1 INFO-VAX	Thu, 22 Sep 2005	Volume 2005 : Issue 530       Contents:4 Re: freeware openVMS plugin for Microsoft Virtual PC Kea and VMS ' Re: Parent procedures and f$environment ' Re: Parent procedures and f$environment ' Re: Parent procedures and f$environment A Re: Pathworks for the hobbyist, and general pricing of VMS things 0 Poor I/O performance (was poor SCSI performance) Remote X Session to PC thru SSH # Re: Remote X Session to PC thru SSH # Re: Remote X Session to PC thru SSH # Re: Remote X Session to PC thru SSH # Re: Remote X Session to PC thru SSH / Strange SMTP problem receiving from MS-exchange 5 TCP/IP Config Setting to set Network speed to 100Mbps 9 Re: TCP/IP Config Setting to set Network speed to 100Mbps 9 Re: TCP/IP Config Setting to set Network speed to 100Mbps 9 Re: TCP/IP Config Setting to set Network speed to 100Mbps C Re: Tentative DCL suggestion: string return value from .COM (Hello, C Re: Tentative DCL suggestion: string return value from .COM (Hello, O Re: Tentative DCL suggestion: string return value from .COM (Hello, Guy Peleg?)   F ----------------------------------------------------------------------  % Date: Thu, 22 Sep 2005 09:28:51 +0200  From: S <soterroatyahoodotcom>= Subject: Re: freeware openVMS plugin for Microsoft Virtual PC & Message-ID: <43325d30$1@news1.ethz.ch>   Mister Scary wrote: K > They sent me a license, but the only way I see to get the software is to   > spend $30 on that CD.   I Or get a perfectly legal copy from anybody who has it already. Where are   you located?   S    ------------------------------    Date: 22 Sep 2005 08:42:22 -0700 From: contracer11@gmail.com  Subject: Kea and VMSC Message-ID: <1127403742.948177.221490@o13g2000cwo.googlegroups.com>   D I'm using Kea for VAX emulator to access our VAX/VMS system. Can you> tell me how ( in a procedure) can I get default background andE foreground color ? I need create a procedure to get default FG and BG 
 KEA's color.   thanks.    ------------------------------  % Date: Thu, 22 Sep 2005 08:47:15 +0100 0 From: Chris Sharman <chris.sharman@sorry.nospam>0 Subject: Re: Parent procedures and f$environment4 Message-ID: <dgtni3$3at$1$8302bc10@news.demon.co.uk>   Jeff Cameron wrote: 7 > Can anyone help me with this DCL programming request?  > L > F$ENVIRONMENT("PROCEDURE") returns the file specification of the currently, > executing command procedure. That is cool! > H > F$ENVIRONMENT("DEPTH") returns the integer depth of nested DCL command
 > procedures.  > L > Given that your depth is 6, how can you find the file specification of theN > parent calling procedure (level 5), and it's parent (level 4), and on up the > tree to level 1?  E Here it is, from here, to here - there may be a less mangled version  6 around somewhere, since I mangled this to suit myself.H Works on VAX & Alpha, could doubtless be easily made to work on Itanium.   Chris    $ vfy = f$verify('debug'+0)  $! $! PROCEDURE: GET_CALLER.COM $! $! ABSTRACT: $!L $!    This code fragment obtains the name of the procedure which called this $!    procedure. $!/ $!From: Ken Bosward <bosward.ken.kw@bhp.com.au>  $!Newsgroups: comp.os.vms  $!Subject: DCL Tricks ' $!Date: Fri, 01 Nov 1996 12:35:58 +1000 ( $!Message-ID: <3279620E.3091@bhp.com.au> $! $! HISTORY:  $!/ $!    15-MAY-1993  MDC  Written by Michael Clay 7 $!     1-Nov-1996  Ken Bosward, clean up temporary file = $!    06-Nov-1996  Dave Cantor, shorten report code using !AC + $!                      Added nesting level B $!    20-JUL-2000 Chris Sharman, Alpha-ize (without breaking VAX),A $!                      add parms to suppress self (P1 boolean) & I $!                      make global symbols get_caller__N (if p1 or p2 =   S*),G $!                      where N=0 is this proc, 1 is calling proc, and   so on;K $!                      ie inverse of depth: self-relative seemed more use.  $!O $!----------------------------------------------------------------------------- 6 $  fac =  f$parse(f$environment("procedure"),,,"name") $  pid =  f$getjpi(0,"pid") * $  temp =  "sys$scratch:''fac'_''pid'.tmp" $  on warning then goto FINISH  $  on control_y then goto FINISH $!! $! Define known symbol values ...  $!" $  arch := 'f$getsyi("arch_name")' $  vax = "!" $  alpha = "!" $  'arch' :=@ $  'vax'ctl$ag_clidata =  %x7FFE2A5C   ! From SYS$SYSTEM:SYS.STB% $  'alpha'ctl$ag_clidata = %x7ffce260 > $  ppd$l_prc      =  %x00000008   ! From SYS$SYSTEM:DCLDEF.STB> $  idf_l_lnk      =  %x00000000   ! From SYS$SYSTEM:DCLDEF.STB> $  idf_l_filename =  %x00000068   ! From SYS$SYSTEM:DCLDEF.STB $!F $! Get value of PRC_L_IDFLNK from DCLDEF.STB (changes with version of  VMS) ... $!+ $  call fetchsym_'arch' dcldef prc_l_idflnk  $! $! Traverse IDF list $! $  ppd =  ctl$ag_clidata3 $  prc =  f$cvui(0,32,f$fao("!AD",4,ppd+ppd$l_prc)) 6 $  idf =  f$cvui(0,32,f$fao("!AD",4,prc+prc_l_idflnk))  $  lev =  f$environment("depth") $  baselev = lev? $  makesym = (f$ext(0,1,p1).eqs."S").or.(f$ext(0,1,p2).eqs."S")  $  if p1 then $ goto tonext  $next_filename: $ $     if idf .eq. 0 then goto FINISHL $     procname = f$fao("!AC",f$cvui(0,32,f$fao("!AD",4,idf+idf_l_filename))) $     if makesym6 $     then get_caller__'f$str(baselev-lev) == procname> $     else write sys$output f$fao(" !2UL -> !AS",lev,procname) $     endif = $tonext:     idf =  f$cvui(0,32,f$fao("!AD",4,idf+idf_l_lnk))  $     lev = lev - 1  $     goto next_filename $!& $fetchsym_alpha: subroutine ! file symH $  search sys$loadable_images:'p1'.stb 'p2' /exa/form=passall/out='temp' $! record too big for dcl E $  set file/attr=(rfm:fix,mrs:8,lrl:8) 'temp' ! Quadword aligned (+2)  $  sym = "   " + p2  $  sym[0,16] = 0 $  sym[16,8] = f$length(p2) 
 $  ver = 1
 $  before = 3  $  after = (f$len(sym)-1)/8 : $symloop: ! search for each quadword of the symbol in turn( $  sym8 = """"+f$ext(8*ver-8,8,sym)+"""" $  siz = f$len(sym8)-2 $  sea 'temp';'ver' &sym8 I /key=(pos:1,siz:'siz')/exa/form=passall/win=('before','after')/out='temp'  $  ver = ver + 1 $  before = before + 1 $  after = after - 1$ $  if after.ge.0 then $ goto symloop $! $  open/read temp 'temp' $  read temp valquad
 $  close temp  $  delete/nolog 'temp';*! $  'p2' ==  f$cvui(16,32,valquad)  $  endsubroutine $!$ $fetchsym_vax: subroutine ! file symG $  search sys$system:'p1'.stb 'p2' /exact/format=passall /output='temp'  $  open/read temp 'temp' $  read temp symbols
 $  close temp  $  delete/nolog 'temp';*' $  loc_sym =  f$locate("''p2'",symbols) , $  'p2' ==  f$cvui((loc_sym-5)*8,32,symbols) $finish:  endsubroutine  $! $FINISH:
 $ set noonE $ if f$type(prc_l_idflnk).nes."" then $ deletexx/sym/glo prc_l_idflnk 2 $ if f$trnlnm("TEMP").nes."" then close/nolog temp7 $ if f$search(temp) .nes. "" then delete/nolog 'temp';*    ------------------------------  # Date: Thu, 22 Sep 2005 10:07:22 GMT " From:   VAXman-  @SendSpamHere.ORG0 Subject: Re: Parent procedures and f$environment0 Message-ID: <00A4A2C4.5C57597D@SendSpamHere.ORG>  [ In article <BF5792BC.1454E%roktsci@comcast.net>, Jeff Cameron <roktsci@comcast.net> writes: 6 >Can anyone help me with this DCL programming request? > K >F$ENVIRONMENT("PROCEDURE") returns the file specification of the currently + >executing command procedure. That is cool!  > G >F$ENVIRONMENT("DEPTH") returns the integer depth of nested DCL command  >procedures. > K >Given that your depth is 6, how can you find the file specification of the M >parent calling procedure (level 5), and it's parent (level 4), and on up the  >tree to level 1?  >  >Thank you in advance. > 
 >Jeff Cameron    I have a procedure here:    >    http://tmesis.com/symbol/using/PREVIOUS_PROCEDURE_NAME.COM;  J that will get you the previous calling procedure name.  In fact, this will return the entire call stack.   I It requires TMESIS SYMBOL to set the values for some DCL internal symbols I but you could look these up in DCLDEF.STB and hardcode them if necessary. J TMESIS SYMBOL can look these up on the fly making the procedure functional on any version of VMS.   --  K VAXman- A Bored Certified VMS Kernel Mode Hacker   VAXman(at)TMESIS(dot)COM              5   "Well my son, life is like a beanstalk, isn't it?"     ------------------------------    Date: 22 Sep 2005 05:59:34 -0700# From: "Galen" <gltackett@gmail.com> 0 Subject: Re: Parent procedures and f$environmentC Message-ID: <1127393974.573952.189030@z14g2000cwz.googlegroups.com>   K > It requires TMESIS SYMBOL to set the values for some DCL internal symbols K > but you could look these up in DCLDEF.STB and hardcode them if necessary.   G Even looking them up in DCLDEF.STB could be scripted in DCL along these ? lines, though in pure DCL it would probably look pretty clunky:   " $    DCL_SYMBOL_NAMES := a,b,c,...2 $    LINK/MAP/NOEXE SYS$LOADABLE_IMAGES:DCLDEF.STB
 $    I = 0 $ LOOP: / $    NAME = F$EXTRACT(I, ",", DCL_SYMBOL_NAMES) * $    IF NAME .EQS. "," THEN GOTO LOOP_EXIT% $!  <search map file for symbol name> ; $!  <extract symbol value from first line of search result> 6 $!  <store value in appropriately named global symbol> $    I = I + 1 $    GOTO LOOP $ EXIT_LOOP:  < Using a tool like gawk would make this much easier to write.   Galen    ------------------------------    Date: 22 Sep 2005 06:01:51 -0700/ From: "Martin Walker" <martin.walker@csf.co.uk> J Subject: Re: Pathworks for the hobbyist, and general pricing of VMS thingsC Message-ID: <1127394111.485801.202650@z14g2000cwz.googlegroups.com>    Tom Garcia wrote:  > Hello, > H > Firstly, thank you to all those who arranged the hobbyist programme --H > having previously taken advantage of it to run a simple home PWS 500auL > web/mail server, I'm now really getting *into* VMS - this time with a herdD > of dump-liberated VAXstation 4000/60s - though I think power/noise> > requirements may preclude running *all* the latter 24/7 :-(. > K > I'm enjoying reading this newsgroup, and it's likely I'll be throwing out N > questions from time to time :-). First a simple licensing thing re PathworksN > (specifically v6.1 for VAX, though I think it's considered identical to 7.3a? > for licensing purposes?) -- is it accessible to the hobbyist?  > M > Here is what I have found out so far (this is probably all obvious but it's K > the first time I've investigated how VMS licensing works beyond executing ) > the DCL to register the hobbyist PAKs):  > M > (1) Some basic authentication services work without any extra licenses, but  > file sharing services do not.  > 1 > (2) The appropriate license is PWLMXXXCA07.03 .  > G > (3) The full commercial prices for the client licensing method are at F > http://h71000.www7.hp.com/swcat/us/pwrksasca.html -- ie $149 for oneK > concurrent PC access, though I have found around $105 advertised in other 	 > places.  > F > (4) New servers come with one or two client access licenses, and theH > September 2003 DECcampus provided perpetual licensing for some limitedL > number of clients (35?), so they *are* being bundled in some places -- butB > are they being bundled with anything accessible to the hobbyist? > M > Wondering out of interest how much everything costs to the commercial user, J > I see the price for an ActualRealCopy of VMS is around $500, but I thinkI > this is just the base system, i.e. not even TCP/IP let alone stuff like I > Pathworks? And the EIP which I think *does* include lots of fun toys is M > about three times this, i.e. .gt.budget. Are both these licenses valid just L > for some version x.y-z -- e.g. would I have to pay the whole $500 again toL > upgrade, say, base from 8.2 to 8.2-1, or is there discount? I am in UK, if > it makes any difference. >  > Thanks for your time,  >  > --9 > Tom Garcia | tgarcia@hivemind.remove-this-spam-trap.org     E For VMS/Alpha, when an "actual real user" (commercial user) buys VMS, ) they buy license(s) and media separately.   D Media kits are either the base operating system (with DECnet, TCPIP,F DECwindows but not much more) or a layered product CD kit which is allD the layered products.  Most customers need both.  (Some products are$ available as individual media kits).  G When you but a server machine you get the base license and EIP package. @  EIP includes DECnet, TCPIP, DECwindows, a couple of Pathworks /D Advanced Server licenses and a few other things.  With a workstationG system, you get the base license, a single user and NAS-150 (similar to C EIP but slightly different products and with a single license PAK). F You are entitled to use all the versions of those products released up to the purchase date.   C If you want a higher version, you must either have the product on a @ Right to Update (RTU) support contract or purchase a RTU licenseE (typically about half the price of a new license).  You don't need an C update license to go up a "-" version but you do for a "." upgrade. D (E.g. 8.2 to 8.2-1 does not require a RTU license; 8.2 to 8.3 will).F Pathworks / Advanced Server is one of the few products which require a! new license PAK for each upgrade.   E Licenses are either user based (e.g. additional VMS users, Pathworks) @ or machine based (e.g. network protocols).  Some products can beE purchased in either variant.  Machine based licenses are based on the E machines license class - either workgroup, departmental or enterprise 7 with prices typically doubling as you go up each class.   A The rules for VMS/I64 are different but similar in some respects.   G If you require UK prices contact me off-list - CSF are an authorised HP , reseller in the UK and we specialise in VMS.   Regards,   Martin   ------------------------------  % Date: Thu, 22 Sep 2005 14:01:22 +0200 ( From: "Rudolf Wingert" <win@fom.fgan.de>9 Subject: Poor I/O performance (was poor SCSI performance) 3 Message-ID: <000001c5bf6d$5979eb70$994614ac@wat153>    Hello,  G in case of the predecessor stream (poor SCSI performance), I did made a H few tests with COPY and BACKUP. For the test I created a 128MB file withF fixed record size. To measure the I/O performance I used the following
 two commands: & 1. $PIPE SH TI ; COPY file NL: ; SH TI, 2. $PIPE SH TI ; BACK file NL:T/SAVE ; SH TIA With BACKUP the performance was independent of record size (about H 30MB/s). Not so COPY; the performance varying from 60MB/s (record size =E 16384) to 0.6MB/s (record size = 16). In case of this I think COPY is ! not match intelligent programmed. H Copy is used to copy files, so it should use block i/o instead of recordF i/o if it is possible (disk, tape). An other question is, what's aboutE applications? IMHO sequential i/o should be every time block i/o from 3 media to cache memory. Or would this make problems? G One more question is, why BACKUP is slower then COPY if the record size  is greater then 1024?    Best regards R. Wingert    ------------------------------  % Date: Thu, 22 Sep 2005 13:35:21 +0100 % From: "issinoho" <issinoho@gmail.com> ( Subject: Remote X Session to PC thru SSH0 Message-ID: <11j598atd7dtnea@corp.supernews.com>  K Apologies if this is old news but I tried this today as a "I wonder if..."  3 exercise, and lo-and-behold it worked like a charm.   G Alpha sitting at home behind firewall, port-forwarding SSH; PC at work   behind corporate firewall.J Fired up eXcursion on PC at work, then started PuTTY & set X11 forwarding " 'on' and SSH'd into Alpha at home.K Typed 'mcr decw$clock' in PuTTY session and ... XClock appears on work PC.   Lovely.   H Not only that but all the X applications work seamlessly and very, very  quickly - even Mozilla.   M No need for complex 'set display' commands, no need to open X11 ports on any  % firewalls, just SSH and away you go.     ------------------------------  % Date: Thu, 22 Sep 2005 09:47:43 -0400 0 From: BRAD <bMradAhamPiltSon-at-coMmcAast.nPeSt>, Subject: Re: Remote X Session to PC thru SSH0 Message-ID: <nuGdnU4K_dCbK6_eRVn-pg@comcast.com>   issinoho wrote: M > Apologies if this is old news but I tried this today as a "I wonder if..."  5 > exercise, and lo-and-behold it worked like a charm.  > I > Alpha sitting at home behind firewall, port-forwarding SSH; PC at work   > behind corporate firewall.L > Fired up eXcursion on PC at work, then started PuTTY & set X11 forwarding $ > 'on' and SSH'd into Alpha at home.M > Typed 'mcr decw$clock' in PuTTY session and ... XClock appears on work PC.  	 > Lovely.  > J > Not only that but all the X applications work seamlessly and very, very  > quickly - even Mozilla.  > O > No need for complex 'set display' commands, no need to open X11 ports on any  ' > firewalls, just SSH and away you go.    D I have heard of folks doing this, but if you clear up just one tiny 7 problem for me, I'll see if this works for me, as well.   H When you say, "port-forwarding SSH", do you mean that the firewall does + the "port-forward", or that the Alpha does?      --  . Bradford J. Hamilton "All opinions are my own"@                       "Lose the MAPS, and replace '-at-' with @"   ------------------------------    Date: 22 Sep 2005 16:56:11 +01006 From: peter@langstoeger.at (Peter 'EPLAN' LANGSTOEGER), Subject: Re: Remote X Session to PC thru SSH, Message-ID: <4332e22b$1@news.langstoeger.at>  c In article <nuGdnU4K_dCbK6_eRVn-pg@comcast.com>, BRAD <bMradAhamPiltSon-at-coMmcAast.nPeSt> writes: K >> Not only that but all the X applications work seamlessly and very, very   >> quickly - even Mozilla. >>  P >> No need for complex 'set display' commands, no need to open X11 ports on any ( >> firewalls, just SSH and away you go.  > E >I have heard of folks doing this, but if you clear up just one tiny  8 >problem for me, I'll see if this works for me, as well. > I >When you say, "port-forwarding SSH", do you mean that the firewall does  , >the "port-forward", or that the Alpha does?  H Yes, you need a port forward in the (NAT) Firewall to allow incoming SSH4 at your alpha (all at home). An this was meant here.  G There is no need for extra configuration work, except you need to click I in PuTTY (a recent one of course) on "Enable X11 forwarding" in the menue D "Connection" "SSH" "X11" (and this you can save as default/session).  J It works with TCPIP V5.4 and up and TCPware V5.6-2 and up (both with ECOs)   --   Peter "EPLAN" LANGSTOEGER % Network and OpenVMS system specialist  E-mail  peter@langstoeger.atF A-1030 VIENNA  AUSTRIA              I'm not a pessimist, I'm a realist   ------------------------------  # Date: Thu, 22 Sep 2005 15:57:59 GMT " From:   VAXman-  @SendSpamHere.ORG, Subject: Re: Remote X Session to PC thru SSH0 Message-ID: <00A4A2F5.570FC47C@SendSpamHere.ORG>  X In article <11j598atd7dtnea@corp.supernews.com>, "issinoho" <issinoho@gmail.com> writes:L >Apologies if this is old news but I tried this today as a "I wonder if..." 4 >exercise, and lo-and-behold it worked like a charm. > H >Alpha sitting at home behind firewall, port-forwarding SSH; PC at work  >behind corporate firewall. K >Fired up eXcursion on PC at work, then started PuTTY & set X11 forwarding  # >'on' and SSH'd into Alpha at home. L >Typed 'mcr decw$clock' in PuTTY session and ... XClock appears on work PC.  >Lovely. > I >Not only that but all the X applications work seamlessly and very, very   >quickly - even Mozilla. > N >No need for complex 'set display' commands, no need to open X11 ports on any & >firewalls, just SSH and away you go.    ... and your point is?   --  K VAXman- A Bored Certified VMS Kernel Mode Hacker   VAXman(at)TMESIS(dot)COM              5   "Well my son, life is like a beanstalk, isn't it?"     ------------------------------  % Date: Thu, 22 Sep 2005 08:58:42 -0700 # From: "Tom Linden" <tom@kednos.com> , Subject: Re: Remote X Session to PC thru SSH( Message-ID: <opsxihb4mvzgicya@hyrrokkin>  : On 22 Sep 2005 16:56:11 +0100, Peter 'EPLAN' LANGSTOEGER   <peter@langstoeger.at> wrote:   9 > In article <nuGdnU4K_dCbK6_eRVn-pg@comcast.com>, BRAD   / > <bMradAhamPiltSon-at-coMmcAast.nPeSt> writes: K >>> Not only that but all the X applications work seamlessly and very, very  >>> quickly - even Mozilla.  >>> K >>> No need for complex 'set display' commands, no need to open X11 ports   
 >>> on any( >>> firewalls, just SSH and away you go. >>F >> I have heard of folks doing this, but if you clear up just one tiny: >> problem for me, I'll see if this works for me, as well. >>J >> When you say, "port-forwarding SSH", do you mean that the firewall does. >> the "port-forward", or that the Alpha does? > J > Yes, you need a port forward in the (NAT) Firewall to allow incoming SSH6 > at your alpha (all at home). An this was meant here. > I > There is no need for extra configuration work, except you need to click K > in PuTTY (a recent one of course) on "Enable X11 forwarding" in the menue F > "Connection" "SSH" "X11" (and this you can save as default/session). > H > It works with TCPIP V5.4 and up and TCPware V5.6-2 and up (both with   > ECOs)  >  OK, I have a related question   J I use PuTTY v58 from W2k to connect using SSH2 to a VMS7.3-2 machine and   thenK DECnet to a VMS7.3-1 machine where I run emacs 19.22  I have set enable X11 ! forwarding on the SSH connnection        When I try to run emacs I get        NORNS> emacs -nw FOO.TXT   [Spawning a new Kept EMACS] /   emacs: Terminal type vt102-80 is not defined. 6   If that is not the actual type of terminal you have,?   use the Bourne shell command `TERM=... export TERM' (C-shell: F   `setenv TERM ...') to specify the correct type.  It may be necessary>   to do `unset TERMCAP' (C-shell: `unsetenv TERMCAP') as well..   [Attached to DCL in directory $7$DKA0:[TOM]]       Now in the TERM dir there is  H   VT100.ELC;1         VT102.EL;1          VT125.EL;1          VT200.EL;1H   VT200.ELC;1         VT201.EL;1          VT220.EL;1          VT240.EL;1     NORNS> type VT102.EL   (load "term/vt100" nil t)   ?   Is the problem that the above lisp file is not byte compiled? .   Who is claiming that it is VT102-80,  PuTTY?  8   SHO TERM/FULL doesn't supply that form of designation.  L   I also tried just copying VT102.EL to VT102-80.EL under the obvious theory@   to no avail.  The '80' is likely just the width, I would guess   Tom    ------------------------------  + Date: Thu, 22 Sep 2005 13:14:07 +0000 (UTC) . From: klewis@LUMINA.MITRE.ORG (Keith A. Lewis)8 Subject: Strange SMTP problem receiving from MS-exchange. Message-ID: <dguamv$lf6$1@newslocal.mitre.org>  . I'm having a problem with the SMTP receiver.  - This is OpenVMS/Alpha 7.3-1, TCPIP 5.4 ECO 2.   K When I get a large incoming message from Microsoft Exchange 6.0.6603.0, the G SMTP reciever hangs, and eventually times out.  If I watch the files in I SYS$SPECIFIC:[TCPIP$SMTP], the hang seems to coincide with an increase in I the "used" size of the TCPIP$SMTP_RECV_RUN.LOG file -- as if the reciever G process knows it's done writing at the beginning of the hang.  The only L thing written after that are the SYSTEM-F-TIMEOUT and job termination stats.  G The final size of the file is (for example) 11285 blocks, and there are $ almost 1,000,000 free on the disk.    I Is anybody else having this problem?  I think it started after my upgrade . from TCPIP 5.3 to 5.4, and ECO 2 did not help.  0 --Keith Lewis              klewis {at} mitre.org> The above may not (yet) represent the opinions of my employer.   ------------------------------    Date: 22 Sep 2005 08:45:03 -0700/ From: "Shahin Yaz" <Shahinyaz@blueyonder.co.uk> > Subject: TCP/IP Config Setting to set Network speed to 100MbpsC Message-ID: <1127403903.577745.209510@g43g2000cwa.googlegroups.com>   C Our server was not starting up TCP/IP on VMS 7.1-2 (Old version 4.2 B which is UCX), and the server was not getting onto the network. WeC found it was due the speed setting of the Ethernet Port, and it was / resolved when the port was switched to 100Mbps. G I thought that VMS (or the Ethernet Card ?) would detect the port speed E and align itself. Is there a setting in TCP/IP Config (UCX for 7.1-2) = that forces this setting ? Where and How is this configured ?   & Thanks very much for any help on this.   Regards    Shahin   ------------------------------  # Date: Thu, 22 Sep 2005 16:23:37 GMT A From: "Colin Butcher" <colin_DOT.butcher_AT@xdelta_DOT.co_DOT.uk> B Subject: Re: TCP/IP Config Setting to set Network speed to 100Mbps> Message-ID: <dUAYe.113400$G8.77526@text.news.blueyonder.co.uk>  K Alpha console level. SET EWAx_MODE <value> (typically FASTFD). Use a proper 7 managed switch with the speed & duplex set accordingly.   D It's a hardware level NIC thing, not a TCPIP (or DECnet or whatever) protocol thing.    --     Hope this helps, Colin. ) colin DOT butcher AT xdelta DOT co DOT uk E It's not mine, but I like this definition: Legacy = stuff that works.    ------------------------------  + Date: Thu, 22 Sep 2005 16:51:08 +0000 (UTC) . From: klewis@LUMINA.MITRE.ORG (Keith A. Lewis)B Subject: Re: TCP/IP Config Setting to set Network speed to 100Mbps. Message-ID: <dgunds$roq$1@newslocal.mitre.org>   "Shahin Yaz" <Shahinyaz@blueyonder.co.uk> writes in article <1127403903.577745.209510@g43g2000cwa.googlegroups.com> dated 22 Sep 2005 08:45:03 -0700: D >Our server was not starting up TCP/IP on VMS 7.1-2 (Old version 4.2C >which is UCX), and the server was not getting onto the network. We D >found it was due the speed setting of the Ethernet Port, and it was0 >resolved when the port was switched to 100Mbps.H >I thought that VMS (or the Ethernet Card ?) would detect the port speedF >and align itself. Is there a setting in TCP/IP Config (UCX for 7.1-2)> >that forces this setting ? Where and How is this configured ?  I For the Alpha, you set the speed at the SRM console before you boot.  The  most common command is:    >>>SET EWA0_MODE AUTO-NEGOTIATE   F which causes the ethernet card EWA0 to determine its speed through the auto-negotiation process.     ' To get a list of candidate devices, use    >>>SHOW *_MODE  0 --Keith Lewis              klewis {at} mitre.org> The above may not (yet) represent the opinions of my employer.   ------------------------------    Date: 22 Sep 2005 09:44:32 -0700/ From: "Shahin Yaz" <Shahinyaz@blueyonder.co.uk> B Subject: Re: TCP/IP Config Setting to set Network speed to 100MbpsC Message-ID: <1127407472.301576.174480@g47g2000cwa.googlegroups.com>    Thanks for your reply.  * What would the 10Bps setting be ? SLOWHD ?  ? I am still puzzled as to why the alpha / ethernet card does not ' automatically adjust to the port speed.   G So everytime I move an Alpha from a 10Mbps to 100Mbps port I would have  to do this ?  2 Is this nothing to do with the driver installed ?    Thanks again   Shahin   ------------------------------    Date: 22 Sep 2005 07:39:54 -0500 From: briggs@encompasserve.orgL Subject: Re: Tentative DCL suggestion: string return value from .COM (Hello,3 Message-ID: <X9rPqE5LCU+M@eisner.encompasserve.org>   L In article <43321078.9020801@triumf.ca>, Fred Bach <music@triumf.ca> writes:! > briggs@encompasserve.org wrote:  > A >> In article <0pfYe.893$L45.422@fe07.lga>, Z <Z@no.spam> writes:  >>   >>>Galen wrote:  >>> I >>>>Just as EXIT currently does, it would set the reserved symbol $STATUS K >>>>to the numeric value status_code. But it would also optionally set some E >>>>new reserved symbol (perhaps $STATUS_STRING?) to the string value 
 >>>>"string".  >>>>I >>>>This would be cleaner than assigning a value to a global symbol as we  >>>>must currently do. >>>  >>>FYI, I use a /JOB logical.  >>> ; >>>You _could_  DEFINE/JOB STATUS_STRING "my return string"  >>   >>  F >> I guess you're free to use whatever you like.  However, using a jobD >> logical name seems to me to be just asking to be bitten by a race
 >> condition.  >> > 
 >    John, > @ >    I have seen what I think are race conditions in VMS before,@ >    mostly having to do with a delay in closing a file, another@ >    having to do with SET TERMINAL not working reliably.  CouldB >    you please more fully describe this particular race condition@ >    you're talking about wrt /job logicals?  We use them a lot.  @ Say you decide to have a DCL .COM file named FOO.COM and it sets a job logical name FOO_STATUS.  D FOO is a handy utility and you find yourself using it a lot in otherD code that you write.  And being a good programmer you include status! checks each time that it is used.   B One of your users decides to use $ SPAWN /NOWAIT or otherwise getsD two concurrent threads of execution going in the same job tree.  And1 he runs some of your standard code that uses FOO.   C If those two threads use FOO at the same time, one thread might see - the status code produced by the other thread.   , That's the race condition I'm talking about.   	John Briggs   ------------------------------    Date: 22 Sep 2005 09:56:24 -0500- From: Kilgallen@SpamCop.net (Larry Kilgallen) L Subject: Re: Tentative DCL suggestion: string return value from .COM (Hello,3 Message-ID: <BDckqf4aMlXb@eisner.encompasserve.org>   T In article <X9rPqE5LCU+M@eisner.encompasserve.org>, briggs@encompasserve.org writes:  D > One of your users decides to use $ SPAWN /NOWAIT or otherwise getsF > two concurrent threads of execution going in the same job tree.  And3 > he runs some of your standard code that uses FOO.  > E > If those two threads use FOO at the same time, one thread might see / > the status code produced by the other thread.  > . > That's the race condition I'm talking about.  C I believe that is a programming error so grievous that it cannot be  called a "race condition".  = One could use global variables safely by qualifying each name : with the name of the invoking procedure and making them beA process logical names.  Have the invoking procedure pass the name  in as a symbol or parameter.  + At that point this discussion becomes moot.   E A similar discussion has taken place for compiled languages regarding F the passing of complex data structures to exception handlers.  At someE point the handling of exceptional circumstances becomes top heavy and 3 introduces more vulnerabilities than it eliminates.    ------------------------------    Date: 22 Sep 2005 05:24:42 -0700# From: "Galen" <gltackett@gmail.com> X Subject: Re: Tentative DCL suggestion: string return value from .COM (Hello, Guy Peleg?)B Message-ID: <1127391882.794740.57220@g44g2000cwa.googlegroups.com>   > FYI, I use a /JOB logical.: > You _could_  DEFINE/JOB STATUS_STRING "my return string"  ? Yeah, I too have command procedures that do that sort of thing.   D The idea behind my suggestion was to allow procedures to return moreB information while not altering the global context, by which I meanA logical names plus global symbols. A command procedure could then > invoke another procedure with fewer concerns for name clashes.  B > I don't think it's lame, I just think its "the long way around".  D > Introduces lots of questions, also. For example, does the value ofF > $STATUS_STRING change anytime the value of $STATUS changes? Does theI > depth invoking the proc. see $STATUS_STRING as a local symbol? Can that $ > value be passed out of a PIPEline?  G Good points, David. If such a feature existed, I think it would be wise F to have $STATUS_STRING's properties basically mirror those of $STATUS.F That would make $STATUS_STRING a global symbol which, I believe, wouldD change any time $ EXIT was executed. I suppose it would have "" as aE default value and be passed out of a PIPE similarly to $STATUS (which + I'm not quite clear on in PIPE commands...)   F Also, what happens when an executable image exits? Would there have toC be a new string argument to SYS$EXIT as well? I doubt the VMS folks  would want to do that.    > There maybe a better approach.  . There always is. :-) What about the following?  C As things stand in current versions of DCL: While $STATUS_STRING is D officially a reserved symbol name, there's not much to stop one from
 merely doing:      $ $STATUS_STRING [:]== ...   $ EXIT [exit_status]  C I guess even this fairly small (?) concern might be lessened if VMS A explicitly reserved $STATUS_STRING for this purpose, or if such a E practice became widely adopted by community custom or as site policy.    ------------------------------   End of INFO-VAX 2005.530 ************************                                                                                                                                                                                                                                                                                                                                                                                      