1 INFO-VAX	Tue, 08 Oct 2002	Volume 2002 : Issue 555       Contents:! Re: "Intel's Huge Bet Turns Iffy" 0 Re: "Underqualified tipster backed HP merger"..., A warning to use bounds checking in BASIC(?)0 Re: A warning to use bounds checking in BASIC(?)0 Re: A warning to use bounds checking in BASIC(?)! Re: Announcing WHOIS V1.5 for VMS ! Re: Announcing WHOIS V1.5 for VMS  Apparently SUN listens Re: Apparently SUN listens Re: Apparently SUN listens* C - overlaying variables in the same PSECT. Re: C - overlaying variables in the same PSECT. Re: C - overlaying variables in the same PSECT# Re: C: sizeof question with structs # Re: C: sizeof question with structs # Re: C: sizeof question with structs # Re: C: sizeof question with structs # Re: C: sizeof question with structs # Re: C: sizeof question with structs # Re: C: sizeof question with structs # Re: C: sizeof question with structs # RE: C: sizeof question with structs # Re: C: sizeof question with structs  Re: Disable access log on CSWS! Re: e-mail web client (SilkyMail) ! Re: e-mail web client (SilkyMail) ! Re: e-mail web client (SilkyMail) ! Re: e-mail web client (SilkyMail) ! Re: e-mail web client (SilkyMail)  Re: FDDI LINKS FAILED  FMS: missing RPC$SHARE Re: For a new alpha server Hobbyist kit Contents  Re: Hobbyist kit Contents 7 Re: HP tells VMS users "Relax, you are sitting pretty"! 7 Re: HP tells VMS users "Relax, you are sitting pretty"! 7 Re: HP tells VMS users "Relax, you are sitting pretty"! 7 Re: HP tells VMS users "Relax, you are sitting pretty"! 7 Re: HP tells VMS users "Relax, you are sitting pretty"!  Re: HP to push .NET big time Re: HP to push .NET big time Re: mouning floppy drive Re: MySQL for VMS?& OpenVMS 7.3-1 and TCPware 5.6 question- Re: Problem with DEC TCP/IP and about patches - Re: Problem with DEC TCP/IP and about patches  Re: Proprietary or Open Source Re: Proprietary or Open Source Proprietary or Open Source Re: Proprietary or Open Source Re: Proprietary or Open Source Re: Proprietary or Open Source Re: Proprietary or Open Source RE: Proprietary or Open Source3 Tally Printer causing overflow data underrun errors ) Re: TPU access to DCL symbols or logicals  Re: VMS 7.3 DCL / MX  F ----------------------------------------------------------------------  $ Date: Mon, 7 Oct 2002 22:05:17 -0400) From: "Neil Rieck" <n.rieck@sympatico.ca> * Subject: Re: "Intel's Huge Bet Turns Iffy"8 Message-ID: <J3ro9.918$Iw5.181172@news20.bellglobal.com>   According to this link... ( http://www.theinquirer.net/?article=5699K it looks like Intel has some chipset related problems which could result in I more rollout delays. This make me think that we'll see EV7 and EV79 speed   records for a few years to come.  I On the flip side, I have no doubt that the DEC/Compaq engineers that went L over to Intel, as well as the the ones who stayed behind, will make IA-64 an eventual success.   
 Neil Rieck Kitchener/Waterloo/Cambridge,  Ontario, Canada.! http://www3.sympatico.ca/n.rieck/    ------------------------------  % Date: Mon, 07 Oct 2002 22:50:14 -0400 ( From: David Froble <davef@tsoft-inc.com>9 Subject: Re: "Underqualified tipster backed HP merger"... , Message-ID: <3DA247E6.3050501@tsoft-inc.com>   Jan C. Vorbrggen wrote:  G >>When I was hired, one of the documents they required was a copy of my E >>degree. Fortunately I could still find the original paper and got a  >>whole passel of copies made. >> > J > That's not SOP in the US? My, are you believing in the integrity of your > future employees.     M Actually, it is not common in the US.  In any case paper can be forged.  The  L best check is to contact the institution.  Good thing, since I'm not sure I J could lay hands on my degree.  Not sure if I still deserve it either.  :-)   Dave   ------------------------------  % Date: Mon, 07 Oct 2002 22:08:04 +0100 - From: Gerald Marsh <gerald@cyfer.demon.co.uk> 5 Subject: A warning to use bounds checking in BASIC(?) 8 Message-ID: <u4t3qus2ldmnq8rl5ualph58c6odrv2094@4ax.com>  , I hope a compiler guru can help on this one.  A The following is a cut-down version of a program to demonstrate a $ potential problem with the debugger:   $ type big_array_exception.bas ! ' ! Create large integer array and access     5         declare INTEGER Int_Array1(5%),Int_Array2(3%)              N% = 0%          FOR I%=1% TO 15%1                 Int_Array1(I%) = I%*2+123123+I%/3 ?                 PRINT "Element ";I%;" contains ";Int_Array1(I%)          NEXT I%              Int_Array2(0%) = 15%    '         PRINT "Array2 =";Int_Array2(0%)                PRINT  "Finished..."  D I know that N% had might as well not been there and the program nameF is a bit of a misnomer but a compilation without array bounds checking6 I thought would result in an ACCVIO where I% became 6.   The results are:    
 $ reca bas( $ basic big_array_exception/chec=nobound
 $ reca lin $ link big_array_exception	 $ reca ru  $ run big_array_exception  Element  1  contains  123125   Element  2  contains  123127   Element  3  contains  123130   Element  4  contains  123132   Element  5  contains  123134   Element  6  contains  123137   Element  7  contains  123139   Element  8  contains  123141   Element  9  contains  123144   Element  10  contains  123146    Element  11  contains  123148    Element  12  contains  123151    Element  13  contains  123153   ! Element  14  contains  17694728    Element  15  contains  65584  
 Array2 = 15    Finished...     E Up to the non-existent element 13 the calculation works but where the $ hell did the value for 14 come from?  E I noted that there was no accvio produced but if I%'s loop goes to 16 @ above the program accvios after the string "Finished" is output.  D Is this a feature of Alpha architecture and is it a method to ensure we all use /CHECK=BOUNDS?   F I declared the second array to see if it would encroach on the first'sB boundaries - I suspect that a 64 byte boundary has something to do
 with this.  ? (I've logged a call with CSC regarding strange behaviour of the E debugger with this program - SET BREAK/EXCEPTION doesn't! - The stack  gets corrupted.)  ' Bye for now and keep the flag a-flying!    Gerald.    Gerald Marsh  / gerald -at- cyfer -dot- demon -dot- co -dot- uk    ------------------------------  % Date: Mon, 07 Oct 2002 14:39:38 -0700 + From: "Barry Treahy, Jr." <Treahy@MMaz.com> 9 Subject: Re: A warning to use bounds checking in BASIC(?) ' Message-ID: <3DA1FF1A.6030204@MMaz.com>    Gerald Marsh wrote:   - >I hope a compiler guru can help on this one.  > B >The following is a cut-down version of a program to demonstrate a% >potential problem with the debugger:  >  >$ type big_array_exception.bas  >!( >! Create large integer array and access >  > 6 >        declare INTEGER Int_Array1(5%),Int_Array2(3%) >  >  >        N% = 0% >        FOR I%=1% TO 15% 2 >                Int_Array1(I%) = I%*2+123123+I%/3@ >                PRINT "Element ";I%;" contains ";Int_Array1(I%) >        NEXT I% >  >  >        Int_Array2(0%) = 15%  >  > ( >        PRINT "Array2 =";Int_Array2(0%) >  >  >  >        PRINT  "Finished..."  > E >I know that N% had might as well not been there and the program name G >is a bit of a misnomer but a compilation without array bounds checking 7 >I thought would result in an ACCVIO where I% became 6.  >  >The results are:  >  >  >$ reca bas ) >$ basic big_array_exception/chec=nobound  >$ reca lin  >$ link big_array_exception 
 >$ reca ru >$ run big_array_exception >Element  1  contains  123125    >Element  2  contains  123127    >Element  3  contains  123130    >Element  4  contains  123132    >Element  5  contains  123134    >Element  6  contains  123137    >Element  7  contains  123139    >Element  8  contains  123141    >Element  9  contains  123144     >Element  10  contains  123146    >Element  11  contains  123148    >Element  12  contains  123151    >Element  13  contains  123153  " >Element  14  contains  17694728   >Element  15  contains  65584    >Array2 = 15   >Finished... >  > F >Up to the non-existent element 13 the calculation works but where the% >hell did the value for 14 come from?  > F >I noted that there was no accvio produced but if I%'s loop goes to 16A >above the program accvios after the string "Finished" is output.  > E >Is this a feature of Alpha architecture and is it a method to ensure  >we all use /CHECK=BOUNDS?   > G >I declared the second array to see if it would encroach on the first's C >boundaries - I suspect that a 64 byte boundary has something to do  >with this.  > @ >(I've logged a call with CSC regarding strange behaviour of theF >debugger with this program - SET BREAK/EXCEPTION doesn't! - The stack >gets corrupted.)  > ( >Bye for now and keep the flag a-flying! >    >  Gerald,   I without bounds checking, you'll only get an access violation if you step  G on a page you do not have access to.  Put another way, these variables  B had no problem existing in the same page in a psect that you were  permitted to write to.  < If you really want to see what the compiler is doing, put a G /MACHINE/LIST/NOOPT on the basic compile statement.  You'll be able to  I see the layout of the variables as well as how the computations occurred  H and you'll probably find that at the point of element 14, the temporary I index for I% was overwritten which ended up altering the results of your   equation...    Barry    --    @ Barry Treahy, Jr  *  Midwest Microwave  *  Vice President & CIO   A E-mail: Treahy@mmaz.com * Phone: 480/314-1320 * FAX: 480/661-7028    ------------------------------  $ Date: Mon, 7 Oct 2002 21:42:54 -07002 From: "Randy Park" <rjpark@mindspring.nospaam.com>9 Subject: Re: A warning to use bounds checking in BASIC(?) 2 Message-ID: <antnq6$9fi$1@slb4.atl.mindspring.net>  8 Gerald Marsh <gerald@cyfer.demon.co.uk> wrote in message2 news:u4t3qus2ldmnq8rl5ualph58c6odrv2094@4ax.com.... > I hope a compiler guru can help on this one. > C > The following is a cut-down version of a program to demonstrate a & > potential problem with the debugger: >   > $ type big_array_exception.bas > ! ) > ! Create large integer array and access  >  > 7 >         declare INTEGER Int_Array1(5%),Int_Array2(3%)  >  >  >         N% = 0%  >         FOR I%=1% TO 15%3 >                 Int_Array1(I%) = I%*2+123123+I%/3 A >                 PRINT "Element ";I%;" contains ";Int_Array1(I%)  >         NEXT I%  >  >  >         Int_Array2(0%) = 15% >  > ) >         PRINT "Array2 =";Int_Array2(0%)  >  >  >  >         PRINT  "Finished..." > F > I know that N% had might as well not been there and the program nameH > is a bit of a misnomer but a compilation without array bounds checking8 > I thought would result in an ACCVIO where I% became 6. >  > The results are: >  >  > $ reca bas* > $ basic big_array_exception/chec=nobound > $ reca lin > $ link big_array_exception > $ reca ru  > $ run big_array_exception  > Element  1  contains  123125 > Element  2  contains  123127 > Element  3  contains  123130 > Element  4  contains  123132 > Element  5  contains  123134 > Element  6  contains  123137 > Element  7  contains  123139 > Element  8  contains  123141 > Element  9  contains  123144 > Element  10  contains  123146  > Element  11  contains  123148  > Element  12  contains  123151  > Element  13  contains  123153 ! > Element  14  contains  17694728  > Element  15  contains  65584
 > Array2 = 15 
 > Finished...  >  > G > Up to the non-existent element 13 the calculation works but where the & > hell did the value for 14 come from? > G > I noted that there was no accvio produced but if I%'s loop goes to 16 B > above the program accvios after the string "Finished" is output. > F > Is this a feature of Alpha architecture and is it a method to ensure > we all use /CHECK=BOUNDS?  > H > I declared the second array to see if it would encroach on the first'sD > boundaries - I suspect that a 64 byte boundary has something to do > with this. > A > (I've logged a call with CSC regarding strange behaviour of the G > debugger with this program - SET BREAK/EXCEPTION doesn't! - The stack  > gets corrupted.) > ) > Bye for now and keep the flag a-flying!  > 	 > Gerald.  >  > Gerald Marsh > 1 > gerald -at- cyfer -dot- demon -dot- co -dot- uk   : Based upon your output, I would guess that INTEGER default is LONG (4 bytes).  = NoBounds checking means that array bounds will not be checked = at run time.  If you exceed the array bound, you get what you 8 get.  You may stomp on other data fields, you may access9 allocated but unused memory, or you may attempt to access = (read or write) outside your allocated memory.  If the latter ; happens you will get an ACCVIO error.  You might even stomp  on string headers.  7 Int_Array1 has 6 elements (0 to 5) and Int_Array2 has 4 > elements (0 to 3).  The compiler probably allocates Int_Array2: immediatly after Int_Array1 in the same PSECT.  There's no: gap in allocation between these two arrays.  Other integer; data may follow in the same PSECT, but aligned on an 8 byte 	 boundary.   ; I have a rather large program that uses no bounds checking, 8 but the code that accesses arrays is well controlled and9 doesn't exceed array bounds and on occasion does it's own : array bounds checking.  I get a 10% reduction in CPU usage in some modules by doing this.   ------------------------------  $ Date: Mon, 7 Oct 2002 15:31:35 -0400; From: "Brian Tillman" <tillman_brian@notnoone.notnohow.com> * Subject: Re: Announcing WHOIS V1.5 for VMS$ Message-ID: <3da1e15c$1@news.si.com>  K >So I set out to build a better WHOIS client which can handle today's needs ? >better. It is fully native to VMS, designed for TCPIP Services  > ! >It is available for download at: / http://pages.infinit.net/jfmezei/vms/index.html   J Could you make this a little more compatible with other TCP/IP stacks?  In+ particular, the /SERVER qualifier is better J written /HOST (to correspond with the qualfier in Multinet and TCPware andF to be more in line with the -h (i.e., host) option in Unix versions of whois).  --A Brian Tillman                   Internet: tillman_brian at si.com A Smiths Aerospace                          tillman at swdev.si.com = 3290 Patterson Ave. SE, MS      Addresses modified to prevent < Grand Rapids, MI 49512-1991     SPAM.  Replace "at" with "@"8        This opinion doesn't represent that of my company   ------------------------------  % Date: Mon, 07 Oct 2002 16:16:37 -0400 - From: JF Mezei <jfmezei.spamnot@videotron.ca> * Subject: Re: Announcing WHOIS V1.5 for VMS, Message-ID: <3DA1EBA0.8B190C04@videotron.ca>   Brian Tillman wrote:L > Could you make this a little more compatible with other TCP/IP stacks?  In- > particular, the /SERVER qualifier is better L > written /HOST (to correspond with the qualfier in Multinet and TCPware andH > to be more in line with the -h (i.e., host) option in Unix versions of	 > whois).    Suggestion noted.    ------------------------------  # Date: Mon, 07 Oct 2002 18:42:05 GMT # From: "John Smith" <a@nonymous.com>  Subject: Apparently SUN listens I Message-ID: <1Ako9.201958$8b1.29245@news01.bloor.is.net.cable.rogers.com>   2 http://www.eweek.com/article2/0,3959,588641,00.asp   October 3, 2002 * In Reversal, Sun to Ship $99 Solaris 9 x86   By  Peter Galli   K Sun Microsystems Inc. on Friday will announce plans to release an unbundled K version of the Solaris 9 operating environment for non-Sun x86 hardware for  $99 for a single CPU system.  G After months of indecision, Sun will now ship Solaris 9 x86, unbundled, F supporting both the Sun hardware platform, for both current and futureK products, as well as the same list of all hardware supported for Solaris 8, K according to Sun officials in Palo Alto, Calif. Early access to the Solaris L 9 bits will come in the next four to eight weeks, and the final product will* ship in the January time-frame, they said.  C The announcement brings to an end months of uncertainty about Sun's J intentions in this regard. In early January the Palo Alto, Calif., companyJ announced that it had decided to defer the productization of the Solaris 9K Intel version; a week later it announced plans to meet a group of customers J and developers to try to work out a compromise; and in August it announced9 that it would only support Solaris 9 on Sun x86 hardware.   K John Loiacono, vice president of the operating platforms group at Sun, told K eWeek in an interview Thursday that the company remained fully committed to J supporting the Solaris x86 platform going forward, which supports both Sun and non-Sun hardware.   G Sun will also work with the community to provide the compatibility test I suite it used internally, and was considering opening up the code for the < product to the open source community sometime in the future.  I "Our announcement earlier this year that we might not be able to continue F with Solaris 9 x86 caused a huge upheaval, and people wanted an answerH sooner than we could give it to them," he said. "We always said we wouldG continue to talk to the community and find a viable business model that 1 would support it unbundled. We've now done that."   H The Solaris 9 x86 rollout is exactly the same delivery time frame as forJ Sun's own x86 offering, the LX50. Sun will also continue with its plans toK deploy its entire SunONE (Open Network Environment) stack, as well as other I key software components on the Solaris, Linux and Solaris x856 platforms,  Loiacono said.  J "We now have a new plan and business model around Solaris x86, and, unlikeF with Solari8 x86 which was available for free download, we will now beJ selling Solaris 9 x86 starting at $99," he said. "The customers we talk toF feel this still offers good value, especially as it runs on relativelyD low-cost hardware. I also believe it is viable as a long-term viable business we can sustain."   J In the past Sun did not ship Solaris bundled on its own hardware. But withE the advent of the LX50, follow on products and diversification of the J Solaris x86 product line, "you see that the whole company is now behind itL rather than trying to sell it as an unbundled product," he said. "That's whyF we see it as a sustainable business for us moving forward and actuallyL addressing the unbundled platforms is something we can now sustain as well."  K Graham Lovell, director of marketing for operating systems and availability F at Sun, said the business model around this was buoyed by a ramping ofI interest in the platform from its existing installed base, new prospects, , ISVs and those interested in software sales.  H "So this is now a vastly different business proposal than before when weJ were just essentially selling an operating system on an x86 box," he said.  E Sun will continue to engage with the community and will work with the K community to provide the Sun compatibility test suite used to validate that # x86-based products and systems run.   J Sun will also probably open source this product sometime in the future. As@ such, it will work with the community to put together a hardwareL compatibility list that expand the range of systems known to work on Solaris on x86.   H Sun was also exploring a more collaborative working arrangement with theI community involving things like code exchanges, information exchanges and I the like, Lovell said. But he pointed out that while Sun would supply and L support its SunONE software products on the x86 operating system as a stack,J some of this was pre-bundled for free on its LX50 hardware along with someB other components that were supplied under various licensing terms.  J "So there won't be parity between Sun hardware and non-Sun hardware. ThereF will still be an element of software that the LX50 comes with that theK non-Sun hardware won't get. Things we might include in the LX50 as a bundle K are the streaming server. The general principle we are operating on is that L if it available for SPARC it should be available for the LX50," Lovell said.  H While the community had predicted that Sun could sell as many as 100,000L copies of the unbundled Solaris 9, Lovell said he expected that number to beI closer to between 20,000 and 50,000 for single CPU systems, tailing as it   moved into two-plus CPU systems.  G Looking to the future, Sun's Loiacono said Solaris 10 would continue to I build on a lot of the investment that had gone into Solaris two and three J years ago, which would be manifested in Solaris 10. This would not just beI incremental increases in functionality, but include new technologies that H takes the container concept in Solaris 9 further and delivers a complete+ solution in that space around availability.   = "We'll see things in the file system area that are phenomenal H breakthroughs," he said. "There are five or six substantial areas aroundD availability, security and file system functionality that will be an absolute breakthrough."    ------------------------------  # Date: Mon, 07 Oct 2002 23:21:19 GMT $ From: "Ed Wilts" <ewilts@ewilts.org># Subject: Re: Apparently SUN listens * Message-ID: <PFoo9.1153$hb4.777@sccrnsc02>   John Smith wrote: 4 > http://www.eweek.com/article2/0,3959,588641,00.asp >  > October 3, 2002 , > In Reversal, Sun to Ship $99 Solaris 9 x86  E I snipped the entire copyrighted article which was most likely posted  without permission.   L This has absolutely nothing whatsoever to do with VMS.  Please keep the list topics somewhat VMS-related.  0     .../Ed (who's been here for about 15 years!)   -- Ed Wilts Mounds View, MN, USA mailto:ewilts@ewilts.org   ------------------------------  # Date: Tue, 08 Oct 2002 01:23:51 GMT # From: "John Smith" <a@nonymous.com> # Subject: Re: Apparently SUN listens G Message-ID: <Hsqo9.42536$Aiq1.192@news04.bloor.is.net.cable.rogers.com>   / "Ed Wilts" <ewilts@ewilts.org> wrote in message $ news:PFoo9.1153$hb4.777@sccrnsc02... > John Smith wrote: 6 > > http://www.eweek.com/article2/0,3959,588641,00.asp > >  > > October 3, 2002 . > > In Reversal, Sun to Ship $99 Solaris 9 x86 > G > I snipped the entire copyrighted article which was most likely posted  > without permission.  > I > This has absolutely nothing whatsoever to do with VMS.  Please keep the  list > topics somewhat VMS-related. > 2 >     .../Ed (who's been here for about 15 years!)  - Nice that you've been around for 15 years ED.   I Just curious....How many times have you written to Ken Olson, Bob Palmer, K Capellas, or Carly telling them that what you see them doing to VMS and not F listening to the customer base is harming VMS? And how often have they" listend to what you've had to say?  D The article is relevant because it is the antithesis of how ChumHPaq behaves.   ------------------------------   Date: 7 Oct 2002 14:31:27 -0700 " From: cstranslations@msn.com (Joe)3 Subject: C - overlaying variables in the same PSECT = Message-ID: <d56d1c2d.0210071331.403cdbe7@posting.google.com>   ' Compaq C V6.5-001 on OpenVMS Alpha V7.3   E Sent the below to support last Thursday. Haven't heard anything back.  They must be rather busy...   B Sorry about any wrap on this. Supposing that I want to overlay twoF variables on the same memory (do the same thing as set up multiple MAPD statements in BASIC where the MAP statements have the same "name")  how might I do this in C?   > From the docs what I seem to want is some variation of #pragmaD extern_model. I am trying the following but this doesn't seem end up@ with the two variables in question starting at the same address:  > #include "socket.h"     // TPCIP headers among other things...  
 struct IOSB {     short  status;     short  XferSize;     void   *ptr;  };   typedef struct FTPSRECORD {     int                 RecNum;    struct IOSB         iosb;&    char                DeviceName[16];?    struct sockaddr_in  rhost;   // the standard TCPIP struct... )    char                DottedAddress[16]; $    char                FillSpace[3];"    short               PortNumber;
 } FTPSRECORD;    #pragma extern_model save , #pragma extern_model strict_refdef "GblSect" nopic,ovr,noexe,wrt,shr,novec  FTPSRECORD  ftps[64];     , #pragma extern_model strict_refdef "GblSect" nopic,ovr,noexe,wrt,shr,novec ' char        GlobalSectionOverlay[8192];  #pragma extern_model restore   void main(void) {      }   C According to the linker map I'm ending up with the variables "ftps" D and "GlobalSectionOverlay" concatenated in memory. This is confirmedC if I run the above under the debugger and look at the locations via F EV/ADDRESS(GlobalSectionOverlay starts roughly where I would expect it0 to start given the size of the variable "ftps").  A Any pointers in the right direction examples greatly appreciated.      tia  joe    ------------------------------  % Date: Mon, 07 Oct 2002 16:55:00 -0600  From: Kevin Handy <kth@srv.net> 7 Subject: Re: C - overlaying variables in the same PSECT $ Message-ID: <94oo9.2242$G5.217@fe01>  
 Joe wrote:) > Compaq C V6.5-001 on OpenVMS Alpha V7.3  > G > Sent the below to support last Thursday. Haven't heard anything back.  > They must be rather busy...  > D > Sorry about any wrap on this. Supposing that I want to overlay twoH > variables on the same memory (do the same thing as set up multiple MAPF > statements in BASIC where the MAP statements have the same "name")  > how might I do this in C?  > @ > From the docs what I seem to want is some variation of #pragmaF > extern_model. I am trying the following but this doesn't seem end upB > with the two variables in question starting at the same address: > @ > #include "socket.h"     // TPCIP headers among other things... >  > struct IOSB {  >    short  status;  >    short  XferSize;  >    void   *ptr;  > }; >  > typedef struct FTPSRECORD {   >    int                 RecNum; >    struct IOSB         iosb;( >    char                DeviceName[16];A >    struct sockaddr_in  rhost;   // the standard TCPIP struct... + >    char                DottedAddress[16]; & >    char                FillSpace[3];$ >    short               PortNumber; > } FTPSRECORD;  >  > #pragma extern_model save . > #pragma extern_model strict_refdef "GblSect" > nopic,ovr,noexe,wrt,shr,novec  > FTPSRECORD  ftps[64];  >  > . > #pragma extern_model strict_refdef "GblSect" > nopic,ovr,noexe,wrt,shr,novec ) > char        GlobalSectionOverlay[8192];  > #pragma extern_model restore >  > void main(void) {  >  >  > }  > E > According to the linker map I'm ending up with the variables "ftps" F > and "GlobalSectionOverlay" concatenated in memory. This is confirmedE > if I run the above under the debugger and look at the locations via H > EV/ADDRESS(GlobalSectionOverlay starts roughly where I would expect it2 > to start given the size of the variable "ftps"). > C > Any pointers in the right direction examples greatly appreciated.   D Wouldn't a 'union' work? That's the normal way to overlap using 'C',= and it doesn't require any pragma's or setting up any special  sections to make it work.    ------------------------------  % Date: Mon, 07 Oct 2002 23:58:56 -0000 ! From: Z  <zarlenga@conan.ids.net> 7 Subject: Re: C - overlaying variables in the same PSECT / Message-ID: <uq47u03lh7bn73@corp.supernews.com>   # Joe <cstranslations@msn.com> wrote: D : Sorry about any wrap on this. Supposing that I want to overlay twoH : variables on the same memory (do the same thing as set up multiple MAPF : statements in BASIC where the MAP statements have the same "name")  : how might I do this in C?   
 With a union.    union  {      short  var1;     long   var2;     double var3; } vars;   E var1, var2 and var3 will all start at the same address in memory.  If E all the members of tehunion are the same ize, they will overlay each   other completely.    ------------------------------  % Date: Mon, 07 Oct 2002 20:16:52 +0200 6 From: Arne =?ISO-8859-1?Q?Vajh=F8j?= <arne@vajhoej.dk>, Subject: Re: C: sizeof question with structs) Message-ID: <3DA1CF94.9050309@vajhoej.dk>    JF Mezei wrote:     > Lets say I have the following: >  > struct recipe_struct  {    > 					char ingredient[30] ; > 					int  amount ; > 					short metric_flg ;  > 					char comment[100] ; > 				 } >  > L > I know that in a program, I can get the size of the whole structure with : > % > 	x = sizeof(struct recipe_struct) ;  > > > But is there a way to get the size of an individual member ?  5 I think most people just use sizeof on the data types  of the members.   9 Tne NULL pointer trick suggested by others will allow you 7 to change the data type of members without changing the - sizeof code. But how often do you need that ?    Arne   ------------------------------  % Date: Mon, 07 Oct 2002 11:40:19 -0000 ! From: Z  <zarlenga@conan.ids.net> , Subject: Re: C: sizeof question with structs/ Message-ID: <uq2sl35bd1iubc@corp.supernews.com>   . JF Mezei <jfmezei.spamnot@videotron.ca> wrote:  : Lets say I have the following: : struct recipe_struct {    - :                       char ingredient[30] ; % :                       int  amount ; * :                       short metric_flg ;+ :                       char comment[100] ;  : 		       }  > One thing to keep in mind when dealing with sizeof structures:= the size may vary (in the case above, it will vary) depending - on the alignment of the structure's elements.     L : I know that in a program, I can get the size of the whole structure with :% : 	x = sizeof(struct recipe_struct) ; > : But is there a way to get the size of an individual member ? : I know I can do:( : 	struct recipe_struct chocolate_cake ;. : 	keylen = sizeof(chocolate_cake.ingredient);H : But I'd like to get the size without having to allocate the structure.  @ Cast a NULL pointer to a pointer to the structure containing theE element, then dereference to the element you want and take the sizeof  that.    Do you need an example?    ------------------------------  % Date: Mon, 07 Oct 2002 11:50:48 -0000 ! From: Z  <zarlenga@conan.ids.net> , Subject: Re: C: sizeof question with structs/ Message-ID: <uq2t8o86uefj0e@corp.supernews.com>   ! Z <zarlenga@conan.ids.net> wrote: B : Cast a NULL pointer to a pointer to the structure containing theG : element, then dereference to the element you want and take the sizeof  : that.    : Do you need an example?    Here's an example:       #include <stdio.h>   struct example_s {          char   some_text[100];         int    an_integer;         double two_doubles[2]; };   main() { 
 int     size;   A     size = sizeof( ( (struct example_s *)(NULL) )->two_doubles ); 3     printf("size of two doubles : %d\n", size);      }    ------------------------------  % Date: Mon, 07 Oct 2002 15:09:27 -0400 - From: JF Mezei <jfmezei.spamnot@videotron.ca> , Subject: Re: C: sizeof question with structs, Message-ID: <3DA1DBE6.956E5041@videotron.ca>  = >  keylen = sizeof(((struct recipe_struct *) 0)->ingredient);     ' Thanks to all who answered. Neat trick.    ------------------------------  % Date: Mon, 07 Oct 2002 15:26:08 -0400 - From: JF Mezei <jfmezei.spamnot@videotron.ca> , Subject: Re: C: sizeof question with structs, Message-ID: <3DA1DFCE.27B8EE8E@videotron.ca>   Arne Vajhj wrote:; > Tne NULL pointer trick suggested by others will allow you 9 > to change the data type of members without changing the / > sizeof code. But how often do you need that ?   K Not often, but yesterday, I stumbled on a situation where I needed to zap X M bytes with spaces into some string that wasn't part of a structure, but the X G was the size of a member of a structure. Seemed rather silly to have to F allocate a "temp" structure just so I could so sizeof(temp.ingredient)   ------------------------------  % Date: Mon, 07 Oct 2002 22:02:14 +0200 6 From: Arne =?ISO-8859-1?Q?Vajh=F8j?= <arne@vajhoej.dk>, Subject: Re: C: sizeof question with structs) Message-ID: <3DA1E846.4030505@vajhoej.dk>    JF Mezei wrote:    > Arne Vajhj wrote:; >>Tne NULL pointer trick suggested by others will allow you 9 >>to change the data type of members without changing the / >>sizeof code. But how often do you need that ?     M > Not often, but yesterday, I stumbled on a situation where I needed to zap X O > bytes with spaces into some string that wasn't part of a structure, but the X I > was the size of a member of a structure. Seemed rather silly to have to H > allocate a "temp" structure just so I could so sizeof(temp.ingredient)     True.   6 But if the data-structure in question is a char array,* then often a #define could have been used.   Arne   ------------------------------  % Date: Mon, 07 Oct 2002 19:57:32 -0000e! From: Z  <zarlenga@conan.ids.net>e, Subject: Re: C: sizeof question with structs/ Message-ID: <uq3ppctj1hjgbf@corp.supernews.com>e  $ Arne Vajhj <arne@vajhoej.dk> wrote:; : Tne NULL pointer trick suggested by others will allow you 9 : to change the data type of members without changing theo/ : sizeof code. But how often do you need that ?N  D You never _need_ it, it just makes the code much easier to maintain.   And it's not a "trick."g   ------------------------------  % Date: Mon, 07 Oct 2002 20:06:59 -0000T! From: Z  <zarlenga@conan.ids.net>t, Subject: Re: C: sizeof question with structs/ Message-ID: <uq3qb3f8b62n52@corp.supernews.com>u  . JF Mezei <jfmezei.spamnot@videotron.ca> wrote:> :>  keylen = sizeof(((struct recipe_struct *) 0)->ingredient);) : Thanks to all who answered. Neat trick.n  B You _could_ make a macro so that you have easy and quick access to$ that code ...                         8 #define    sizeof_el(s,e)    sizeof(((struct s *) 0)->e)  9 ... then you can use sizeof_el(recipe_struct,ingredient);    0    ------------------------------  $ Date: Mon, 7 Oct 2002 17:00:30 -0700# From: "Tom Linden" <tom@kednos.com>A, Subject: RE: C: sizeof question with structs9 Message-ID: <CIEJLCMNHNNDLLOOGNJIIEPBFNAA.tom@kednos.com>s  7 Do all C compilers support the casting of null pointer?"= (Of course, on some machines, in the past anyway, NULL wasn't,? always zero)  Upon first looking at it I would have thought thee2 compiler would have generated a diagnostic.  Cute.   >-----Original Message-----P( >From: Z [mailto:zarlenga@conan.ids.net]' >Sent: Monday, October 07, 2002 1:07 PM  >To: Info-VAX@Mvb.Saic.Com- >Subject: Re: C: sizeof question with structsC >P > / >JF Mezei <jfmezei.spamnot@videotron.ca> wrote:n? >:>  keylen = sizeof(((struct recipe_struct *) 0)->ingredient); * >: Thanks to all who answered. Neat trick. > C >You _could_ make a macro so that you have easy and quick access toi% >that code ...                       r >e9 >#define    sizeof_el(s,e)    sizeof(((struct s *) 0)->e)s > ; >.... then you can use sizeof_el(recipe_struct,ingredient);  >l >0 >  >---' >Incoming mail is certified Virus Free.-; >Checked by AVG anti-virus system (http://www.grisoft.com).,A >Version: 6.0.394 / Virus Database: 224 - Release Date: 10/3/2002a >  ---c& Outgoing mail is certified Virus Free.: Checked by AVG anti-virus system (http://www.grisoft.com).@ Version: 6.0.394 / Virus Database: 224 - Release Date: 10/3/2002   ------------------------------  % Date: Tue, 08 Oct 2002 04:15:53 -0000i! From: Z  <zarlenga@conan.ids.net>m, Subject: Re: C: sizeof question with structs/ Message-ID: <uq4mvptbf4f9e2@corp.supernews.com>e  " Tom Linden <tom@kednos.com> wrote:9 : Do all C compilers support the casting of null pointer?s? : (Of course, on some machines, in the past anyway, NULL wasn'tlA : always zero)  Upon first looking at it I would have thought thew4 : compiler would have generated a diagnostic.  Cute.  ? All compilers should suppoort casting the NULL pointer, and the<7 actual value of NULL is unimportant for this statement:r  :    sizeof( ( (struct [struct_name] *)(NULL) )->[element] )  8 You can substititute any # for NULL and it'll stll work.   Neat, hunh?  God, I love C!d   ------------------------------  % Date: Mon, 07 Oct 2002 20:37:34 +0200a6 From: Arne =?ISO-8859-1?Q?Vajh=F8j?= <arne@vajhoej.dk>' Subject: Re: Disable access log on CSWSe) Message-ID: <3DA1D46E.7010507@vajhoej.dk>e   Rick Barry wrote:l  I > Sending to NL: will work, as Alan suggests, and will improve throughputt; > considerably for those who don't need access information.n    	 Will it ?l  7 It will obvious improve the situation if the bottlenecki is the disk in question.  4 But I am not so sure it will improve performance for the web-server.m  1 I just made a small test on my home system and myA% disks are 4 times faster than NLA0: !C  1 (test program attached below - yes it is Pascal -@+ I was in the mood for writing Pascal today)n   Arne  7 =======================================================e     program z(input,output);   vart     i,t1,t2 : integer;     f       : text;s   beginr     open(f,'z.z',new);     rewrite(f);      t1 := clock;     for i:=1 to 100000 do begin%$        writeln(f,'This is a test!');     end;     t2 := clock;
     close(f);)     writeln(t2-t1);      open(f,'NLA0:',new);     rewrite(f);i     t1 := clock;     for i:=1 to 100000 do begin $        writeln(f,'This is a test!');     end;     t2 := clock;
     close(f);      writeln(t2-t1);  end.   ------------------------------  % Date: Mon, 07 Oct 2002 20:22:54 +0200o6 From: Arne =?ISO-8859-1?Q?Vajh=F8j?= <arne@vajhoej.dk>* Subject: Re: e-mail web client (SilkyMail)) Message-ID: <3DA1D0FE.2070202@vajhoej.dk>   % winston@SSRL.SLAC.STANFORD.EDU wrote:c  J > It can get confused about the total number of unread emails, but I'll beL > surprised to hear that it ever deletes the wrong message, because it seemsJ > to keep track of the messages not by sequential position in the newmail H > folder but by RMS record number.  (You say "delete 3", but the client P > translates that into an RMS record number and RMS will report an error if it'sL > already been deleted, rather than deleting the new 3d entry in the newmail > folder.)      % OK, that was what I was asking about.1  I > POP users seem to default to unified download and delete operations, oriK > download and never delete - the clients remember something about the lastrQ > message they read that keeps the same client from downloading the same messagesoM > over and over; as a result, there's little incentive to do mailbox cleanup."    > Actually most of the webmail interfaces using POP3 allows both= read but leave on server (for peeking in ones email when awayr: from the PC permanently storing the email) and delete (for getting rid of spam).e   Arne   ------------------------------  % Date: Mon, 07 Oct 2002 20:28:22 +0200i6 From: Arne =?ISO-8859-1?Q?Vajh=F8j?= <arne@vajhoej.dk>* Subject: Re: e-mail web client (SilkyMail)) Message-ID: <3DA1D246.9020503@vajhoej.dk>    Francesco Gennai wrote:i  T > Arne Vajhj <arne@vajhoej.dk> wrote in message news:<3D9FF6E7.50209@vajhoej.dk>...- >>To elaborate a bit: the reason that this is . >>usually not a problem is that email accounts/ >>is usually personal. It is probably extremely - >>rare that two web-sessions would access the3 >>same POP-account.c > E > Please take a look at RFC 1939 for UIDL commad and all of you coulde( > find a good answer to this discussion.     ????  5 Since the unique ID can be used for the DELE command,g6 then it does actually not solve the problem discussed.   Arne   ------------------------------  $ Date: Mon, 7 Oct 2002 15:41:39 -0400; From: "Brian Tillman" <tillman_brian@notnoone.notnohow.com>T* Subject: Re: e-mail web client (SilkyMail)$ Message-ID: <3da1e3b8$1@news.si.com>  H >But don't they first move the email from the remote server via pop to a localAC >mailbox before serving that local mailbox to the client via http ?   I Endymion's offering uses POP to move the files to a separate server-basedlE store (not the local client) and then manipulates the copy.  DeletingoL message three deletes the mailman copy, not message three on the POP server. --A Brian Tillman                   Internet: tillman_brian at si.combA Smiths Aerospace                          tillman at swdev.si.comr= 3290 Patterson Ave. SE, MS      Addresses modified to preventt< Grand Rapids, MI 49512-1991     SPAM.  Replace "at" with "@"8        This opinion doesn't represent that of my company   ------------------------------  % Date: Mon, 07 Oct 2002 21:57:03 +0200 6 From: Arne =?ISO-8859-1?Q?Vajh=F8j?= <arne@vajhoej.dk>* Subject: Re: e-mail web client (SilkyMail)) Message-ID: <3DA1E70F.7000106@vajhoej.dk>t   Brian Tillman wrote:  K > Endymion's offering uses POP to move the files to a separate server-based G > store (not the local client) and then manipulates the copy.  Deleting N > message three deletes the mailman copy, not message three on the POP server.  0 Not the version of Endymion Mailman my ISP uses.  3 If I delete email via the web-interface, then it isi gone when I download via POP3.   Arne   ------------------------------  # Date: Mon, 07 Oct 2002 21:03:06 GMTe$ From: winston@SSRL.SLAC.STANFORD.EDU* Subject: Re: e-mail web client (SilkyMail)8 Message-ID: <00A15192.B9680E59@SSRL04.SLAC.STANFORD.EDU>  b In article <3DA1D0FE.2070202@vajhoej.dk>, Arne =?ISO-8859-1?Q?Vajh=F8j?= <arne@vajhoej.dk> writes:& >winston@SSRL.SLAC.STANFORD.EDU wrote: >dK >> It can get confused about the total number of unread emails, but I'll beOM >> surprised to hear that it ever deletes the wrong message, because it seemsaK >> to keep track of the messages not by sequential position in the newmail cI >> folder but by RMS record number.  (You say "delete 3", but the client tQ >> translates that into an RMS record number and RMS will report an error if it'stM >> already been deleted, rather than deleting the new 3d entry in the newmaile
 >> folder.)    >  >C& >OK, that was what I was asking about. >eJ >> POP users seem to default to unified download and delete operations, orL >> download and never delete - the clients remember something about the lastR >> message they read that keeps the same client from downloading the same messagesN >> over and over; as a result, there's little incentive to do mailbox cleanup. >a > ? >Actually most of the webmail interfaces using POP3 allows both3> >read but leave on server (for peeking in ones email when away; >from the PC permanently storing the email) and delete (forj >getting rid of spam). >>  J By "POP users" I really meant "the people who were using POP here at SSRL C before most of them bailed to Exchange and we got the rest to go tonN IMAP-on-VMS", since that's the only population of POP users I got to watch. WeO didn't have a webmail-to-POP interface for them  -- since we were using YahMAILeJ there wasn't really any need to get POP involved -- so they mostly chose aN setting in Eudora and stuck with it.  Those who only wanted to read their mailK at work did unified download-and-delete; those who wanted to read mail from M home as well tended to do download and _never_ delete, which meant that theiraN newmail folders grew really huge.  Sometimes their POP processes couldn't readI through all the old newmail to get to the new newmail without blowing out O process memory; sometimes it would just take so long to do it that their "check"H mail every minute" settings would result in spinning up multiple new POPJ processes before their old ones got back to them, by which time the client! wasn't paying attention any more.u   I don't like POP much.   -- Alan    ------------------------------  $ Date: Mon, 7 Oct 2002 21:25:32 +0200" From: "Hans Vlems" <hvlems@iae.nl> Subject: Re: FDDI LINKS FAILED5 Message-ID: <ansn3e$h05sk$1@ID-143435.news.dfncis.de>s  : "Fred Jiang" <r54123@email.sps.mot.com> schreef in bericht% news:anqr85$36e$1@newshost.mot.com...1 > buddy,J > We have two ES40 servers and a DS10 (quorum node) consisted as a cluster	 > system.9I > We use DEFPAAA (PCI TO FDDI adapter) to connect them together. However,  LastJ > time re rebooted the two ES40 while keep DS10 on line. Somehow, We found one5I > of the members joined and Both ES 40 hung.the LED of the DEFPAAA showed  thatH > links were failling and then link failed.Does any friend know the root	 > reason?e >eE It would certainly help if you could post the output of the followingr	 commands:o   $ mc sysgen sho votese $ mc sysgen sho expected_votes  E Otherwise you leave us to guessing what happened; could be a hardwareiJ faillure (how are the systems connected, in a ring or point-to-point to an FDDI concentrator) or software.H   ------------------------------  % Date: Mon, 07 Oct 2002 17:52:20 -0400 - From: JF Mezei <jfmezei.spamnot@videotron.ca>  Subject: FMS: missing RPC$SHARE , Message-ID: <3DA20212.FD5A132E@videotron.ca>   VAX VMS 7.2, FMS 2.4  1 FMS/CONVERT fails due to a missing RPC$SHARE.EXE T  K I have not used FMS/CONVERT in a long long time, so it is possible that thenF last time I used it, I was still at VMS 5.5-2 (went from 5.5-2 to 7.2)  H I just checked and RPC$SHARE.EXE doesn't exist in the V7.2 distribution.  M If I find it in an earlier distribution, will it hurt to place it on my 7.2 ?h5 Does anyone know what RPC$SHARE.EXE did in the past ?y  N FMS/CONVERT just takes a .FLG text source file and converts it into a compiled
 .FRM form.  % Am I the only one with this problem ?    ------------------------------  # Date: Tue, 08 Oct 2002 01:15:55 GMT.# From: "John Smith" <a@nonymous.com>t# Subject: Re: For a new alpha servermI Message-ID: <flqo9.204606$8b1.85502@news01.bloor.is.net.cable.rogers.com>   0 "Yong Liu" <fdu9774@rogers.com> wrote in messageB news:E9po9.155306$q41.5844@news02.bloor.is.net.cable.rogers.com... > Hi,  > I > Here is the question. The lab needs a new alpha server. I got a list of L > record of what we have (from a three year old machine). If I list shopping& > list as below. Am I over specifying?K > Please go line by line and let me know which one I don't need to specify.a BeI > honest, I don't know what some of the item is for. Thanks for the help.  >u( >       CAS ASVR DS20E 67/667 256 MB VMS- >       3 Channel Raid Controller 8MB Cache??a( >       Battery Backup for KZPAC Cache??, >       Ultra 68VHD 2 metre cable assembly??< >       VMS Style Keyboard: US ( Does that come as default?)) >       CAS 1.8M VIDEO EXTENSION CABLE ??u5 >       Compaq Cable Extension KB/MON/MOU (default??)', >       Compaq S510 15" MPRII COLOUR MONITOR( >       CAS DS20E 67/667 SMP UPG VMS ???. >       1 Channel WIDE ULTRA-2 (LVD) Adapter??4 >       3D Lab OXGEN VXI 32MB PCI Graphics Adapter??( >       PCI to 10/100 Ethernet NIC 8255x& >       CSW 12/24GB Internal DAT Drive >       DS20E SCSI Cable??' >       18GB 10K ULTRA3 UNI Hot Plug HDu1 >       BA36R-RD RH ULTRA SCSI Shelf CAB SplitB?? % >       9.1GB 10K RPM Ultra SCSI Disk ) >       3X-DEFPA UNI=PCI to FDDI DAS w/sc ) >       CAS PCL LVD SCSI 10/100 ENET 2D??m* >       CAS DS20E/ES40 6 SLOT DRIVE CAGE??" >       CSW 12/24GB DAT DRIVE TTOP    H I would not worry about specific components until you clearly understand your needs.p  @ The machine above has several high-availability features - RAID,D battery-backed cache, fiber-channel. Are you in fact using all theseI capabilities? If so, then you need a newer version of the same. But don't:I forget that you can cluster the old machine with a new machine to provide>5 better availability and/or compute farm capabilities.e  J You may want to consider DLT for backup if you are acquiring lots of data., DAT is ok but slow for high-capacity backup.   ------------------------------  $ Date: Mon, 7 Oct 2002 17:08:46 -05002 From: "Stuart Johnson" <ssj152 AT charter DOT net> Subject: Hobbyist kit Contents/ Message-ID: <uq41ff3giqqf69@corp.supernews.com>   J Does anyone have any information on what is going to be in the forthcomingG hobbyist kits from Montgar? I am hoping that it will contail the entire.C layered products disks, ie. be the full OS distribution and layereda products. I can dream, can't I?a   Thanks in advance, Stuart Johnson ssj152 AT charter DOT nets   ------------------------------  % Date: Mon, 07 Oct 2002 23:00:41 -0400  From: RCS / RI <rcs@osfn.org>r" Subject: Re: Hobbyist kit Contents' Message-ID: <3DA24A59.E9F7E31@osfn.org>o   Stuart Johnson wrote:i > L > Does anyone have any information on what is going to be in the forthcomingI > hobbyist kits from Montgar? I am hoping that it will contail the entiredE > layered products disks, ie. be the full OS distribution and layered,! > products. I can dream, can't I?   B Has there been any discussion of how the HP merger will affect the Hobbyist program?h  B Was there any news about the Hobbyist at the Encompass conference?   -mikeu Michael Umbricht   --  H The Retro-Computing Society of RI, Inc.                   (401) 861-1977H 25 Eagle St Bldg 5 Ste 206                      http://www.osfn.org/rcs/H Providence RI 02908                       telnet://kronos.egr-ri.ids.net   ------------------------------  $ Date: Mon, 7 Oct 2002 17:07:03 -0400* From: "Bill Todd" <billtodd@metrocast.net>@ Subject: Re: HP tells VMS users "Relax, you are sitting pretty"!6 Message-ID: <JqadnWBnmeT6ajygXTWc3w@News.GigaNews.Com>  : "JF Mezei" <jfmezei.spamnot@videotron.ca> wrote in message& news:3DA1DE8B.370DBDC8@videotron.ca... > Fred Kleinsorge wrote: > >SI > > I'd hate to insult morons by calling you one.  You claimed that HP-UXi2 > > requires Itanium-2, which is completely false. >n? > Is it possible that IA64 version 1.0 was only good enough foreI > workstation/small systems and that they have to wait for the version IIs before> > they can start thinking about building large scale machines?  K While it's entirely reasonable to characterize Itanic1 as not good for muchhL of anything, that doesn't bear on the question of whether HP-UX would run on it.a   >lK > Isn't it also true that they have to wait for another generation of IA64s J > before they can think of building the larger VMS machines that some hope > customers will migrate to ?   K It had better not be, because from all current indications there will be nokI substantive changes in Itanic (save for process-shrinks and on-chip cacherC expansion) until 2005 (and likely later for any real core changes).e  L Assuming that's the case (and assuming that EV79 doesn't get canned), ItanicL will remain inferior in most performance respects to Alpha over that period,K but not so much inferior that a customer couldn't use it in a large machinegI if s/he wanted to (after all, older Alphas with less performance than theeK new Itanics will have were usable in large configurations - the result justeJ wouldn't be state-of-the-art today, and thus it's not clear why a customer. would choose Itanic over Alpha in such cases).  2  Is IA64 ready to run Tandem's NSK or do they also( > have to wait for the next generation ?  I See above:  given that no significant changes appear scheduled for ItaniceK before at least 2005, NSK (ISTR scheduled for Itanic release in 2004) wouldi, seem to depend only on current capabilities.   >sL > Is so, that stating that IA64 isn't ready for prime time would not be such ad > moronic thing.  K But that's not the statement Fred was responding to.  Nor is it necessarilyiJ a reasonable one in any event:  the fact that Itanic shows no sign of everC having had the potential to become competitive with Alpha had AlphaeD development continued doesn't mean that it can't obtain some kind ofK foothold in the marketplace:  it just means that choosing it over Alpha was I a stupid decision for both Compaq and Intel (though Intel had the problemmF that its partner HP was counting on a compatible PA-RISC replacement).  E Of course, there still seems to be some question about whether actual F production Itanic2 systems are generally available (I've heard severalI people claim they've tried to obtain them from HP and been told they were L back-ordered, and have heard no one state that they've actually succeeded inI acquiring one through normal channels) - so in *that* sense they may welltI not be ready for prime time (and of course only HP and perhaps a Far EastnG supplier or two are even claiming to offer them yet:  Intel's and IBM'stB supporting chipsets and systems won't hit the market for a while).   - bill   ------------------------------  % Date: Mon, 07 Oct 2002 15:20:45 -0400e- From: JF Mezei <jfmezei.spamnot@videotron.ca>e@ Subject: Re: HP tells VMS users "Relax, you are sitting pretty"!, Message-ID: <3DA1DE8B.370DBDC8@videotron.ca>   Fred Kleinsorge wrote: > G > I'd hate to insult morons by calling you one.  You claimed that HP-UXv0 > requires Itanium-2, which is completely false.  = Is it possible that IA64 version 1.0 was only good enough for5N workstation/small systems and that they have to wait for the version II before< they can start thinking about building large scale machines?  I Isn't it also true that they have to wait for another generation of IA64sCH before they can think of building the larger VMS machines that some hopeM customers will migrate to ? Is IA64 ready to run Tandem's NSK or do they alsoo& have to wait for the next generation ?  L Is so, that stating that IA64 isn't ready for prime time would not be such a moronic thing.   ------------------------------  % Date: Mon, 07 Oct 2002 17:57:27 -0400 - From: JF Mezei <jfmezei.spamnot@videotron.ca>r@ Subject: Re: HP tells VMS users "Relax, you are sitting pretty"!, Message-ID: <3DA20344.423ED67F@videotron.ca>   Bill Todd wrote:M > It had better not be, because from all current indications there will be no K > substantive changes in Itanic (save for process-shrinks and on-chip cacheeE > expansion) until 2005 (and likely later for any real core changes).   L But is it possible that Intel intends to produce a IA64 model that is betterK suited for enterprise computing (same core, but better suited for multi-cpue systems etc  ?)e  L In other words, could HP produce today a wildfire/marvel class machine basedJ on Itanium 1.0 (Merced) or now on Itanium II ? or does it need to wait forH some new version of it with to make the chip suitable for such systems ?   ------------------------------  $ Date: Mon, 7 Oct 2002 18:38:59 -0400* From: "Bill Todd" <billtodd@metrocast.net>@ Subject: Re: HP tells VMS users "Relax, you are sitting pretty"!6 Message-ID: <SdqcnTNeXZNukT-gXTWcow@News.GigaNews.Com>  : "JF Mezei" <jfmezei.spamnot@videotron.ca> wrote in message& news:3DA20344.423ED67F@videotron.ca... > Bill Todd wrote:L > > It had better not be, because from all current indications there will be noG > > substantive changes in Itanic (save for process-shrinks and on-chip  cacheiG > > expansion) until 2005 (and likely later for any real core changes).  > G > But is it possible that Intel intends to produce a IA64 model that is  betterC > suited for enterprise computing (same core, but better suited foru	 multi-cpun > systems etc  ?)a  D Not by any indication.  IIRC Intel is claiming that both Madison andB Montecito processors will be able to plug into McKinley sockets asL upward-compatible replacements, and while this would not absolutely precludeL having included, say, some dummy pins on McKinley that could be used for newE purposes later (such that Madison or Montecito could have a differenttH relationship with a new chipset than with a McKinley chipset) there's no& indication that this in fact happened.   >sH > In other words, could HP produce today a wildfire/marvel class machine based-0 > on Itanium 1.0 (Merced) or now on Itanium II ?  E I believe it could, though why one would even think of doing so using J Itanic1 is unclear.  In fact, IIRC one of the Far Eastern companies (NEC?)I is already demoing (and may even be selling, or will soon) a 32-processorp McKinley-based system.  J But there probably won't be any Marvel-class Itanic systems before 2005 atE the earliest, because there's no indication that Itanics will acquireeE on-chip routing and memory controllers before then - and without themrK calling a system organized like Wildfire a 'Marvel-class' system would be arG definite stretch.  While Itanic may continue to nip at Alpha's heels inlL SPECint2K during this period (which considering that Alpha will be a processK generation behind it and the basic EV6 core will be well into school-age byCA then is not particularly impressive), when you start getting intosK high-processor-count SMP-like systems there's no way Itanic will be able to I compete with EV7's fast on-chip inter-processor communications (and EV7's E very low memory access latency will also be a differentiator for somed" applications in all environments).   - bill   ------------------------------  # Date: Tue, 08 Oct 2002 01:26:22 GMT # From: "John Smith" <a@nonymous.com> @ Subject: Re: HP tells VMS users "Relax, you are sitting pretty"!J Message-ID: <2vqo9.204699$8b1.127678@news01.bloor.is.net.cable.rogers.com>  5 "Bill Todd" <billtodd@metrocast.net> wrote in message,0 news:JqadnWBnmeT6ajygXTWc3w@News.GigaNews.Com... >  >aG > Assuming that's the case (and assuming that EV79 doesn't get canned),e ItanicF > will remain inferior in most performance respects to Alpha over that period,aE > but not so much inferior that a customer couldn't use it in a large  machineaK > if s/he wanted to (after all, older Alphas with less performance than thewH > new Itanics will have were usable in large configurations - the result justL > wouldn't be state-of-the-art today, and thus it's not clear why a customer0 > would choose Itanic over Alpha in such cases).  F Maybe lower support/licensing costs? That was very typical of why manyI companies upgraded to the next genration VAXen and Alphas over the years.    ------------------------------  % Date: Mon, 07 Oct 2002 20:29:28 -0000l- From: wspencer@ap.nospam.org (Warren Spencer)h% Subject: Re: HP to push .NET big time.5 Message-ID: <92A0AFB6Dwarrenspencer1977@216.168.3.30>I  * mathog@caltech.edu (David Mathog) wrote in! <3D91D6DC.23A8E7FB@caltech.edu>:     >"Terry C. Shannon" wrote: >>  / >> <lbohan@dbc..spamless..com> wrote in message 5 >> news:opf0puc690g4ntiotr8ab8m7tv52gi7kmi@4ax.com...I1 >> > On Tue, 24 Sep 2002 00:53:42 -0400, JF Mezein* >> > <jfmezei.spamnot@videotron.ca> wrote: >> >H >> > >  HP, based in Palo Alto, Calif., and Redmond-based Microsoft plan >> > >  to   >> spendF >> > >about $25 million  each, said Rick Fricchione, vice president of
 >> Enterprise % >> > >Ready Microsoft Services at HP.v >> >B >> > I'd wonder why HP wants to donate $25M towards a MSFT effort;A >> > (You'd think Billy's pockets would be deep enough to buy the  >> >  whole lunch? ...)  >> >A >> > Terry Shannon's remarks inre MSFT organ donors come to mind; * >> > that much hasn't changed, apparently. >>  7 >> Same concept, diffferent donor. But that might be anm@ >> oversimplification as both firms are investing in the effort. >x; >I've seen this show twice already and hardly need to watch,> >this tired plot unfold yet again.  Remember when Digital made@ >a big show of how it was training all those MSCEs?  Deja vu all< >over again.  Much of the cast has even be retained from the >two earlier efforts (sic).  >  >Here are the Cliff notes: >o3 >1.  Company is infected with managerial parasites.s> >2.  Infected company loses ability to make rational decisionsB >    and hands over intellectual capital and markets to Microsoft.E >3.  Remains of the infected company are devoured by other companies.I< >4.  Parasites move on to kill other companies.  (EspeciallyC >    the one foolish enough to have ingested the most pieces of thep  >    previous infected company.) >?	 >Regards,  >m
 >David Mathoga >mathog@caltech.edud >n  J So it's not so much that MS products are prone to viruses; instead you're L asserting that Microsoft, in and of itself, is a virus?  Interesting theory  indeed!<   ws   -- n   Warren Spencer' Senior Software Engineer (not a writer)m The Associated Press  < ** Time flies like an arrow.  Fruit flies like a bananna. **   ------------------------------  % Date: Mon, 07 Oct 2002 20:31:48 -0000l- From: wspencer@ap.nospam.org (Warren Spencer)s% Subject: Re: HP to push .NET big timet5 Message-ID: <92A0AEC6Ewarrenspencer1977@216.168.3.30>"  & John@mvpsi.com (John Vottero) wrote in% <up3q9mp38gchec@news.supernews.com>:    ; >"JF Mezei" <jfmezei.spamnot@videotron.ca> wrote in messagea' >news:3D914D5E.A2312D3F@videotron.ca...d >> "Terry C. Shannon" wrote:F >> > Color me stupid, but might this be a way for Microsoft to achieve= >> > its ambition of RENTING software rather than selling it?c >>D >> But isn't that the main goal of .NET to begin with ? I had always >understood:% >> that this was the purpose of .NET.e >gI >That is a common misconception.  .NET is a new development environment. v; >New from the ground up.  I wish VMS had something similar.a >n  L New to Microsoft, perhaps.  Many few it as a less mature knock-off of Java,  and I suspect I'm one of them.   ws   --     Warren Spencer' Senior Software Engineer (not a writer)h The Associated Press  < ** Time flies like an arrow.  Fruit flies like a bananna. **   ------------------------------  $ Date: Mon, 7 Oct 2002 15:56:11 -0400; From: "Brian Tillman" <tillman_brian@notnoone.notnohow.com>e! Subject: Re: mouning floppy drivec$ Message-ID: <3da1e720$1@news.si.com>  ' >Then what you probably want is mtools.d  H or MGPCX from http://vms.process.com/ftp/vms-freeware/FILESERV/MGPCX.ZIP --  A Brian Tillman                   Internet: tillman_brian at si.comRA Smiths Aerospace                          tillman at swdev.si.comp= 3290 Patterson Ave. SE, MS      Addresses modified to preventt< Grand Rapids, MI 49512-1991     SPAM.  Replace "at" with "@"8        This opinion doesn't represent that of my company   ------------------------------  % Date: Mon, 07 Oct 2002 18:35:11 -0000g- From: wspencer@ap.nospam.org (Warren Spencer)  Subject: Re: MySQL for VMS?o5 Message-ID: <92A09040Dwarrenspencer1977@216.168.3.30>   / colive@technologEase.com (Chris Olive) wrote ine3 <b10654c6.0209240752.299a54db@posting.google.com>:    F >Michiel Erens <I.dont.want.spam@this.mailaddress.is.invalid> wrote in= >message news:<3D8F6926.3FB0@this.mailaddress.is.invalid>... eG >Well, bummer.  The Alpha images were built on VMS v7.2 and I'm runningsC >Alpha VMS v7.1 (I know it's way behind, but I have no control overk< >that), so I'm getting sharable image ident mismatch errors: >a# >Error activating image PTHREAD$RTL ? >Image file OMEGA$DKB5:[SYS0.SYSCOMMON.][SYSLIB]PTHREAD$RTL.EXEe$ >Ident mismatch with shareable image > ( >Unless someone knows a way around this? >aF >I'll try and get ahold of the author, although he hasn't yet answeredG >my FIRST message, so I'm not even sure if he still works at Process ore >not...h >r >Chris  I If you can get your hands on the appropriate version of PTHREAD$RTL.EXE, wI you should be able to INSTALL it on your 7.1 system,and successfuly link  L against it.  No guarantees, however, that the older version of the rtl will $ operate correctly under OpenVMS 7.1.   ws   -- m   Warren Spencer' Senior Software Engineer (not a writer)  The Associated Press  < ** Time flies like an arrow.  Fruit flies like a bananna. **   ------------------------------  $ Date: Mon, 7 Oct 2002 23:22:47 -0400) From: "Neil Rieck" <n.rieck@sympatico.ca>"/ Subject: OpenVMS 7.3-1 and TCPware 5.6 question-8 Message-ID: <dcso9.977$Iw5.211027@news20.bellglobal.com>   Facts:K 1. Every 6 months a batch job on my OpenVMS systems will run to switch botheH the OpenVMS system clock, as well as the TCPware "time zone", "time zoneK name", MULTINET_TIMEZONE (for email stamps), to/from daylight savings time.h 2. I'm not using NTP. 2 3. I occasionally call NTPdate from a login script  
 Questions:I 1. I've set sysgen parameter 'AUTO_DLIGHT_SAV' to 1 so that OpenVMS makes L changing the system clock unnecessary, but is there anything I need to do to TCPware?" 2. Is it capable of auto-changing?  
 Neil Rieck Kitchener/Waterloo/Cambridge,g Ontario, Canada.! http://www3.sympatico.ca/n.rieck/i   ------------------------------  # Date: Mon, 07 Oct 2002 11:57:48 GMT . From: peter@langstoeger.at (Peter LANGSTOEGER)6 Subject: Re: Problem with DEC TCP/IP and about patches3 Message-ID: <0Feo9.19660$M52.213414@news.chello.at>o  Z In article <3DA16344.3050800@beenet.ru>, Konstantin Klubnichkin <kostik@beenet.ru> writes:G >I'm observing strange bug in TCP/IP stack on my OpenVMS 7.2 VAX (from 0: >Hobbyist CD) when I try to telnet to VAX from Linux box .J >After TCP connection is estableshed (SYN-ACK procedure) there is strange E >packet flow between linux (client) and VAX (server). And looks like 9 >output from VMS is buffered somewhere. I mean following:   J If it is TCPIP V5.0-9 (aka V5.0) or V5.0-10 (aka V5.0A) what you run, then@ the problem is well known (a endless loop in negotiation phase).D Use TCPIP V5.0-11 (unfortunately also known as V5.0A and not V5.0B),F V5.1-15 or V5.3-18 and install also the appropriate ECOs afterwards...   --   Peter "EPLAN" LANGSTOEGERt% Network and OpenVMS system specialist  E-mail  peter@langstoeger.atF A-1030 VIENNA  AUSTRIA              I'm not a pessimist, I'm a realist   ------------------------------  # Date: Tue, 08 Oct 2002 00:19:42 GMTv! From: rob.buxton@wcc.spam.govt.nz-6 Subject: Re: Problem with DEC TCP/IP and about patches& Message-ID: <3da22480.2321357525@news>  : On Mon, 07 Oct 2002 17:10:14 +0400, Konstantin Klubnichkin <kostik@beenet.ru> wrote:s   >Peter LANGSTOEGER wrote:a] >> In article <3DA16344.3050800@beenet.ru>, Konstantin Klubnichkin <kostik@beenet.ru> writes:  >> tI >>>I'm observing strange bug in TCP/IP stack on my OpenVMS 7.2 VAX (from e< >>>Hobbyist CD) when I try to telnet to VAX from Linux box .L >>>After TCP connection is estableshed (SYN-ACK procedure) there is strange G >>>packet flow between linux (client) and VAX (server). And looks like i; >>>output from VMS is buffered somewhere. I mean following:g >>   >> bM >> If it is TCPIP V5.0-9 (aka V5.0) or V5.0-10 (aka V5.0A) what you run, then:C >> the problem is well known (a endless loop in negotiation phase).hG >> Use TCPIP V5.0-11 (unfortunately also known as V5.0A and not V5.0B),tI >> V5.1-15 or V5.3-18 and install also the appropriate ECOs afterwards...  >> n > % >So I have to find corresponding ECO?  >Now going to find it...   A very useful site is:  9 http://ftp.digital.com.au/pub/ecoinfo/ecoinfo/version.htma   Rob, >O >Thanks! >: >--  >Best regards, >Konstantin Klubnichkin  >j   ------------------------------  # Date: Mon, 07 Oct 2002 21:25:26 GMTJ$ From: winston@SSRL.SLAC.STANFORD.EDU' Subject: Re: Proprietary or Open Sourcei8 Message-ID: <00A15195.D7CAB2B8@SSRL04.SLAC.STANFORD.EDU>  o In article <Aflo9.154425$q41.77869@news02.bloor.is.net.cable.rogers.com>, "John Smith" <a@nonymous.com> writes:g >tM >"Kevin Handy" <kth@srv.net> wrote in message news:eZko9.2167$G5.1003@fe01.... >>? >> How many times does this article need to be posted to all ofc2 >> the various newsgroups anyway?  Is spamming the> >> list with the same off-topic message over-and-over supposed >> to make it sound better?i >sK >My apologies for posting something that might provide some ammunition thatYI >could be used by VMS users to keep VMS around when the hue and cry is too >'open source'.  >   7 It's like neither of you read the article to the end.     F My impression is that the editor who put on the headline (and probablyJ rewrote the first paragraph) is anti-open-source, pro-Microsoft, while theB author wrote a fairly balanced article including quotes from real,G non-ax-grinding, computer security experts (Denning, Spafford, Neumann)iI and concludes that having open-source/closed-source is irrelevant and theiK important thing is to have a development process where people who know what E they're doing write code and other people who know what they're doing0
 review it.  C If you manage to get people to say "open source sucks" that doesn'tsJ necessarily mean they'll flock to VMS; you're better off pointing out thatJ VMS engineers are porting a lot of "industry standard" open source to VMS.   -- Alan    ------------------------------  $ Date: Mon, 7 Oct 2002 17:31:12 -0400* From: "Bill Todd" <billtodd@metrocast.net>' Subject: Re: Proprietary or Open Sourceo6 Message-ID: <SNKdnbJJrtyNYDygXTWc3Q@News.GigaNews.Com>  . "John Smith" <a@nonymous.com> wrote in messageC news:Aflo9.154425$q41.77869@news02.bloor.is.net.cable.rogers.com...] >(. > "Kevin Handy" <kth@srv.net> wrote in message news:eZko9.2167$G5.1003@fe01...  > >t@ > > How many times does this article need to be posted to all of3 > > the various newsgroups anyway?  Is spamming the8? > > list with the same off-topic message over-and-over supposed  > > to make it sound better? >6L > My apologies for posting something that might provide some ammunition thatJ > could be used by VMS users to keep VMS around when the hue and cry is to > 'open source'.  K No apology required:  Handy's clearly a Linux bigot, since the article (hadtC he bothered to read it in its entirety) was reasonably even-handed.p   - bill   ------------------------------  # Date: Mon, 07 Oct 2002 18:45:39 GMTm# From: "John Smith" <a@nonymous.com>a# Subject: Proprietary or Open SourcebJ Message-ID: <nDko9.201981$8b1.161735@news01.bloor.is.net.cable.rogers.com>  2 http://www.eweek.com/article2/0,3959,562220,00.asp   September 30, 2002' Open Source: A False Sense of Security?    By  Dennis Fisher   K Over the last couple of years, as security vulnerability reports have piled I up on products from such big vendors as Microsoft Corp., Oracle Corp. and G Cisco Systems Inc., open-source advocates have snickered. If only those'L vendors would release their source code and let the open-source community atJ it, all their problems would go away, they said. And when the Code Red andG Nimda worms chewed their way through hundreds of thousands of unpatched K Microsoft Internet Information Services servers last year, Apache users sat=B back and smiled, believing nothing like that could happen to them.   Then it did.  L In late July, researchers found several flaws in the OpenSSL tool kit, whichL is commonly used for secure transmissions on Apache servers. About six weeks@ later, someone released a worm called Slapper that exploited theL vulnerability and not only installed a back door on each infected server butD also turned machines using OpenSSL into a waiting army of zombies byD dropping in a DDoS (distributed-denial-of-service) tool kit as well.  E The infected machines can communicate with one another via a private, J peer-to-peer network. Security experts predicted that it was only a matterL of time before someone used the thousands of compromised servers to launch a devastating DDoS attack.  I Despite the mantra that open-source software is more secure thanks to itseL communal writing and review process, the vulnerabilities in OpenSSL were allI buffer overruns, the most common and, many say, most preventable flaws in 	 software.a  F That such flaws were found in an open-source tool kit and subsequentlyL exploited by a destructive worm comes as no surprise to some experts. Still,L it's enough to prompt some to question the long-held belief that open-source software is more secure.  L "Linux is awful. There are no design specs. Everybody and their half-brotherK who knows some [C code] writes code for it, and they all have the same lack J of knowledge," said Gene Spafford, professor of computer science at PurdueG University, in West Lafayette, Ind., and an expert on network security.tK "It's who writes it and whether it's planned [that makes a difference], notr who looks at the code."r  L Despite such rumblings, however, few open-source believers are ready to dropE Linux or other open-source products because of newly spawned securityeH concerns. Mike Prince, for example, thought long and hard about securityG before deciding, in 1999, to roll out Linux companywide to thousands of J users in hundreds of locations across the country. By the time Prince madeL the call, however, the CIO at Burlington Coat Factory Warehouse Corp. had no1 doubts about the reliability of the new software.)  H As a longtime user of a variety of back-office open-source applications,H Prince said he believed the security of the software was a given. And he hasn't changed his mind.  D "The security of the open-source software hasn't been an issue. It'sI excellent," said Prince, at Burlington's headquarters in Burlington, N.J.hL "On the operating system side, although there are loopholes found, the speedK with which they're fixed and the commitment to making the problem known andaH resolved are excellent. The stability rivals the best of the proprietaryK Unix systems. The whole security model in Linux is better than in Windows."   E So who's right? Does patent-protected development behind closed doorsoD produce more secure software? Or does the collaborative, open-sourceI community, where thousands of smart, independent developers are poised toh spot and fix security problems?:  I Many IT managers and security experts say it's not that simple. Security,tF they insist, comes down to attention to detail and careful coding, notL whether the code is freely available on the Internet or locked in a vault on a corporate campus.l  F "Unless there's a great deal of discipline underlying the development,E there's no difference in the security [of proprietary and open-sourceIJ software]. Open source is not inherently more secure," said Peter Neumann,F principal scientist at SRI International, in Menlo Park, Calif., and aL security and networking expert who in 1965 helped design the file system forF Multics, which is still considered one of the most secure and reliableI operating systems ever written. "If everyone has the same bad skills, allFH the eyeballs in the world won't help you. Unless there's discipline, you still come up with garbage."  K Advocates of Linux and other open-source software often cite users' abilityeK to modify the code and adapt it to their environments as a key advantage ofsG open-source applications. However, that can be a drawback if the peoplec, doing the modifications aren't well-trained.  K Some devotees say the real strength of open source lies in its transparency ' and the flexibility it gives customers.   K "The transparency gives you security because you can pick and choose what'saH in your environment," said John Alberg, co-founder and vice president ofC engineering at Employease Inc., an Atlanta-based developer of humantC resources software and a user of numerous open-source applications.a  H "Commercial software tends to have a lot of doors you don't know about,"H Alberg said. "What open source does is allow you to manage a more secureJ environment. There are fewer moving parts in the products, and, hence, you have fewer problems."i  D "Open-source software is developed by people who are more attuned toH security. Commercial software vendors are trying to hit feature sets andH target dates," said Dan Agronow, vice president of technology at WeatherJ Channel Enterprises Inc.'s Weather.com site, in Atlanta, which uses Linux,K Apache and other open-source software. "With open source, it isn't releasedPG until it's ready, and that's it. But we still pay a lot of attention to  security. You have to."v  K To the extent that open-source products such as Linux still suffer securitynF holes, however, they may soon get help from a small number of startups, dedicated to hardening the operating system.  K Guardian Digital Inc., of Allendale, N.J., recently released EnGarde SecureoL Linux Professional, which features a litany of added security functionality,F such as a network gateway firewall, a network IDS (intrusion detectionH system) and a host IDS, and a security control center. Even the NationalL Security Agency, of Fort Meade, Md., has gotten in on the act, producing its) own Security Enhanced Linux distribution.   H For as much criticism as Microsoft takes for the lack of security in itsJ products, some Linux distributions have begun to experience more problems.D Red Hat Inc., of Raleigh, N.C., for example, has issued fixes for 35J security problems in its Red Hat Linux 7.3 since June, while Microsoft, ofL Redmond, Wash., has released six patches in the same time period for WindowsK XP Pro. However, the list of patches included in the new Service Pack 1 for I XP Pro shows 30 security-related fixes, including several that were neverF  publicized or issued separately.  I But, some observers say, comparisons of bug reports simply prove that all.C software is insecure. The real determinant of security is competent.& programming and code review, they say.  J "I don't think it's a good idea to have one rule as to whether code shouldG be open. If Microsoft opened the [Internet Explorer] code now, it wouldlJ probably be very bad because it's full of all kinds of bugs. But if it hadG been open from the start, that would have been good," said Avi Rubin, aiF principal researcher in the secure systems research department at AT&T$ Labs-Research, in Florham Park, N.J.  I "Apache is a good example. Anything like that that has a formal structure I and people working on it is good," Rubin said. "Part of the beauty of theHL open-source process is that they take into account that vulnerabilities willC happen, so they're prepared for it. The people making decisions are : responding out of pride, not from a business perspective."  I Indeed, the response to open-source software security problems that Rubin C has experienced is one of the things that convinced Burlington Coat.E Factory's Prince that the open-source community was more dedicated ton! security than commercial vendors.P  F Prince once found a bug in an open-source operating system utility andI posted a question about it to a newsgroup. The author of the utility soonhI replied, confirming the problem, telling Prince how to work around it and2H saying he had a new version of the utility on the way that would fix the bug.  H "That's what open source does. They have brilliant people who, once theyJ understand the problem, are probably in competition with each other to fixI it," Prince said. "There hasn't been a minute of time wasted being jerkedi around."  F However, even hard-core advocates of open-source software concede thatL simply making source code available doesn't make an application more secure.  H "What really makes a difference is having someone who knows what they'reJ doing writing the code and looking at the code," said Crispin Cowan, chiefI scientist at WireX Communications Inc., in Portland, Ore., a developer of F secure Linux solutions. "But I think that the open-source process does enable greater security."l   ------------------------------  # Date: Mon, 07 Oct 2002 19:28:32 GMT # From: "John Smith" <a@nonymous.com>.' Subject: Re: Proprietary or Open SourcepI Message-ID: <Aflo9.154425$q41.77869@news02.bloor.is.net.cable.rogers.com>O  L "Kevin Handy" <kth@srv.net> wrote in message news:eZko9.2167$G5.1003@fe01... >A> > How many times does this article need to be posted to all of1 > the various newsgroups anyway?  Is spamming the = > list with the same off-topic message over-and-over supposede > to make it sound better?  J My apologies for posting something that might provide some ammunition thatH could be used by VMS users to keep VMS around when the hue and cry is to 'open source'.   ------------------------------  % Date: Mon, 07 Oct 2002 13:22:48 -0600  From: Kevin Handy <kth@srv.net> ' Subject: Re: Proprietary or Open Sourceo% Message-ID: <eZko9.2167$G5.1003@fe01>r   John Smith wrote:n4 > http://www.eweek.com/article2/0,3959,562220,00.asp >  > September 30, 2002) > Open Source: A False Sense of Security?e >  > By  Dennis Fishere  - [snip long anti-linux/pro-Microsoft diatribe]:  = So, since Linux has had one virus written for it, for which at< patch existed before the virus did, we should now switch all? Linux boxes to the much safer MicroSoft Windows(tm), which onlyi8 has 70,000+ virus (and rapidly increasing), because they+ obviously do a much better job at security.e  < How many times does this article need to be posted to all of/ the various newsgroups anyway?  Is spamming then; list with the same off-topic message over-and-over supposeda to make it sound better?  ; Obviously he thinks it is such an important message that heM) has to hide behind a fake e-mail address.e   ------------------------------  % Date: Mon, 07 Oct 2002 16:00:37 -0600e From: Kevin Handy <kth@srv.net>I' Subject: Re: Proprietary or Open Source & Message-ID: <3DA20405.7010603@srv.net>   John Smith wrote: N > "Kevin Handy" <kth@srv.net> wrote in message news:eZko9.2167$G5.1003@fe01... > > >>How many times does this article need to be posted to all of1 >>the various newsgroups anyway?  Is spamming thee= >>list with the same off-topic message over-and-over supposedo >>to make it sound better? >  > L > My apologies for posting something that might provide some ammunition thatJ > could be used by VMS users to keep VMS around when the hue and cry is to > 'open source'.  A I don't see how this article is very good ammunition against OpenhA Source. If you ignore all the ranting, it just says that there iskC ONE virus that can attack SOME unpached Linux systems, and claiming 9 that this proves that Open Source is insecure. Microsoft,6? on the other hand, is wonderfully secure because it is not Open0= Source. Blah! Lets just ignore the 70,000+ virus that Windows-? blissfully runs as insignificiant against this one Linux virus.   C It states many good points about Open Source, including many peoples@ who have had good experiences with it, but immediately dismisses each one with a snide comment.  A I don't see how you could show this article to someone and really8; expect them to switch to VMS because of anything in it.  It 8 sounds like a MicroSoft fluff piece against Open Source:  9 	"Sure you can get things fixed by the author in an hour,a4 	and you get to see the source code, and you can fix; 	problems yourself if you really need to, and it's cheaper,e3 	and it works better, and is more secure from virusn4 	attacks; but ours isn't Open Source and we would be7 	buried by even more virus if we made it so, thus we'rea 	obviously a better choice."  C Would an article like that convince you to switch from Open Source?tB Isn't that what this article says?  First time I read it I thought- it was someones idea of a joke.  Maybe it is.b   ------------------------------  % Date: Mon, 07 Oct 2002 16:48:34 -0600  From: Kevin Handy <kth@srv.net>s' Subject: Re: Proprietary or Open Source % Message-ID: <8_no9.2239$G5.1217@fe01>s   Bill Todd wrote:0 > "John Smith" <a@nonymous.com> wrote in messageE > news:Aflo9.154425$q41.77869@news02.bloor.is.net.cable.rogers.com...( > . >>"Kevin Handy" <kth@srv.net> wrote in message > ! > news:eZko9.2167$G5.1003@fe01...e > ? >>>How many times does this article need to be posted to all of 2 >>>the various newsgroups anyway?  Is spamming the> >>>list with the same off-topic message over-and-over supposed >>>to make it sound better?e >>L >>My apologies for posting something that might provide some ammunition thatJ >>could be used by VMS users to keep VMS around when the hue and cry is to >>'open source'. >  > M > No apology required:  Handy's clearly a Linux bigot, since the article (hadeE > he bothered to read it in its entirety) was reasonably even-handed.l  > Wasn't trying to sound anti-VMS here, just tired of repeatedly? seeing this article plastered in every newsgroup with a commentlE about how it clearly proves Closed Source is superior to Open Source.eA I can't see why they think so highly of this article in that way,e? especially with the last line "But I think that the open-sourcee& process does enable greater security".  : I'll use whichever OS looks like it will work better for a< particular job, be it Windows, Linux, BSD, VMS, CP/M, Mumps,4 Plan9, Pick, OS/278, RSTS/E, DOS, MS-DOS, Emacs, ...   ------------------------------  $ Date: Mon, 7 Oct 2002 16:52:41 -0700# From: "Tom Linden" <tom@kednos.com>o' Subject: RE: Proprietary or Open Sourcen9 Message-ID: <CIEJLCMNHNNDLLOOGNJICEPBFNAA.tom@kednos.com>    >-----Original Message----- ' >From: Kevin Handy [mailto:kth@srv.net]r' >Sent: Monday, October 07, 2002 3:49 PMa >To: Info-VAX@Mvb.Saic.Com( >Subject: Re: Proprietary or Open Source >m >  >Bill Todd wrote:w1 >> "John Smith" <a@nonymous.com> wrote in messagetF >> news:Aflo9.154425$q41.77869@news02.bloor.is.net.cable.rogers.com... >> n/ >>>"Kevin Handy" <kth@srv.net> wrote in messagei >> m" >> news:eZko9.2167$G5.1003@fe01... >> r@ >>>>How many times does this article need to be posted to all of3 >>>>the various newsgroups anyway?  Is spamming the ? >>>>list with the same off-topic message over-and-over supposed- >>>>to make it sound better? >>>:> >>>My apologies for posting something that might provide some  >ammunition thatK >>>could be used by VMS users to keep VMS around when the hue and cry is tov >>>'open source'.d >> n >>  B >> No apology required:  Handy's clearly a Linux bigot, since the 
 >article (hadvF >> he bothered to read it in its entirety) was reasonably even-handed. >A? >Wasn't trying to sound anti-VMS here, just tired of repeatedlym@ >seeing this article plastered in every newsgroup with a commentF >about how it clearly proves Closed Source is superior to Open Source.B >I can't see why they think so highly of this article in that way,@ >especially with the last line "But I think that the open-source' >process does enable greater security".o >n; >I'll use whichever OS looks like it will work better for ae= >particular job, be it Windows, Linux, BSD, VMS, CP/M, Mumps, 5 >Plan9, Pick, OS/278, RSTS/E, DOS, MS-DOS, Emacs, ...-  D Emacs is big and powerful, but calling it an OS, is maybe stratching it :-)   >. >m >---' >Incoming mail is certified Virus Free.s; >Checked by AVG anti-virus system (http://www.grisoft.com).2A >Version: 6.0.394 / Virus Database: 224 - Release Date: 10/3/2002e >s ---m& Outgoing mail is certified Virus Free.: Checked by AVG anti-virus system (http://www.grisoft.com).@ Version: 6.0.394 / Virus Database: 224 - Release Date: 10/3/2002   ------------------------------  $ Date: Mon, 7 Oct 2002 16:48:19 -0400. From: "Jerry Alan Braga" <jabraga@flanagan.ca>< Subject: Tally Printer causing overflow data underrun errors3 Message-ID: <Xqmo9.110324$C8.313585@nnrp1.uunet.ca>e  I I have a tally MT691 printer that is configured for telnet based printingoJ and when you place the printer off-line for a period of time while data isH being sent or when it runs out of paper and then you reset it causes the
 same problem.7  H We have genicom printers on the same terminal server DECServer 900TM andJ they are running fine,  and to the point as when I plug into the tally theL working cable of the genicom printer (port settings and telnet listeners areI identical) it also has the same problems.  So this must be something with 8 the terminal server and this particular type of printer.   Settings on portsl  5 Port 24: KIT_PRT_11                    Server: TKIT02c  F Character Size:            8           Input Speed:              19200F Flow Control:            XON           Output Speed:             19200F Parity:                 None           Signal Control:        DisabledG Stop Bits:           1                  Signal Select:  CTS-DSR-RTS-DTR?  F Access:               Remote           Local Switch:              NoneF Backwards Switch:       None           Name:                   PORT_24F Break:              Disabled           Session Limit:                1F Forwards Switch:        None           Type:                      AnsiF Default Protocol:     Telnet           Default Menu:              NoneF Autolink Timer One:10 Two:10           Dialer Script:             None  " Dedicated Service: REN.FLANAGAN.CA Authorized Groups:   0 (Current)  Groups:   0   Enabled Characteristics:E Autoconnect,  Autoprompt,  Broadcast,  Failover,  Input Flow Control,I? Lock,  Loss Notification,  Message Codes,  Output Flow Control,g Verification  K Printer is setup for XON/XOFF flow with the proper baud rates / stop bits /r parity  G I have tried on the printer RDY/BSY flow but the same problem is there.d   Any ideas ?    -- Jerry Alan Braga Senior Programmer Analystt Flanagan Foodservice (519) 748-2190 Ext. 3361 jabraga@flanagan.caI   ------------------------------  % Date: Mon, 07 Oct 2002 16:37:27 -0400a& From: David M Smith <dsmit115@csc.com>2 Subject: Re: TPU access to DCL symbols or logicals' Message-ID: <ansra7$evc$1@lore.csc.com>o   JF Mezei wrote:p  G > Looked in Google, and it seemed that back in the early 1990s, DigitalwT > engineers were aware of a big need to give TPU procedures access to logical names. > S > Current documentation isn't very obvious  if such a feature was added since then.s >   H I found a pretty nifty collection of code and procedures which implementH a "general-purpose" CALL_USER routine which is capable (according to theB author) of calling any shareable image routine. I found it on Andy' Harper's OpenVMS VMS_SHARE archive, at:-  8         http://www.agh.cc.kcl.ac.uk/files/vms/vms_share/  F and the name is tpu_calluser.PACKAGE. (If you need help with VMS_SHARE@ format, ask a separate question about it.) You will need a BASICE compiler if you want to modify and/or recompile the code, but I foundd> the image file included works on my OpenVMS V7.3 system as is.  D After defining the TPU$CALLUSER logical name to refer to this image, e.g.  6 $ define TPU$CALLUSER disk:[directory]TPU_CALLUSER.EXE  F you can then use a TPU command similar to the following within TPU (myF example is from standard TPU EVE$SECTION with SET KEYPAD EDT). I press< GOLD KP7 to get a Command: prompt, then enter the following:  G Command: TPU EXECUTE (CALL_USER(0,'LIB$GET_LOGICAL "SYSTEM1", TEST$'));   H This says "execute the TPU command CALL_USER passing the arguments 0 (asC explained in the README file, this doesn't appear to be modified ora@ needed) and a call to the LIB$GET_LOGICAL RTL routine to get theD equivalence string associated with the logical name SYSTEM1 (in thisD case) and return it to the TPU string symbol TEST (the suffix '$' isF BASIC convention for a string symbol; it took me awhile to figure thatE out!). Then, I can subsequently retrieve the value by referencing the@9 TPU symbol TEST. As demonstration, doing another command:n   Command: TPU COPY_TEXT(TEST);:  < inserts the equivalence string into the current text buffer.I -------------------------------------------------------------------------mE David M. Smith 302.391.8533                       dsmit115 at csc dot. comoC Computer Sciences Corporation     (Opinions are those of the writer  only)mI -------------------------------------------------------------------------e   ------------------------------  $ Date: Mon, 7 Oct 2002 16:11:40 -0400; From: "Brian Tillman" <tillman_brian@notnoone.notnohow.com>s Subject: Re: VMS 7.3 DCL / MXe$ Message-ID: <3da1eac2$1@news.si.com>  & >That's interesting - what sets them ?   SYLOGICALS.COM, of course.  J >The documentation suggests that the mail$protocol_PREFIX is needed for an0 >alternate prefix (ie, not MX%). We've just got: >,! >"MAIL$INTERNET_TRANSPORT" = "MX"e >m) >"MAIL$PROTOCOL_GFAX" = "GOLDFAX_MAILSHR"n  I We found some behavioral issues with mail when we installed VMS V7.2 thatd< using the MAIL$PROTOCOL_xxx logicals and doing away with the& MAIL$INTERNET_TRANSPORT logical fixed. --A Brian Tillman                   Internet: tillman_brian at si.comhA Smiths Aerospace                          tillman at swdev.si.comr= 3290 Patterson Ave. SE, MS      Addresses modified to prevente< Grand Rapids, MI 49512-1991     SPAM.  Replace "at" with "@"8        This opinion doesn't represent that of my company   ------------------------------   End of INFO-VAX 2002.555 ************************