someone in the UNIX/C world will need to either redeclare time_t and time_b J > structures as unsigned. Hopefully long before 2106 those people will be 5 > smart enough to create a new 64-bit clock variable.   E    Have you even _looked_ at "time_t" (and friends) in DECC$TYPES.H?  F (Or the equivalent on any UNIX system in the past five years?)  HavingB some grip on reality would let you redirect your fears more toward actual problems.      (What's "time_b"?)   H ------------------------------------------------------------------------  3    Steven M. Schweda               sms@antinode-org 4    382 South Warwick Street        (+1) 651-699-9818    Saint Paul  MN  55105-2547    ------------------------------   End of INFO-VAX 2006.608 ************************