INFO-VAX Fri, 18 Jul 2008 Volume 2008 : Issue 399 Contents: Re: "Network tape drive" for VMS Re: "Network tape drive" for VMS Re: "Network tape drive" for VMS Re: "Network tape drive" for VMS Re: "Network tape drive" for VMS Re: "Network tape drive" for VMS Re: Any known BASIC/SMG issues on Itanium? Re: Any known BASIC/SMG issues on Itanium? Re: Any known BASIC/SMG issues on Itanium? Re: Any known BASIC/SMG issues on Itanium? Re: c. 2008 Brian Schenkenberger. Re: c. 2008 Brian Schenkenberger. Re: c. 2008 Brian Schenkenberger. Re: c. 2008 Brian Schenkenberger. F$CVTIME inconsistency after all those years Re: F$CVTIME inconsistency after all those years Re: F$CVTIME inconsistency after all those years Re: Imagemagick 6.30 and text drawing ? Question about quoting inside a string Re: Question about quoting inside a string Re: Question about quoting inside a string Re: Question about quoting inside a string Re: Question about quoting inside a string Re: Question about quoting inside a string Re: Question about quoting inside a string Re: TCPIP$TELNETSYM_* logical names Re: TCPIP$TELNETSYM_* logical names Re: TCPIP$TELNETSYM_* logical names Re: TCPIP$TELNETSYM_* logical names Upgrade to Copper Gigabit ---------------------------------------------------------------------- Date: Fri, 18 Jul 2008 12:40:49 +0200 From: Albrecht Schlosser Subject: Re: "Network tape drive" for VMS Message-ID: Tom Linden wrote: > On Thu, 17 Jul 2008 07:01:15 -0700, Tom Linden wrote: > >> On Thu, 17 Jul 2008 06:11:14 -0700, Albrecht Schlosser >> wrote: >> >>> Is there anything available that could be called a >>> "network tape drive" that is supported with >>> >>> (a) OpenVMS / Alpha 7.3-2 >>> (b) OpenVMS / I64, 8.2 or 8.3 ff. ? >>> >>> The reason is to have the backup tape drive separated from the >>> server for easier access to change the backup tapes. >>> >>> Please don't ask, why would you want to do that. It's simply >>> because of a user's request. And this should *not* be another >>> VMS server (cluster node) with a tape drive ;-) >>> >>> Albrecht >> >> If you have fibre, put it on a Modular Data Router >> > You didn't say whether you had fibre or not, but I just checked > an MDR will run you ~$500. You can get a fibre switch for less > than $300 and HBAs can be had for $20-40 each. This is your > best solution, allowing you to put the tape unit as far away as > you please from the computer room Thanks to all who replied. Here's a summary of what I read so far. Fibre channel seems the way to go, if at all. It's supported on both platforms neede, including OpenVMS/Alpha 7.3-2. The server in question doesn't have FC yet, but maybe it could be upgraded. However, I can't believe the prices above. Are these for supported configurations? I had a price of about 4000 EUR (ca. $ 6000 ?) in mind for a FC adapter for Itanium, maybe together with or additional to a FC switch that would be needed. I'll have to check that further. iSCSI seems not to be supported yet. I found something about a demonstration kit for V. 8.3-1H1 (I64 only). Seems to be supported with 8.4 at best. Does anybody know more? Also thanks for all the interesting comments regarding DC and backup management. For the original question, there are two points to mention: (1) The "users" are not necessarily personnel in the same company as the IT people. Today, the server is in the users's office, but they complain about the noise, and the server can't be cooled appropriately. Thus the idea of moving it to the DC/server room. The IT of the other company would host the server, but they wouldn't like to do daily maintenance jobs like changing tapes :-( (2) Everything is a question of costs. There is no cluster, and having another server in the office to serve the backup tape would not help to remove the server from the office. Albrecht ------------------------------ Date: Fri, 18 Jul 2008 11:44:21 +0000 (UTC) From: gartmann@nonsense.immunbio.mpg.de (Christoph Gartmann) Subject: Re: "Network tape drive" for VMS Message-ID: In article , Albrecht Schlosser writes: [...] >Also thanks for all the interesting comments regarding DC and >backup management. For the original question, there are two points >to mention: > >(1) The "users" are not necessarily personnel in the same company >as the IT people. Today, the server is in the users's office, but >they complain about the noise, and the server can't be cooled >appropriately. Thus the idea of moving it to the DC/server room. >The IT of the other company would host the server, but they wouldn't >like to do daily maintenance jobs like changing tapes :-( > >(2) Everything is a question of costs. There is no cluster, and >having another server in the office to serve the backup tape >would not help to remove the server from the office. I woud prefer the suggested solution with an autoloader. Depending on the number of tapes and backup cycles it might be sufficient to exchange tapes once a week. Then, either the IT people change the magazine or one of the office people has to go to the IT department and do it. Regards, Christoph Gartmann -- Max-Planck-Institut fuer Phone : +49-761-5108-464 Fax: -80464 Immunbiologie Postfach 1169 Internet: gartmann@immunbio dot mpg dot de D-79011 Freiburg, Germany http://www.immunbio.mpg.de/home/menue.html ------------------------------ Date: Fri, 18 Jul 2008 11:49:56 +0000 (UTC) From: david20@alpha2.mdx.ac.uk Subject: Re: "Network tape drive" for VMS Message-ID: In article <69755f51-0707-446d-90ee-40054683fbc5@m73g2000hsh.googlegroups.com>, johnwallace4@gmail.com writes: >On Jul 18, 12:01 am, Mark McIntyre >wrote: >> johnwalla...@gmail.com wrote: >"Lights out computing" [1] isn't a phrase you hear much these days, >but once upon a time, there was a widespread goal of keeping people >OUT of the computer room as much as was possible, because a routinely >locked and routinely unoccupied computer room was (and still should >be) a safer computer room. Once upon a time, "managing" a server room >didn't require physical access to PeeCee consoles or whatever, stuff >could mostly be done remotely without requiring people actually >present in the computer room. In principle, systems could be managed >from anywhere on the corporate network using things called VT100s and >terminal servers and VAXcluster Console Systems (if you were lucky), >and related cleverness. These days, with a few honourable exceptions, >you routinely have people in the server room because you routinely >need access to the kbd/mouse/monitor to do routine sysadmin stuff on >x86 boxes and and whatever other PC-centric kit is in there, Nowadays you have a remote-KVM (KVM over IP) solution installed which is basically the same as the Vaxcluster Console system but with a graphical rather than serial/VT100 connection. David Webb Security team leader CCSS Middlesex University >and you >routinely have people in there adding new servers (or maybe blades) >every time a new production app arrives... I believe some people call >it progress. > >You want a shaggy dog true story? A former colleague commissioned the >VMS-based multi-site SCADA network for a major UK utility. One day, >his host for the day on a particular site was off sick, and therefore >because my colleague would have been unescorted, the security folks >didn't allow him on that particular site. Did it stop him doing what >he'd planned for the day? No. He explained the situation to an >understanding manager on a different site, who was happy to allow >access to his network and **the day's work was done remotely** over >the WAN, invisible to the security man who said you can't come in >because you need an escort. How does that work these days with PCs and >the like (unless you've spent a small fortune on Proliant-class high >end server management option cards and high-bandwidth low-latency >inter-site links)? > >Whatever. > >[1] http://www.it-analysis.com/technology/infrastructure/content.php?cid=8169 ------------------------------ Date: 18 Jul 2008 07:36:28 -0500 From: koehler@eisner.nospam.encompasserve.org (Bob Koehler) Subject: Re: "Network tape drive" for VMS Message-ID: In article <1f7df0f8-b35f-4319-b040-028d3383f5f7@59g2000hsb.googlegroups.com>, johnwallace4@gmail.com writes: > A long time ago, when I wanted to solve the problem being addressed > here, ie there was no guarantee that a server-room-authorised person > would be on site to change tapes, the "clustered VMS node with tape > drive, outside the server room" was the exact solution chosen. We put > the tape drive (and its host) where there *was* always going to be a > person available to change tapes, because in the circumstances at that > time, we considered the "security risks" of having a VMS system > outside the computer room were outweighed by the value of having > reasonable confidence that the daily backups would still get completed > even if multiple tapes were needed. Sites with particularly sensitive > data or other different factors to consider might end up making > different decisions. One size does not necessarily suit all. I would be concerned about the physical security of the backup tapes themselves. I would be very concerned about any IT structure that could not provide reliable backups inside a physically secure environment. ------------------------------ Date: Fri, 18 Jul 2008 06:35:14 -0700 From: "Tom Linden" Subject: Re: "Network tape drive" for VMS Message-ID: On Fri, 18 Jul 2008 03:40:49 -0700, Albrecht Schlosser wrote: > However, I can't believe the prices above. Are these for supported > configurations? I had a price of about 4000 EUR (ca. $ 6000 ?) in > mind for a FC adapter for Itanium, maybe together with or additional > to a FC switch that would be needed. I'll have to check that further. Supported doesn't mean you have to buy from HP. Qlogic for Itanium Emulex LP8000 for the Alphas, Brocade Silkworm switch (maybe branded compaq) Google and ebay are your friends. For the Emulex HBAs you will need to upgrade the firmware, which requires an alpha that has wwidmgr in the SRM console. I bought a DS10L just for this purpose (more-or-less) I may have spares lying about. I would put in two HBAs in each node and get two switches, that way you have fully redundant paths. Then you can throw in and HSG80 ... :-) Google and ebay are your friends -- PL/I for OpenVMS www.kednos.com ------------------------------ Date: Fri, 18 Jul 2008 08:00:37 -0700 From: Malcolm Dunnett Subject: Re: "Network tape drive" for VMS Message-ID: <4880B015.3040706@spammers.are.scum> >>> If you have fibre, put it on a Modular Data Router >>> >> You didn't say whether you had fibre or not, but I just checked >> an MDR will run you ~$500. You can get a fibre switch for less >> than $300 and HBAs can be had for $20-40 each. This is your >> best solution, allowing you to put the tape unit as far away as >> you please from the computer room > > Thanks to all who replied. Here's a summary of what I read so far. > > Fibre channel seems the way to go, if at all. It's supported on > both platforms neede, including OpenVMS/Alpha 7.3-2. The server in > question doesn't have FC yet, but maybe it could be upgraded. > > However, I can't believe the prices above. Are these for supported > configurations? I had a price of about 4000 EUR (ca. $ 6000 ?) in > mind for a FC adapter for Itanium, maybe together with or additional > to a FC switch that would be needed. I'll have to check that further. > Those pries will be for a supported configuration using used equipment. There's a modular data router on Ebay right now with a starting bid of $249. http://cgi.ebay.ca/COMPAQ-MODULAR-ROUTER-storageworks-171979-001-163083-B2_W0QQitemZ270255472245QQihZ017QQcategoryZ11183QQssPageNameZWDVWQQrdZ1QQcmdZViewItem One thing to watch when buying a modular data router is the kind of SCSI interface it supports. Many of the used ones on the market support HVD SCSI, whereas I suspect your current tape drive is LVD. The one I mention has LVD SCSCI. The older, supported host apapter for Alpha is the KGPSA-CA/CX. There's an ebay Buy-it-now for $19.95 for these: http://cgi.ebay.ca/COMPAQ-DS-KGPSA-CX-DEC-Fibre-Channel-HBA-PCI-X-CARD_W0QQitemZ110215497592QQihZ001QQcategoryZ11182QQrdZ1QQssPageNameZWD1VQQcmdZViewItemQQ_trksidZp1638Q2em118Q2el1247 a 16 channel 1gb fibrechannel switch is on ebay for $9.99 http://cgi.ebay.ca/Compaq-StorageWorks-SAN-Switch-16-Fibre-Channel-Switch_W0QQitemZ310067388105QQihZ021QQcategoryZ80221QQssPageNameZWDVWQQrdZ1QQcmdZViewItem So your total cost for switch,adpater and data router would be about $280 US (plus the cost of running the fibre). The above is all for 1gb fibrechannel equipment (which should be more than adequate if all you're driving is a tape). I don't know if the modular data router can even support faster fibrechannel. Hope this helps. ------------------------------ Date: Fri, 18 Jul 2008 07:51:07 -0400 From: John Sauter Subject: Re: Any known BASIC/SMG issues on Itanium? Message-ID: Rich Jordan wrote: > One of our VMS customers who recently switched from Alpha to Itanium > just turned up a problem with a not-commonly used feature. The > programs they use are in VMS BASIC and use the SMG libraries for > screen display. One function for print-screen uses the SMG > $PUT_PASTEBOARD call, which uses a user-written subroutine to actually > output each line of the pasteboard in sequence to a file/device/etc. > The code has been working for decades on VAX and Alpha. > > It fails on I64 VMS, both V8.3 and V8.2-1 using the same source files > and the same compile/link commands. We pulled out the two library > modules used and built them independently on each system using the > same BASIC command lines just to be sure (no other user library > functions are involved in the test program). > > The V8.2-1 system is up to date on ECOs including the BASRTL ones. > The V8.3 system is behind but running Update V4 with BASRTL V2. The > test program works fine, built from the same source with the same > BASIC command line options, on Alpha V7.3-2 and V8.3. BASIC is V1.6 > on the itaniums, V1.5 on the V7.3-2 Alpha, and V1.6 on the V8.3 Alpha. > > Running with debug I can trace it into the SMG call. The problem > appears to be the SMG routine's call to the user action routine, which > takes two arguments: > string descriptor (for one line of screen info) > unsigned long user argument by value (defaults to '0') > > The action routine is supposed to output or otherwise deal with the > one screen line per call and is called once for each line in the > pasteboard. Here it is called but the program dies as the routine is > entered, the source for the action routine pops up in the debug > window, but the error has already occurred. > > Error message received from the error handler is: > > %BAS-F-TOOFEWARGS, Too few arguments > -BAS-I-FROLINSUB, from line 2140 in line VT.PRINT_SCREEN > -BAS-I-FROMOD, In module TEST > > I know the argument setup for both the VT.PRINT_SCREEN call, and the > SMG$PUT_PASTEBOARD call are correct and we've updated the two print > subs (including the action routine) to 'modern' standards with > prototyping of arguments to the extent BASIC allows. No change. > > Just thought I'd ask here before talking to HP about it in case anyone > has seen something similar. So far this is the only porting issue > we've had outside of translating some very old VAX MACRO files to > BASIC. > > Thanks > I was the original developer of BASRTL, so I feel obliged to respond, even though I can't contribute much. A Basic subroutine calls the Basic RTL when it is first entered, to handle its arguments. It passes a description, created by the compiler, of what it is expecting to see from its caller. The check for number of arguments is one of the early ones. My experience is with the VAX code from 30 years ago, so things may have changed, but a good place to look is at the subroutine description created by the Basic compiler and passed to the Basic RTL when the subroutine is entered. If it specifies the expected number of arguments incorrectly, that is the problem. I have forgotten the format of that description, but you should be able to deduce it by stepping through the first RTL call that your Basic subroutine makes. John Sauter (John_Sauter@systemeyescomputerstore.com) ------------------------------ Date: Fri, 18 Jul 2008 07:05:30 -0700 (PDT) From: Rich Jordan Subject: Re: Any known BASIC/SMG issues on Itanium? Message-ID: On Jul 18, 6:51=A0am, John Sauter wrote: > Rich Jordan wrote: > > One of our VMS customers who recently switched from Alpha to Itanium > > just turned up a problem with a not-commonly used feature. =A0The > > programs they use are in VMS BASIC and use the SMG libraries for > > screen display. =A0One function for print-screen uses the SMG > > $PUT_PASTEBOARD call, which uses a user-written subroutine to actually > > output each line of the pasteboard in sequence to a file/device/etc. > > The code has been working for decades on VAX and Alpha. > > > It fails on I64 VMS, both V8.3 and V8.2-1 using the same source files > > and the same compile/link commands. =A0We pulled out the two library > > modules used and built them independently on each system using the > > same BASIC command lines just to be sure (no other user library > > functions are involved in the test program). > > > The V8.2-1 system is up to date on ECOs including the BASRTL ones. > > The V8.3 system is behind but running Update V4 with BASRTL V2. =A0The > > test program works fine, built from the same source with the same > > BASIC command line options, on Alpha V7.3-2 and V8.3. =A0BASIC is V1.6 > > on the itaniums, V1.5 on the V7.3-2 Alpha, and V1.6 on the V8.3 Alpha. > > > Running with debug I can trace it into the SMG call. =A0The problem > > appears to be the SMG routine's call to the user action routine, which > > takes two arguments: > > =A0 =A0 =A0string descriptor (for one line of screen info) > > =A0 =A0 =A0unsigned long user argument by value (defaults to '0') > > > The action routine is supposed to output or otherwise deal with the > > one screen line per call and is called once for each line in the > > pasteboard. =A0Here it is called but the program dies as the routine is > > entered, the source for the action routine pops up in the debug > > window, but the error has already occurred. > > > Error message received from the error handler is: > > > %BAS-F-TOOFEWARGS, Too few arguments > > -BAS-I-FROLINSUB, from line 2140 in line VT.PRINT_SCREEN > > -BAS-I-FROMOD, In module TEST > > > I know the argument setup for both the VT.PRINT_SCREEN call, and the > > SMG$PUT_PASTEBOARD call are correct and we've updated the two print > > subs (including the action routine) to 'modern' standards with > > prototyping of arguments to the extent BASIC allows. =A0No change. > > > Just thought I'd ask here before talking to HP about it in case anyone > > has seen something similar. =A0So far this is the only porting issue > > we've had outside of translating some very old VAX MACRO files to > > BASIC. > > > Thanks > > I was the original developer of BASRTL, so I feel obliged to respond, > even though I can't contribute much. =A0A Basic subroutine calls the Basi= c > RTL when it is first entered, to handle its arguments. =A0It passes a > description, created by the compiler, of what it is expecting to see > from its caller. =A0The check for number of arguments is one of the early > ones. > > My experience is with the VAX code from 30 years ago, so things may have > changed, but a good place to look is at the subroutine description > created by the Basic compiler and passed to the Basic RTL when the > subroutine is entered. =A0If it specifies the expected number of argument= s > incorrectly, that is the problem. =A0I have forgotten the format of that > description, but you should be able to deduce it by stepping through the > first RTL call that your Basic subroutine makes. > =A0 =A0 =A0John Sauter (John_Sau...@systemeyescomputerstore.com) John, thanks very much for taking time to respond. If the problem is at that level and not just some source level issue that the itanium doesn't like, I'll have to take it to HP support. The traceback for the program shows the following: image module routine LIBRTL LIB$CALLG LIB$CALLG DEC$BASRTL before tracing back the actual application subroutines. That does make it look like your supposition is likely correct. Rich ------------------------------ Date: Fri, 18 Jul 2008 07:00:52 -0700 (PDT) From: Rich Jordan Subject: Re: Any known BASIC/SMG issues on Itanium? Message-ID: <92802973-3d62-4375-b0cb-bf6ed64722c4@j33g2000pri.googlegroups.com> On Jul 17, 11:01=A0pm, John Santos wrote: > Rich Jordan wrote: > > One of our VMS customers who recently switched from Alpha to Itanium > > just turned up a problem with a not-commonly used feature. =A0The > > programs they use are in VMS BASIC and use the SMG libraries for > > screen display. =A0One function for print-screen uses the SMG > > $PUT_PASTEBOARD call, which uses a user-written subroutine to actually > > output each line of the pasteboard in sequence to a file/device/etc. > > The code has been working for decades on VAX and Alpha. > > > It fails on I64 VMS, both V8.3 and V8.2-1 using the same source files > > and the same compile/link commands. =A0We pulled out the two library > > modules used and built them independently on each system using the > > same BASIC command lines just to be sure (no other user library > > functions are involved in the test program). > > > The V8.2-1 system is up to date on ECOs including the BASRTL ones. > > The V8.3 system is behind but running Update V4 with BASRTL V2. =A0The > > test program works fine, built from the same source with the same > > BASIC command line options, on Alpha V7.3-2 and V8.3. =A0BASIC is V1.6 > > on the itaniums, V1.5 on the V7.3-2 Alpha, and V1.6 on the V8.3 Alpha. > > > Running with debug I can trace it into the SMG call. =A0The problem > > appears to be the SMG routine's call to the user action routine, which > > takes two arguments: > > =A0 =A0 =A0string descriptor (for one line of screen info) > > =A0 =A0 =A0unsigned long user argument by value (defaults to '0') > > > The action routine is supposed to output or otherwise deal with the > > one screen line per call and is called once for each line in the > > pasteboard. =A0Here it is called but the program dies as the routine is > > entered, the source for the action routine pops up in the debug > > window, but the error has already occurred. > > > Error message received from the error handler is: > > > %BAS-F-TOOFEWARGS, Too few arguments > > -BAS-I-FROLINSUB, from line 2140 in line VT.PRINT_SCREEN > > -BAS-I-FROMOD, In module TEST > > > I know the argument setup for both the VT.PRINT_SCREEN call, and the > > SMG$PUT_PASTEBOARD call are correct and we've updated the two print > > subs (including the action routine) to 'modern' standards with > > prototyping of arguments to the extent BASIC allows. =A0No change. > > > Just thought I'd ask here before talking to HP about it in case anyone > > has seen something similar. =A0So far this is the only porting issue > > we've had outside of translating some very old VAX MACRO files to > > BASIC. > > > Thanks > > We had a problem with code in a "USEROPEN" function not executing > correctly that is fixed in BASIC V1.7. =A0Compiling /nooptimize in > V1.6 was a workaround. =A0That was the only porting issue we have > found so far. > > -- > John Santos > Evans Griffiths & Hart, Inc. > 781-861-0670 ext 539 John, thanks. We'll try V1.7 at earliest opportunity. /NOOPTIMIZE doesn't seem to make any difference in this case with either BASIC 1.6 or 1.5 Rich ------------------------------ Date: Fri, 18 Jul 2008 09:04:36 -0700 (PDT) From: Rich Jordan Subject: Re: Any known BASIC/SMG issues on Itanium? Message-ID: <092377aa-a357-489b-99a8-dc5a72a730c6@j22g2000hsf.googlegroups.com> On Jul 18, 9:00=A0am, Rich Jordan wrote: > On Jul 17, 11:01=A0pm, John Santos wrote: > > > > > Rich Jordan wrote: > > > One of our VMS customers who recently switched from Alpha to Itanium > > > just turned up a problem with a not-commonly used feature. =A0The > > > programs they use are in VMS BASIC and use the SMG libraries for > > > screen display. =A0One function for print-screen uses the SMG > > > $PUT_PASTEBOARD call, which uses a user-written subroutine to actuall= y > > > output each line of the pasteboard in sequence to a file/device/etc. > > > The code has been working for decades on VAX and Alpha. > > > > It fails on I64 VMS, both V8.3 and V8.2-1 using the same source files > > > and the same compile/link commands. =A0We pulled out the two library > > > modules used and built them independently on each system using the > > > same BASIC command lines just to be sure (no other user library > > > functions are involved in the test program). > > > > The V8.2-1 system is up to date on ECOs including the BASRTL ones. > > > The V8.3 system is behind but running Update V4 with BASRTL V2. =A0Th= e > > > test program works fine, built from the same source with the same > > > BASIC command line options, on Alpha V7.3-2 and V8.3. =A0BASIC is V1.= 6 > > > on the itaniums, V1.5 on the V7.3-2 Alpha, and V1.6 on the V8.3 Alpha= . > > > > Running with debug I can trace it into the SMG call. =A0The problem > > > appears to be the SMG routine's call to the user action routine, whic= h > > > takes two arguments: > > > =A0 =A0 =A0string descriptor (for one line of screen info) > > > =A0 =A0 =A0unsigned long user argument by value (defaults to '0') > > > > The action routine is supposed to output or otherwise deal with the > > > one screen line per call and is called once for each line in the > > > pasteboard. =A0Here it is called but the program dies as the routine = is > > > entered, the source for the action routine pops up in the debug > > > window, but the error has already occurred. > > > > Error message received from the error handler is: > > > > %BAS-F-TOOFEWARGS, Too few arguments > > > -BAS-I-FROLINSUB, from line 2140 in line VT.PRINT_SCREEN > > > -BAS-I-FROMOD, In module TEST > > > > I know the argument setup for both the VT.PRINT_SCREEN call, and the > > > SMG$PUT_PASTEBOARD call are correct and we've updated the two print > > > subs (including the action routine) to 'modern' standards with > > > prototyping of arguments to the extent BASIC allows. =A0No change. > > > > Just thought I'd ask here before talking to HP about it in case anyon= e > > > has seen something similar. =A0So far this is the only porting issue > > > we've had outside of translating some very old VAX MACRO files to > > > BASIC. > > > > Thanks > > > We had a problem with code in a "USEROPEN" function not executing > > correctly that is fixed in BASIC V1.7. =A0Compiling /nooptimize in > > V1.6 was a workaround. =A0That was the only porting issue we have > > found so far. > > > -- > > John Santos > > Evans Griffiths & Hart, Inc. > > 781-861-0670 ext 539 > > John, > =A0 =A0 =A0thanks. =A0We'll try V1.7 at earliest opportunity. =A0/NOOPTIM= IZE > doesn't seem to make any difference in this case with either BASIC 1.6 > or 1.5 > > Rich BASIC V1.7 worked, after we recompiled the library modules as well as the program source. I'm getting an informational link message now %ILINK-I-DIFTYPE, symbol VT.PRINT_SCREEN_PRT of type FUNC cannot be referenced as type OBJECT module VT.PRINT_SCREEN file: DKA0:[TEST]VT.PRINT_SCREEN.OBJ;3 VT.PRINT_SCREEN_PRT is the user action subroutine passed to SMG $PUT_PASTEBOARD, and VT.PRINT_SCREEN is the actual subroutine called by the test program. I'm sure its something simple, but obviously the itanium is being more picky about it than the VAX or Alpha were (though I haven't installed V1.7 on the test VMS 8.3 Alpha yet so we'll see). Thanks for the suggestions. Rich ------------------------------ Date: Fri, 18 Jul 2008 11:14:21 +0000 (UTC) From: helbig@astro.multiCLOTHESvax.de (Phillip Helbig---remove CLOTHES to reply) Subject: Re: c. 2008 Brian Schenkenberger. Message-ID: > > c. 2008 Brian Schenkenberger. [...] > > Around here, "c." is read as "circa". Perhaps things differ where © That's compose-c-o. :-) ------------------------------ Date: Fri, 18 Jul 2008 11:16:37 +0000 (UTC) From: helbig@astro.multiCLOTHESvax.de (Phillip Helbig---remove CLOTHES to reply) Subject: Re: c. 2008 Brian Schenkenberger. Message-ID: In article , norm.raphael@metso.com writes: > I have seen (c) where the symbol was not appropriate. > > [I could put it here, but the message would then become MIME instead of > text.] Depends on your software. I can use the compose key to create 8-bit characters and my posts appear as plain text, not as MIME. > --=_alternative 0068169D85257488_= > Content-Type: text/html; charset="US-ASCII" > > >

> If you are concerned with avoiding MIME in a text-based newsgroup, please consider also posting the same text in HTML. :-| ------------------------------ Date: Fri, 18 Jul 2008 06:27:18 -0700 From: "Tom Linden" Subject: Re: c. 2008 Brian Schenkenberger. Message-ID: On Fri, 18 Jul 2008 04:14:21 -0700, Phillip Helbig---remove CLOTHES to reply wrote: >> > c. 2008 Brian Schenkenberger. [...] >> >> Around here, "c." is read as "circa". Perhaps things differ where > > © That's compose-c-o. :-) > On mine it is ctrl-alt-c -- PL/I for OpenVMS www.kednos.com ------------------------------ Date: 18 Jul 2008 12:29:06 -0500 From: koehler@eisner.nospam.encompasserve.org (Bob Koehler) Subject: Re: c. 2008 Brian Schenkenberger. Message-ID: <39Da$pQfCBXg@eisner.encompasserve.org> In article , "Tom Linden" writes: > On Fri, 18 Jul 2008 04:14:21 -0700, Phillip Helbig---remove CLOTHES to > reply wrote: > >>> > c. 2008 Brian Schenkenberger. [...] >>> >>> Around here, "c." is read as "circa". Perhaps things differ where >> >> © That's compose-c-o. :-) >> > > On mine it is ctrl-alt-c ctrl-atl-c brings up some screwy Lotus application. I still can't figure out how to get Putty to correctly do compose, but I did find out that it's "VT320e" font is one of the few that displays 8 bit characters using the MS-DOS character set instead of the DEC MCS character set (I'll assume by doing the ISO-LATIN-1 set). ------------------------------ Date: Fri, 18 Jul 2008 02:20:26 -0400 From: JF Mezei Subject: F$CVTIME inconsistency after all those years Message-ID: <4880365a$0$1844$c3e8da3@news.astraweb.com> F$CVTIME requires a time in: 31-FEB-1999:18:57:36 format. But the result is in format: 31-FEB-1999 18:57:36 format What this means is that you can't feed the result of F$CVTIME back into F$CVTIME ! Since it is the only means of doing date calculations in DCL, I would suggest that VMS engineering spend some brain CPU cycles thinking about some way to fix this :-) I built a time value and have to do one calculation on it, and then multiple different calculations from the base result, and since that base result is unusable, I have to first "fix" it to make it usable again. Perhaps add an item code in F$CVTIME that is similar to ABSOLUTE but has the requisite : betwene date and time ? ------------------------------ Date: Fri, 18 Jul 2008 00:07:31 -0700 (PDT) From: Bob Gezelter Subject: Re: F$CVTIME inconsistency after all those years Message-ID: <0dfa082a-9e8b-42a3-a35f-e0ddb20f05b5@v21g2000pro.googlegroups.com> On Jul 18, 2:20 am, JF Mezei wrote: > F$CVTIME requires a time in: 31-FEB-1999:18:57:36 format. > But the result is in format: 31-FEB-1999 18:57:36 format > > What this means is that you can't feed the result of F$CVTIME back into > F$CVTIME ! > > Since it is the only means of doing date calculations in DCL, I would > suggest that VMS engineering spend some brain CPU cycles thinking about > some way to fix this :-) > > I built a time value and have to do one calculation on it, and then > multiple different calculations from the base result, and since that > base result is unusable, I have to first "fix" it to make it usable again. > > Perhaps add an item code in F$CVTIME that is similar to ABSOLUTE but has > the requisite : betwene date and time ? JF, Could you please publish the reproducer for this behavior. On at least two of my test systems (a VAX and an Alpha; at two different release levels), the following test case INCLUDES the cc term: $ y = f$cvtime(f$time(), "absolute") $ show symbol y - Bob Gezelter, http://www.rlgsc.com ------------------------------ Date: Fri, 18 Jul 2008 03:26:54 -0400 From: JF Mezei Subject: Re: F$CVTIME inconsistency after all those years Message-ID: <488045e5$0$18595$c3e8da3@news.astraweb.com> Bob Gezelter wrote: > > Could you please publish the reproducer for this behavior. Nop !!!! Turns out I was using "COMPARISON" instead of "ABSOLUTE" with F$CVTIME and it seems you can't do any date arithmetic with the result of a COMPARISON operation aka: "18-Jul-2008 23:59:59+00:00:01" is valid but "2008-07-18 23:59:59+00:00:01" is NOT valid. So I have to do a F$CVTIME twice, once to generate a displayable date that can be used for further arithmetic and once for a comparison type date that is needed for my logic. Note however: $ c = f$cvtime(B,"COMPARISON") %DCL-W-IVATIME, invalid absolute time - use DD-MMM-YYYY:HH:MM:SS.CC format \2008-07-18 23:59:59.00-04:00\ It tells me to use a format where there is a colon between date and time. But When it generates a "ABSOLUTE" time, it puts a space there (hence my initial confusion). ------------------------------ Date: Fri, 18 Jul 2008 08:21:59 +0200 From: joukj Subject: Re: Imagemagick 6.30 and text drawing ? Message-ID: JF Mezei wrote: > For the sake of deja news archiving: > > I copied all .XML files to the current directory, > then edited type-ghosttscript.xml and changed the > > /usr/share/fonts/default/Type1/ > > to "gs_lib:" > > (which points to the ghostscript fonts) > > And Eureka ! I was able to add text to an image ! > > I would still need to work out how to make the unix software find the > .XML files "naturally" in the software directories instead of having to > use brite force and copy all those files over to the current working > directory to ensure that imagemagick can find them. > > Is there some mechanism on VMS that could help C software lookin for > "X.XML" file use some sort of search path that would look at list of > directories instead of just the current directory ? AS far as I know Imagemagick looks in sys$login for the files. So each user has to cpy them once. Jouk ------------------------------ Date: Fri, 18 Jul 2008 03:15:09 -0400 From: JF Mezei Subject: Question about quoting inside a string Message-ID: <48804324$0$18530$c3e8da3@news.astraweb.com> I need to produce the following: $im_convert temp.gif - -crop 240x300+237+1 - -font Helvetica-Bold -pointsize 20 - -draw "fill black text 300,280 'value' " temp.jpg The above will work correctly since VMS sees a double quote delimited string and inside, having single quotes is OK. However, I need "value" to be a symbol that is substituted. So this would intinctively become: -draw "fill black text 300,280 '''mysymbol'' " temp.jpg (replacing value with ''mysymbol') However, this yields: -draw "fill black text 300,280 mysymbol " temp.jpg First 3 quotes probably cause am emty symbol name to be substituted to nothing, and the ending double single quotes start symbol substitution that gets aborted after the space and are removed, leaving an untranslated symbol name. doing this works: -draw "fill black text 300,280 ' ''mysymbol' ' " temp.jpg Aka, adding a space between the textual quote at start and end so the the symbol's quotes. But this adds spaces to both sides of the value. Is there some way to get a ''symbol' to be substituted and be surrounded by single quotes ? ------------------------------ Date: Fri, 18 Jul 2008 09:26:38 +0200 From: rejoc Subject: Re: Question about quoting inside a string Message-ID: <488045ae$0$24371$426a74cc@news.free.fr> le 18.07.2008 09:15 JF Mezei a écrit: > I need to produce the following: > > $im_convert temp.gif - > -crop 240x300+237+1 - > -font Helvetica-Bold -pointsize 20 - > -draw "fill black text 300,280 'value' " temp.jpg > > > The above will work correctly since VMS sees a double quote delimited > string and inside, having single quotes is OK. > > However, I need "value" to be a symbol that is substituted. > > So this would intinctively become: > > -draw "fill black text 300,280 '''mysymbol'' " temp.jpg > > (replacing value with ''mysymbol') > > However, this yields: > > -draw "fill black text 300,280 mysymbol " temp.jpg > > First 3 quotes probably cause am emty symbol name to be substituted to > nothing, and the ending double single quotes start symbol substitution > that gets aborted after the space and afill black text 300,280 '''mysymbol'' "re removed, leaving an > untranslated symbol name. > > > doing this works: > > -draw "fill black text 300,280 ' ''mysymbol' ' " temp.jpg > > Aka, adding a space between the textual quote at start and end so the > the symbol's quotes. But this adds spaces to both sides of the value. > > > Is there some way to get a ''symbol' to be substituted and be surrounded > by single quotes ? something like f$fao("fill black text 300,280 '!AS'", mysymbol) ?? ------------------------------ Date: Fri, 18 Jul 2008 07:30:09 +0000 (UTC) From: gartmann@nonsense.immunbio.mpg.de (Christoph Gartmann) Subject: Re: Question about quoting inside a string Message-ID: In article <48804324$0$18530$c3e8da3@news.astraweb.com>, JF Mezei writes: >I need to produce the following: > >$im_convert temp.gif - > -crop 240x300+237+1 - > -font Helvetica-Bold -pointsize 20 - > -draw "fill black text 300,280 'value' " temp.jpg > > >The above will work correctly since VMS sees a double quote delimited >string and inside, having single quotes is OK. > >However, I need "value" to be a symbol that is substituted. > >So this would intinctively become: > >-draw "fill black text 300,280 '''mysymbol'' " temp.jpg > >(replacing value with ''mysymbol') > >However, this yields: > > -draw "fill black text 300,280 mysymbol " temp.jpg > >First 3 quotes probably cause am emty symbol name to be substituted to >nothing, and the ending double single quotes start symbol substitution >that gets aborted after the space and are removed, leaving an >untranslated symbol name. > > >doing this works: > > -draw "fill black text 300,280 ' ''mysymbol' ' " temp.jpg > >Aka, adding a space between the textual quote at start and end so the >the symbol's quotes. But this adds spaces to both sides of the value. > > >Is there some way to get a ''symbol' to be substituted and be surrounded >by single quotes ? -draw "fill black text 300,280 ''mysymbol' " temp.jpg should do what you want. Regards, Christoph Gartmann -- Max-Planck-Institut fuer Phone : +49-761-5108-464 Fax: -80464 Immunbiologie Postfach 1169 Internet: gartmann@immunbio dot mpg dot de D-79011 Freiburg, Germany http://www.immunbio.mpg.de/home/menue.html ------------------------------ Date: 18 Jul 2008 11:10:17 GMT From: VAXman- @SendSpamHere.ORG Subject: Re: Question about quoting inside a string Message-ID: <48807a18$0$5023$607ed4bc@cv.net> In article <48804324$0$18530$c3e8da3@news.astraweb.com>, JF Mezei writes: >I need to produce the following: > >$im_convert temp.gif - > -crop 240x300+237+1 - > -font Helvetica-Bold -pointsize 20 - > -draw "fill black text 300,280 'value' " temp.jpg > > >The above will work correctly since VMS sees a double quote delimited >string and inside, having single quotes is OK. > >However, I need "value" to be a symbol that is substituted. > >So this would intinctively become: > >-draw "fill black text 300,280 '''mysymbol'' " temp.jpg > >(replacing value with ''mysymbol') > >However, this yields: > > -draw "fill black text 300,280 mysymbol " temp.jpg > >First 3 quotes probably cause am emty symbol name to be substituted to >nothing, and the ending double single quotes start symbol substitution >that gets aborted after the space and are removed, leaving an >untranslated symbol name. > > >doing this works: > > -draw "fill black text 300,280 ' ''mysymbol' ' " temp.jpg > >Aka, adding a space between the textual quote at start and end so the >the symbol's quotes. But this adds spaces to both sides of the value. > > >Is there some way to get a ''symbol' to be substituted and be surrounded >by single quotes ? Would something simple like this work for you? $ TEST:==XYZZY $ SHOW SYMBOL TEST TEST == "XYZZY" $WRITE SYS$OUTPUT "This is the symbol '",TEST,"' with single quotes." This is the symbol 'XYZZY' with single quotes. -- VAXman- A Bored Certified VMS Kernel Mode Hacker VAXman(at)TMESIS(dot)COM ... pejorative statements of opinion are entitled to constitutional protection no matter how extreme, vituperous, or vigorously expressed they may be. (NJSC) Copr. 2008 Brian Schenkenberger. Publication of _this_ usenet article outside of usenet _must_ include its contents in its entirety including this copyright notice, disclaimer and quotations. ------------------------------ Date: Fri, 18 Jul 2008 14:28:51 +0200 From: Albrecht Schlosser Subject: Re: Question about quoting inside a string Message-ID: <4te6l5-o6c.ln1@news.hus-software.de> JF Mezei wrote: > I need to produce the following: > > $im_convert temp.gif - > -crop 240x300+237+1 - > -font Helvetica-Bold -pointsize 20 - > -draw "fill black text 300,280 'value' " temp.jpg > > > The above will work correctly since VMS sees a double quote delimited > string and inside, having single quotes is OK. > > However, I need "value" to be a symbol that is substituted. > > So this would intinctively become: > > -draw "fill black text 300,280 '''mysymbol'' " temp.jpg > > (replacing value with ''mysymbol') > > However, this yields: > > -draw "fill black text 300,280 mysymbol " temp.jpg > > First 3 quotes probably cause am emty symbol name to be substituted to > nothing, and the ending double single quotes start symbol substitution > that gets aborted after the space and are removed, leaving an > untranslated symbol name. > > > doing this works: > > -draw "fill black text 300,280 ' ''mysymbol' ' " temp.jpg > > Aka, adding a space between the textual quote at start and end so the > the symbol's quotes. But this adds spaces to both sides of the value. > > > Is there some way to get a ''symbol' to be substituted and be surrounded > by single quotes ? $ sq[0,8]=39 ! single quote $ value = "TeSt" $ string = "This is a test "+sq+value+sq+"." $ write sys$output "''string'" This is a test 'TeSt'. $ would this work ? Albrecht ------------------------------ Date: Fri, 18 Jul 2008 11:54:29 -0400 From: JF Mezei Subject: Re: Question about quoting inside a string Message-ID: <4880bcdb$0$18531$c3e8da3@news.astraweb.com> VAXman- @SendSpamHere.ORG wrote: > $WRITE SYS$OUTPUT "This is the symbol '",TEST,"' with single quotes." > This is the symbol 'XYZZY' with single quotes. For a write, yeah this would work. But this would be for a foreign command, so the above wouldn't work :-( ------------------------------ Date: 18 Jul 2008 12:24:14 -0500 From: koehler@eisner.nospam.encompasserve.org (Bob Koehler) Subject: Re: Question about quoting inside a string Message-ID: JF Mezei wrote: > I need to produce the following: > > $im_convert temp.gif - > -crop 240x300+237+1 - > -font Helvetica-Bold -pointsize 20 - > -draw "fill black text 300,280 'value' " temp.jpg > I think this will work, assuming value is a string: $im_convert temp.gif - -crop 240x300+237+1 - -font Helvetica-Bold -pointsize 20 - -draw "fill black text 300,280 " + value + " " temp.jpg But if it doesn't, then this: $temp = "fill black text 300,280 " + value + " " $im_convert temp.gif - -crop 240x300+237+1 - -font Helvetica-Bold -pointsize 20 - -draw 'temp temp.jpg ------------------------------ Date: Fri, 18 Jul 2008 12:46:15 +0200 From: Albrecht Schlosser Subject: Re: TCPIP$TELNETSYM_* logical names Message-ID: Bart.Zorn@gmail.com wrote: > Are you running a cluster? If so, you could consider running each > (class of) queue(s) on a different node, each with the corresponding > system logical names. No, no cluster. But IMHO, that wouldn't help, anyway. Restricting print queues to a particular cluster node would drastically degrade redundancy. And how many cluster nodes would be needed to support all possible combinations of logical names? Albrecht ------------------------------ Date: Fri, 18 Jul 2008 13:17:23 +0200 From: Albrecht Schlosser Subject: Re: TCPIP$TELNETSYM_* logical names Message-ID: <3na6l5-2e7.ln1@news.hus-software.de> Jeff Goodwin wrote: > "Albrecht Schlosser" wrote in message > news:4204l5-jjo.ln1@news.hus-software.de... >> I'm looking for a solution that seems to need the logical names >> >> TCPIP$TELNETSYM_SUPPRESS_FORMFEEDS and >> TCPIP$TELNETSYM_RAW_TCP >> >> preferably in combination with autostart print queues. I know that >> I can use a different logical name for a particular queue, because >> the logical names are read only at the time the queue is started. >> >> But, what would I do if all the queues should be started with >> $ enable autostart/queues >> when the system is booted ? [...] > We have this same issue at our sites with these two logical names and I do > not have an answer for you. I have requested the 'queue specific logicals' > feature for many years via many mechanisms. Just this year I've heard it is > being considered for a future release of TCPIP. > > It's about the most simple thing that can be programmed. Check for the > queue specific logical name like TCPIP$TELNETSYM_RAW_TCP_queue. If it > exists, use it. If it doesn't exist, check for the generic logical name. > DCPS seems to handle queue specific logical names quite nicely. It's really > an obvious problem with a simple solution that can be added to the code. It > is just pathetic it is taking so long. > > -Jeff Yes, that would be a reasonable solution. Having to change the logical name definitions for queue startup is a mess, anyway :-( . There could be race conditions (two operators starting a queue at the same time). Simple maintenance tasks like stopping and restarting a queue would need SYSNAM privilege [1] and the knowledge about the needed logical name values for a particular queue. And, last but not least, there seems to be no way to determine the logical name values that existed, when a queue was started. [1] I didn't check if starting a queue needs SYSNAM otherwise, but I can't imagine a reason for this. VMS engineering? Are there any plans to improve this? Another point to consider would be: If I want to define a system default value for a logical name and some different values when starting a particular queue, then I would do it with, e.q. $ define/system/exec TCPIP$TELNETSYM_SUPPRESS_FORMFEEDS 34 and, when starting a queue with different needs, I would like to define the same logical name without /exec, e.g. $ define/system TCPIP$TELNETSYM_SUPPRESS_FORMFEEDS 35 $ start/queue ... $ deassign/system TCPIP$TELNETSYM_SUPPRESS_FORMFEEDS How would I do this for a logical name like TCPIP$TELNETSYM_RAW_TCP that can only be defined (to 1) or not defined to disable the option? The docs don't say anything about a way to disable the feature. Is it possible to say "$ define/system TCPIP$TELNETSYM_RAW_TCP 0" to disable raw socket mode for a particular queue? Albrecht ------------------------------ Date: Fri, 18 Jul 2008 07:00:41 -0700 (PDT) From: "Bart.Zorn@gmail.com" Subject: Re: TCPIP$TELNETSYM_* logical names Message-ID: <67b2bcec-f161-47da-8bd8-0ad88eade608@a2g2000prm.googlegroups.com> On Jul 18, 12:46=A0pm, Albrecht Schlosser wrote: > Bart.Z...@gmail.com wrote: > > Are you running a cluster? If so, you could consider running each > > (class of) queue(s) on a different node, each with the corresponding > > system logical names. > > No, no cluster. But IMHO, that wouldn't help, anyway. Restricting > print queues to a particular cluster node would drastically degrade > redundancy. And how many cluster nodes would be needed to support > all possible combinations of logical names? > > Albrecht OK, but I think that no cluster is not more redundant than a cluster. What about a sort of supervisor, which stops queues, changes logical names and starts the appropriate other queues to do their work. Once done, on to the next combination... Bart ------------------------------ Date: Fri, 18 Jul 2008 16:51:21 +0200 From: rejoc Subject: Re: TCPIP$TELNETSYM_* logical names Message-ID: <4880ade9$0$23106$426a74cc@news.free.fr> le 18.07.2008 13:17 Albrecht Schlosser a écrit: > Jeff Goodwin wrote: >> "Albrecht Schlosser" wrote in message >> news:4204l5-jjo.ln1@news.hus-software.de... >>> I'm looking for a solution that seems to need the logical names >>> >>> TCPIP$TELNETSYM_SUPPRESS_FORMFEEDS and >>> TCPIP$TELNETSYM_RAW_TCP >>> >>> preferably in combination with autostart print queues. I know that >>> I can use a different logical name for a particular queue, because >>> the logical names are read only at the time the queue is started. >>> >>> But, what would I do if all the queues should be started with >>> $ enable autostart/queues >>> when the system is booted ? Perhaps a solution... 1/ copy tcpip$telnetsym.exe to tcpip$telnetsym2.exe 2/ patch tcpip$telnetsym2.exe to replace all occurences of the logical names you want to be specific with other names (same length)... 3/ define the newly named logical names with the correct values 4/ use this new image as the /processor for the queues that need it... You'll have to deal with new versions of the symbiont but I think they don't evolve often. ------------------------------ Date: Fri, 18 Jul 2008 09:53:18 -0400 From: "David" Subject: Upgrade to Copper Gigabit Message-ID: Island Computers is pleased to offer the following LOW COST Gigabit options for your Alpha IC-DEGXA-TA 1Gbit Copper PCi/PCIX Card (Same as DEGXA-TA) 10 Pack $1500 or $179 each IC-DEGX2-TA 2 Port 1Gbit Copper PCI/PCIX Card 10 Pack $4090 or $449 each These will work perfectly with no issues on VMS 7.3> and Tru64 5.1A> (T64 will require a patch before Build 4 of T64 5.1B) Let us know if interested -- David ============================================= Island Computers US Corp PO Box 86 Tybee GA 31328 Toll Free: 1-877 636 4332 x201, Mobile x251 Email: dturner@islandco.com International & Local: (001)- 404-806-7749 Fax: 912 786 8505 Web: www.islandco.com ============================================= ------------------------------ End of INFO-VAX 2008.399 ************************