1 INFO-VAX	Fri, 29 Jul 2005	Volume 2005 : Issue 419       Contents:) Re: Another VMS system bites the dust :-( ) Re: Another VMS system bites the dust :-( ) Re: Another VMS system bites the dust :-( ! Re: Collapsing a bound volume set ! Re: Collapsing a bound volume set ! Re: Collapsing a bound volume set ! Re: Collapsing a bound volume set ! Re: Collapsing a bound volume set ! Re: Collapsing a bound volume set ! Re: Collapsing a bound volume set ! Re: Collapsing a bound volume set ! Re: Collapsing a bound volume set ! Re: Collapsing a bound volume set ! Re: Collapsing a bound volume set  Re: DSSI EISA config files Re: DSSI EISA config files) How does sys$pal_swasten work on Itanium. L Re: It's that time again for my quarterly "Where is the CONDIST???" posting.L Re: It's that time again for my quarterly "Where is the CONDIST???" posting.P Re: It's that time again for my quarterly "Where is the CONDIST???" posting. posP Re: It's that time again for my quarterly "Where is the CONDIST???" posting. pos Re: LaTeX ---> PDF Re: LaTeX ---> PDF Re: LaTeX ---> PDF Re: LaTeX ---> PDF Re: LaTeX ---> PDF Re: Open VMS TCP / Mail 3 Re: perfomance question about bool variables in cxx 3 Re: perfomance question about bool variables in cxx 8 Re: Rounding v Truncation, was: Re: Platform Support vs.8 Re: Rounding v Truncation, was: Re: Platform Support vs.8 Re: Rounding v Truncation, was: Re: Platform Support vs.8 Re: Rounding v Truncation, was: Re: Platform Support vs.$ SYMVEC: a VMS linker options utility= Re: [OT] Rounding v Truncation, was: Re: Platform Support vs.   F ----------------------------------------------------------------------  % Date: Thu, 28 Jul 2005 14:07:04 -0400 - From: William Webb <william.w.webb@gmail.com> 2 Subject: Re: Another VMS system bites the dust :-(7 Message-ID: <8660a3a105072811075ee9a76b@mail.gmail.com>   : On 7/28/05, JF Mezei <jfmezei.spamnot@teksavvy.com> wrote:
 > Z wrote:L > > > With regards to hardware field service, it's been downhill since Comp= aq > > > bought DEC.  >=20  * Well, Compaq killed Logistics, that's one.  D And they don't have those "kit" trunks like they did in the DEC daysF which contained every conceivable part that was prone to break or wear/ out for a given family of devices...that's two.   J > > We have 4 (?) hour response and have never had to wait more than a few: > > hours for replacement h/w to show up at customer sites >=20 >=20J > Is it possible that in some cities, the quality of field service is much > better than in others ?    Very much so. =20   A If your FSE is two hours away by car, your service will not be as % snappy as if you have an on-site FSE.    >=20L > I know that for sales, when I had to deal with DEC in montreal, it was al= waysL > pretty lame, whereas I kept hearing stories of how great DEC reps were in=  the > kitchener area.  >=20  6 Perhaps it was because it was Montreal (just kidding.)  A I've had Gold support at two sites, consecutively since 1996. =20   ? And HP, if you're paying attention, props to Jim Shane and Clay 9 Williams who epitomize Field Service Engineer excellence.   D And I can say that the only difficulty I've had with Field Service'sE finding of parts has been recently when they took a couple of days to 5 find an RZ26, and were unable to find an RZ23 at all.    : ^ )   B And I anticipated that the RZ23 might be a bit, uh, difficult, andF while logging the call, told them that anything 8GB or less would make an acceptable substitute.    WWWebb   --=20 C NOTE: This email address is only used for noncommerical VMS-related  correspondence. C All unsolicited commercial email will be deemed to be a request for 8 services pursuant to the terms and conditions located at# http://bellsouthpwp.net/w/e/webbww/    ------------------------------  % Date: Thu, 28 Jul 2005 20:09:39 +0200 - From: Didier Morandi <prenom.nom@freesurf.fr> 2 Subject: Re: Another VMS system bites the dust :-(4 Message-ID: <42e91eec$0$28795$626a14ce@news.free.fr>   maher_rj@hotmail.com wrote:  > Hi Didier, > * > Is that the whole article or a summary?    whole.   D.   ------------------------------  # Date: Fri, 29 Jul 2005 03:52:39 GMT 6 From: "Kenneth Farmer" <kfarmer@NOSPAM.spyderbyte.com>2 Subject: Re: Another VMS system bites the dust :-(> Message-ID: <bKhGe.64224$3j2.2188726@twister.southeast.rr.com>  G "Z" <Z@no.spam> wrote in message news:sp5Ge.28697$wj5.12306@fe04.lga... ! > shoppa@trailing-edge.com wrote: $ >>>ALREADY terminated service on it? > J >> With regards to hardware field service, it's been downhill since CompaqI >> bought DEC.  Not so bad at first, but the hill has gotten really steep J >> lately.  We have "Gold level" service and typically it takes HP most ofJ >> a week to find the parts necessary.  It's getting real tempting to just; >> buy a lot of spare hardware and do everything ourselves.  > I > We have 4 (?) hour response and have never had to wait more than a few  I > hours for replacement h/w to show up at customer sites (very important  L > when you stand to pay $10K per minute of downtime in liquidated damages), J > even when it has to be driven to the site by courier van from a depot a  > few hours away.  > L > It's always there, always on time and always the right part(s). One time, M > there was some confusion re: the part no and the depot sent the driver out  + > with one of each variation, just in case.  > 5 > I think very highly of DEC/Compaq/HP field service.  >  > G >> Software support, we haven't used it much lately but in the past few * >> years we've gotten very useful answers. > . > I find the s/w support to be very good, too. > / > It's expensive, but you get what you pay for.       L Same here, flawless experiences with Gold level support.  Parts rarely took G over a few hours.  Many times it was on site in less than an hour.  Of  C course, the labor involved replacing the part could take some time.      Ken   % _____________________________________ " Kenneth R. Farmer <>< 336-736-73763 www.OpenVMS.org | dba.OpenVMS.org | dcl.OpenVMS.org  HP OpenVMS News and Info     ------------------------------  % Date: Thu, 28 Jul 2005 19:26:13 +0200 2 From: martin@radiogaga.harz.de (Martin Vorlaender)* Subject: Re: Collapsing a bound volume set; Message-ID: <42e91535.524144494f47414741@radiogaga.harz.de>   % H Vlems <nospam@what.ever.com> wrote: + > "Ed Wilts" <ewilts@ewilts.org> schreef... - >> I've got a 4-member bound volume set on an H >> HSJ-50 that I'd like to migrate to a new single volume on an EVA.  IsF >> it possible to migrate all of the files to a single member and thenC >> eliminate the volume set by mirroring the remaining volume to an H >> EVA-based disk?  I'm trying to avoid using VMSbackup for this becauseJ >> the directory hierarchy consists of lots and lots of little files and IJ >> don't want to take the volume offline for that long unless I don't have >> a choice. > M > Well, umm, BACKUP/IMAGE would actually be my first choice because it is the  > fastest way.  H Sorry, but no. I have been trying to BACKUP/IMAGE an old volume set to aG shiny new MSA-1000 partition just recently, and BACKUP refused to work: E you can only make an image copy to the same number of disks as are in  the volume set.   % I ended up doing a non-/IMAGE backup.    cu,    Martin --  @   OpenVMS:                | Martin Vorlaender  |  OpenVMS rules!3    The operating system   | work: mv@pdv-systeme.de F    God runs the           |   http://www.pdv-systeme.de/users/martinv/:    earth simulation on.   | home: martin@radiogaga.harz.de   ------------------------------  % Date: Thu, 28 Jul 2005 14:58:04 -0400 - From: JF Mezei <jfmezei.spamnot@teksavvy.com> * Subject: Re: Collapsing a bound volume set, Message-ID: <42E92AB4.5ABAA9B8@teksavvy.com>   Martin Vorlaender wrote:J > Sorry, but no. I have been trying to BACKUP/IMAGE an old volume set to aI > shiny new MSA-1000 partition just recently, and BACKUP refused to work: G > you can only make an image copy to the same number of disks as are in  > the volume set.   L Interesting. Doesn't a bound volume set appear as a single drive to VMS ? Or2 is there an indexf.sys for every physical volume ?  J Could you BACKUP/IMAGE to a save set, and then restore that saveset to the MSA-1000 drive ?  N Not using /IMAGE means that when you have alias directories, you end up havingM truly multiple copies of files, as opposed to the same file being copied over L itself multiple times (unless you use /NOLALIAS in which case, you only have one copy operation per file).    ------------------------------  # Date: Thu, 28 Jul 2005 19:11:10 GMT A From: "Colin Butcher" <colin_DOT.butcher_AT@xdelta_DOT.co_DOT.uk> * Subject: Re: Collapsing a bound volume set= Message-ID: <i5aGe.78821$G8.11477@text.news.blueyonder.co.uk>   F INIT the new volume on the EVA, then move the inactive files across inH groups by BACKUP, then finally quiesce the system and move the remainingJ files, then point everything at the new volume? Depends on the applicationE code and where its logicals point to - you may be able to move things A piecemeal and redirect the logicals as needed 'on the fly'. YMMV.   H Simplest and fastest is probably quiesce the system, then move the filesJ with backup (either directly, or via a disk based saveset), then bring theL applications back into production use. Then backup the HSJ50 bound volume to tape for posterity...    --     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, 28 Jul 2005 19:39:45 +0000 (UTC) P From: helbig@astro.multiCLOTHESvax.de (Phillip Helbig---remove CLOTHES to reply)* Subject: Re: Collapsing a bound volume set$ Message-ID: <dcbca1$ckk$1@online.de>  ; In article <42e91535.524144494f47414741@radiogaga.harz.de>, 5 martin@radiogaga.harz.de (Martin Vorlaender) writes:    ' > H Vlems <nospam@what.ever.com> wrote: - > > "Ed Wilts" <ewilts@ewilts.org> schreef... / > >> I've got a 4-member bound volume set on an J > >> HSJ-50 that I'd like to migrate to a new single volume on an EVA.  IsH > >> it possible to migrate all of the files to a single member and thenE > >> eliminate the volume set by mirroring the remaining volume to an J > >> EVA-based disk?  I'm trying to avoid using VMSbackup for this becauseL > >> the directory hierarchy consists of lots and lots of little files and IL > >> don't want to take the volume offline for that long unless I don't have > >> a choice. > > O > > Well, umm, BACKUP/IMAGE would actually be my first choice because it is the  > > fastest way. > J > Sorry, but no. I have been trying to BACKUP/IMAGE an old volume set to aI > shiny new MSA-1000 partition just recently, and BACKUP refused to work: G > you can only make an image copy to the same number of disks as are in  > the volume set.  > ' > I ended up doing a non-/IMAGE backup.   G Does BACKUP/IMAGE preserve file-IDs?  If so, then it is clear that the  A number of members has to be the same, since the file-ID contains  ( information about volume-set membership.   ------------------------------  + Date: Thu, 28 Jul 2005 19:44:04 +0000 (UTC) P From: helbig@astro.multiCLOTHESvax.de (Phillip Helbig---remove CLOTHES to reply)* Subject: Re: Collapsing a bound volume set$ Message-ID: <dcbci4$ckk$2@online.de>  5 In article <42E92AB4.5ABAA9B8@teksavvy.com>, JF Mezei ' <jfmezei.spamnot@teksavvy.com> writes:   > Martin Vorlaender wrote:L > > Sorry, but no. I have been trying to BACKUP/IMAGE an old volume set to aK > > shiny new MSA-1000 partition just recently, and BACKUP refused to work: I > > you can only make an image copy to the same number of disks as are in  > > the volume set.  > N > Interesting. Doesn't a bound volume set appear as a single drive to VMS ? Or4 > is there an indexf.sys for every physical volume ?  H The latter.  In fact, there are all the .SYS files.  At the user level, B yes, it appears as a single drive, but one can access the members ( individually with sufficient privileges.  L > Could you BACKUP/IMAGE to a save set, and then restore that saveset to the > MSA-1000 drive ? > P > Not using /IMAGE means that when you have alias directories, you end up havingO > truly multiple copies of files, as opposed to the same file being copied over N > itself multiple times (unless you use /NOLALIAS in which case, you only have > one copy operation per file).   ? Yes.  Hopefully, if using /IMAGE, then the setting of /ALIAS is > irrelevant (and it should be an error if /ALIAS or /NOALIAS isF specified), since with an image backup neither ignoring alias entries G nor having multiple copies of the file are acceptable; one should have  H copies of the aliases.  Perhaps /NOALIAS would be more appropriate here,H since presumably the directories (where the information is stored) will D be copied properly, and the backup itself will see that the file is + already there if it tries to copy it again.   G The HELP is confusing, frightening, and differs for different versions   of VMS.    ------------------------------    Date: 28 Jul 2005 13:25:18 -0700$ From: "AEF" <spamsink2001@yahoo.com>* Subject: Re: Collapsing a bound volume setC Message-ID: <1122582318.776084.211550@f14g2000cwb.googlegroups.com>   / Phillip Helbig---remove CLOTHES to reply wrote: = > In article <42e91535.524144494f47414741@radiogaga.harz.de>, 6 > martin@radiogaga.harz.de (Martin Vorlaender) writes: > ) > > H Vlems <nospam@what.ever.com> wrote: / > > > "Ed Wilts" <ewilts@ewilts.org> schreef... 1 > > >> I've got a 4-member bound volume set on an L > > >> HSJ-50 that I'd like to migrate to a new single volume on an EVA.  IsJ > > >> it possible to migrate all of the files to a single member and thenG > > >> eliminate the volume set by mirroring the remaining volume to an L > > >> EVA-based disk?  I'm trying to avoid using VMSbackup for this becauseN > > >> the directory hierarchy consists of lots and lots of little files and IN > > >> don't want to take the volume offline for that long unless I don't have > > >> a choice. > > > Q > > > Well, umm, BACKUP/IMAGE would actually be my first choice because it is the  > > > fastest way. > > L > > Sorry, but no. I have been trying to BACKUP/IMAGE an old volume set to aK > > shiny new MSA-1000 partition just recently, and BACKUP refused to work: I > > you can only make an image copy to the same number of disks as are in  > > the volume set.  > > ) > > I ended up doing a non-/IMAGE backup.  > H > Does BACKUP/IMAGE preserve file-IDs?  If so, then it is clear that the   Yes.  B > number of members has to be the same, since the file-ID contains* > information about volume-set membership.   Indeed.    ------------------------------    Date: 28 Jul 2005 15:29:59 -0500 From: briggs@encompasserve.org* Subject: Re: Collapsing a bound volume set3 Message-ID: <Ko2kpxrNmW55@eisner.encompasserve.org>   \ In article <42E92AB4.5ABAA9B8@teksavvy.com>, JF Mezei <jfmezei.spamnot@teksavvy.com> writes: > Martin Vorlaender wrote:K >> Sorry, but no. I have been trying to BACKUP/IMAGE an old volume set to a J >> shiny new MSA-1000 partition just recently, and BACKUP refused to work:H >> you can only make an image copy to the same number of disks as are in >> the volume set. > N > Interesting. Doesn't a bound volume set appear as a single drive to VMS ? Or4 > is there an indexf.sys for every physical volume ?  @ Yes.  In [000000]000000.DIR on the root volume you have all yourF top level directories catalogues.  As usual, these top level directory( entries contain file names and file IDS.  6 So, for instance, you might have MEZEI.DIR  (3,1234,1)  A So you'd find the file header for MEZEI.DIR on relative volume 3.   @ Pop open MEZEI.DIR and look for [MEZEI]SMTP.DIR.  Let's say that it is (2,2345,3).   E So you find the file header for [MEZEI]SMTP.DIR on relative volume 2.   B Pop open [MEZEI]SMTP.DIR and look for FOO.BAR.  Let's say that you find it at (0,222,1).   G So you find the file header for FOO.BAR on the same volume as SMTP.DIR,  relative volume 2.  F You pop open the file header and find the first two contiguous extentsJ for FOO.BAR.  Those extents must be on the same volume as the file header.  * But there's an extension header, (1,111,2)  I So you open up that file header and find another three contiguous extents ! for FOO.BAR on relative volume 1.    	John Briggs   ------------------------------    Date: 28 Jul 2005 14:49:37 -0700$ From: "Ed Wilts" <ewilts@ewilts.org>* Subject: Re: Collapsing a bound volume setC Message-ID: <1122587377.910249.284170@g44g2000cwa.googlegroups.com>   H > INIT the new volume on the EVA, then move the inactive files across inJ > groups by BACKUP, then finally quiesce the system and move the remaining1 > files, then point everything at the new volume?   C That might be okay if we had rsync for VMS but we don't.  Files are G being created and deleted so this problem is very difficult.  Quiescing B the is non trivial because of the large number of files (>250,000)" involved (I'm not the developer!).  G I'm not looking to minimize the move time - I'm looking to minimize the $ time for application unavailability.  	    .../Ed    ------------------------------  + Date: Thu, 28 Jul 2005 22:03:10 +0000 (UTC) P From: helbig@astro.multiCLOTHESvax.de (Phillip Helbig---remove CLOTHES to reply)* Subject: Re: Collapsing a bound volume set$ Message-ID: <dcbkmt$o02$1@online.de>  G In article <1122587377.910249.284170@g44g2000cwa.googlegroups.com>, "Ed # Wilts" <ewilts@ewilts.org> writes:    J > > INIT the new volume on the EVA, then move the inactive files across inL > > groups by BACKUP, then finally quiesce the system and move the remaining3 > > files, then point everything at the new volume?  > E > That might be okay if we had rsync for VMS but we don't.  Files are I > being created and deleted so this problem is very difficult.  Quiescing D > the is non trivial because of the large number of files (>250,000)$ > involved (I'm not the developer!). > I > I'm not looking to minimize the move time - I'm looking to minimize the & > time for application unavailability.  F How much downtime can you afford?  Suppose you can afford the time to G shut down the application and reboot.  Then you could do the following:   $    o  shadow the current volume set:  F       The following command creates a volume set with the logical nameE       TEST3013. The volume set TEST3013 is shadowed, and each element F       of the shadowset (TEST3011 and TEST3012) is itself a volume set.  C       $ MOUNT/BIND=TEST3013 DSA3011/SHADOW=($1$DUA402:,$1$DUA403:), G       DSA3012/SHADOW=($1$DUA404:,$1$DUA405:) TEST3011,TEST3012 TEST3013   0    o  after the shadow copy completes, shut down  8    o  physically disconnect one member of the volume set      o  reboot  G    o  copy the files from the removed member to the new storage at your 
       leisure       o  shut down the application   6    o  BACKUP/SINCE=BOOT DISK:[*...] NEW_FILES.BCK/SAVE  3    o  transfer the NEW_FILES.BCK to the new storage       o  restore NEW_FILES.BCK   2    o  restart the application with the new storage  I If you want to get rid of files which have been deleted in the meantime,  & experiment with an incremental backup.   ------------------------------    Date: 28 Jul 2005 15:25:39 -0700$ From: "AEF" <spamsink2001@yahoo.com>* Subject: Re: Collapsing a bound volume setC Message-ID: <1122589539.084022.157430@g47g2000cwa.googlegroups.com>   / Phillip Helbig---remove CLOTHES to reply wrote: 7 > In article <42E92AB4.5ABAA9B8@teksavvy.com>, JF Mezei ( > <jfmezei.spamnot@teksavvy.com> writes: > > Martin Vorlaender wrote:N > > > Sorry, but no. I have been trying to BACKUP/IMAGE an old volume set to aM > > > shiny new MSA-1000 partition just recently, and BACKUP refused to work: K > > > you can only make an image copy to the same number of disks as are in  > > > the volume set.  > > P > > Interesting. Doesn't a bound volume set appear as a single drive to VMS ? Or6 > > is there an indexf.sys for every physical volume ? > I > The latter.  In fact, there are all the .SYS files.  At the user level, C > yes, it appears as a single drive, but one can access the members * > individually with sufficient privileges. > N > > Could you BACKUP/IMAGE to a save set, and then restore that saveset to the > > MSA-1000 drive ? > > R > > Not using /IMAGE means that when you have alias directories, you end up havingQ > > truly multiple copies of files, as opposed to the same file being copied over P > > itself multiple times (unless you use /NOLALIAS in which case, you only have! > > one copy operation per file).  > A > Yes.  Hopefully, if using /IMAGE, then the setting of /ALIAS is @ > irrelevant (and it should be an error if /ALIAS or /NOALIAS isG > specified), since with an image backup neither ignoring alias entries H > nor having multiple copies of the file are acceptable; one should haveJ > copies of the aliases.  Perhaps /NOALIAS would be more appropriate here,I > since presumably the directories (where the information is stored) will E > be copied properly, and the backup itself will see that the file is - > already there if it tries to copy it again.  > H > The HELP is confusing, frightening, and differs for different versions	 > of VMS.   F Frightening? No. Confusing, yes. But I think the bottom line is to useB /NOALIAS except when restoring from pre-/ALIAS-BACKUP save sets. IF would guess that you could just not use any alias qualifier for V7 andG up. I know for v6.2 there is a small difference between the default and F using /NOALIAS. I get a few extra copies of a few aliased files when ID don't include /NOALIAS. I assume that's fixed in some later version.   ------------------------------  % Date: Thu, 28 Jul 2005 22:16:39 -0400 - From: JF Mezei <jfmezei.spamnot@teksavvy.com> * Subject: Re: Collapsing a bound volume set, Message-ID: <42E99185.96643E2C@teksavvy.com>   Ed Wilts wrote: E > That might be okay if we had rsync for VMS but we don't.  Files are I > being created and deleted so this problem is very difficult.  Quiescing D > the is non trivial because of the large number of files (>250,000)$ > involved (I'm not the developer!).  K Have you considered using the magic of logical names to solve the problem ?   O if you have   MYLOGICAL = $DISK1:[directory]  which contains a gazillion files.   
 You could   7 DEFINE MYLOGICAL $DISK2:[directory], $DISK1:[DIRECTORY]   L Then, the application would be creating new files in $DISK2, but would still) be able to access/delete files in $DISK1.   K You could then use a DCL file that does a copy to $disk2 and then a $delete I from $disk1. (this may have better control over backup since you would be K deleting the old copy sooner. The application would see locked files during . the process if it tries to access those files.  M Once you've migrated all the files over, you can have the logical pointing to  the $disk2 directory.    ------------------------------    Date: 28 Jul 2005 12:58:08 -0500; From: koehler@eisner.nospam.encompasserve.org (Bob Koehler) # Subject: Re: DSSI EISA config files 3 Message-ID: <ttsr9or264nY@eisner.encompasserve.org>   K In article <87fytzcirz.fsf@prep.synonet.com>, prep@prep.synonet.com writes: C > Does anyone have the EISA config file for a EISA-DSSI controller?   6    Didn't know there was such a critter.  Who made it?   ------------------------------  # Date: Thu, 28 Jul 2005 19:05:01 GMT A From: "Colin Butcher" <colin_DOT.butcher_AT@xdelta_DOT.co_DOT.uk> # Subject: Re: DSSI EISA config files < Message-ID: <x%9Ge.78815$G8.7851@text.news.blueyonder.co.uk>  L Don't you mean the ISA configuration utility floppy that you boot from? IIRCD that sets up the EISA config data for the Alpha console firmware (or something like that).    --     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: 28 Jul 2005 13:51:35 -0700$ From: "roger" <roger.tucker@mci.com>2 Subject: How does sys$pal_swasten work on Itanium.B Message-ID: <1122583895.010440.98310@g44g2000cwa.googlegroups.com>  D On alpha, sys$setast() calls the pal code routine PAL_SWASTEN.  ThisG PAL code routine swaps the ast enable bit in the ASTEN register for the G current mode (and current process since it's currently in the CPU).  If D enabled and ast's are pending, then SWASTEN caused an IPL 2 deliveryG interrupt request.  Since PAL code runs with interrupts turned off, and F I assume, ASTs are enabled/disabled for each kernel thread and since aC single CPU can be running only one thread, this is atomic.  I would G also assume, the ASTEN bits are saved back to the thread data structure < if the currently running thread is context switched.  So theG PHD$Q_ASTSR_ASTEN might not match the register for a while, but it will  be updated later.   D OK.  I think I understand this, but correct me if I'm wrong, or feelC free to add any other comments about this.  But my main question is < trying to understand Itanium and this little example routineF sys$setast().  Since there is no PAL code routines on Itanium how doesE the sys$pal_swasten() routine work.  Does it turn off all interrupts? G or take out the sched spin lock to prevent scheduling (i guess it would E just have to raise IPL to sched), or some other method?  I just don't D see it yet, so I'm sure I'm missing something, so any information in* this area might help me understand this...   ------------------------------  # Date: Thu, 28 Jul 2005 19:02:55 GMT A From: "Colin Butcher" <colin_DOT.butcher_AT@xdelta_DOT.co_DOT.uk> U Subject: Re: It's that time again for my quarterly "Where is the CONDIST???" posting. = Message-ID: <zZ9Ge.78813$G8.52822@text.news.blueyonder.co.uk>   I Nothing yet, not even a notification from DSPP EMEA that it's "come out".    --     Hope this helps, Colin. ) colin DOT butcher AT xdelta DOT co DOT uk E It's not mine, but I like this definition: Legacy = stuff that works.    ------------------------------  # Date: Fri, 29 Jul 2005 00:06:51 GMT " From:   VAXman-  @SendSpamHere.ORGU Subject: Re: It's that time again for my quarterly "Where is the CONDIST???" posting. 0 Message-ID: <00A47738.5BA2B362@SendSpamHere.ORG>   In article <zZ9Ge.78813$G8.52822@text.news.blueyonder.co.uk>, "Colin Butcher" <colin_DOT.butcher_AT@xdelta_DOT.co_DOT.uk> writes: J >Nothing yet, not even a notification from DSPP EMEA that it's "come out".  M Jeez Colin, how do you rate?  I get nothing, nunca, nada, ziltch, zero, etc.  L from DSPP.  I wouldn't know they existed had it not been for some posting inK comp.os.vms.  The fact that there is a $524.70 debit in my business account L on 31-March-2005 for the VMS SDK would have me believe that DSPP is still inL existance.  Actually, the latter still reports as Compaq Computer Corp. and + not DSPP (assuming we were being pedantic).    --  K VAXman- A Bored Certified VMS Kernel Mode Hacker   VAXman(at)TMESIS(dot)COM              5   "Well my son, life is like a beanstalk, isn't it?"     ------------------------------  % Date: Fri, 29 Jul 2005 00:22:04 -0400 - From: William Webb <william.w.webb@gmail.com> Y Subject: Re: It's that time again for my quarterly "Where is the CONDIST???" posting. pos 6 Message-ID: <8660a3a10507282122a41b986@mail.gmail.com>  - On 7/28/05, John Santos <john@egh.com> wrote:  > Colin Butcher wrote:L > > Nothing yet, not even a notification from DSPP EMEA that it's "come out= ". >=20H > Me neither, though at least one of my customers has it.  The key might > be that I'm also on DSPP.  >=20> > I did get the VAX  SPL Q2CY2005.  The CD's in it are labeled$ > VAXBINJUN05x, and dated june 2005. >=20$ > Nothing for the Alpha since March. >=20 >=20 > --
 > John Santos  > Evans Griffiths & Hart, Inc. > 781-861-0670 ext 539 >=209 I got both VAX and Alpha 2QCY2005 condists some time ago.    WWWebb   --=20 C NOTE: This email address is only used for noncommerical VMS-related  correspondence. C All unsolicited commercial email will be deemed to be a request for 8 services pursuant to the terms and conditions located at# http://bellsouthpwp.net/w/e/webbww/    ------------------------------  # Date: Fri, 29 Jul 2005 00:16:23 GMT   From: John Santos <john@egh.com>Y Subject: Re: It's that time again for my quarterly "Where is the CONDIST???" posting. pos + Message-ID: <rzeGe.19234$sc3.6433@trnddc07>    Colin Butcher wrote:K > Nothing yet, not even a notification from DSPP EMEA that it's "come out".   F Me neither, though at least one of my customers has it.  The key might be that I'm also on DSPP.   = I did get the VAX  SPL Q2CY2005.  The CD's in it are labeled  " VAXBINJUN05x, and dated june 2005.  " Nothing for the Alpha since March.     --   John Santos  Evans Griffiths & Hart, Inc. 781-861-0670 ext 539   ------------------------------    Date: 28 Jul 2005 12:56:44 -0500; From: koehler@eisner.nospam.encompasserve.org (Bob Koehler)  Subject: Re: LaTeX ---> PDF 3 Message-ID: <1Asi3jk$SDuw@eisner.encompasserve.org>   w In article <dcaf2d$10k$1@online.de>, helbig@astro.multiCLOTHESvax.de (Phillip Helbig---remove CLOTHES to reply) writes: I > I have a somewhat outdated but working LaTeX installation (the [TEXMF]  G > stuff from the freeware CD which Ralf Grtner put together).  When I  J > have time, I plan to update it, but now I don't have the time and don't H > want to mess with anything for fear of changing something which works  > and which I need.  > G > However, I need a method to produce PDF files, either from the LaTeX  F > source or from the .DVI or from the .PS produced by DVIPS.  I would C > prefer some sort of standalone solution, independent of my LaTeX  K > installation (except of course that it would take some file generated by  2 > the LaTeX stuff and transform it to a PDF file). > E    I use DVIPS all the time, but I would have no fear of using DVIPDF &    unless I could prove it was broken.   ------------------------------  + Date: Thu, 28 Jul 2005 18:36:12 +0000 (UTC) P From: helbig@astro.multiCLOTHESvax.de (Phillip Helbig---remove CLOTHES to reply) Subject: Re: LaTeX ---> PDF $ Message-ID: <dcb8ir$5q9$1@online.de>  @ In article <slrndehhjf.16f.thierry@MARS.Family>, Thierry Dussuet" <thierry@dussuet.lugs.ch> writes:   - > ftp://ftp.ctan.org/pub/tex/systems/OpenVMS/  > Q > There is a file called DVIPDF in the archive, which sounds to me as if it would " > create a PDF from a DVI file :-)  P $ type DISK$SOFT:[TEXMF.SRC.PACKAGES.LATEX.CONTRIB.SUPPORTED.HYPERREF]DVI2PDF.;1	 #!/bin/sh  #  # dvi2pdf. SPQR 15/01/95 # > # This is a Unix shell script demonstrating how to use repere.; # It can easily be adapted to a DOS batch file or the like. C # The idea is to take the output from dvihps, preceded and followed > # by copies of the .rep file produced by the hyperref package,A # and process it with the repere program. This puts the main body = # of the PostScript back, but follows it by a set of pdfmarks ! # which form a table of contents.  # 5 # repere is compiled from the Flex program repere.lex  #  F=`basename $1 .dvi` shift  dvihps -z $* $F.dvi -o $F.ps( cat $F.rep $F.ps $F.rep | repere > $$.ps mv $$.ps $F.ps! echo $F.ps ready for distillation    :-(    ------------------------------  + Date: Thu, 28 Jul 2005 18:36:58 +0000 (UTC) P From: helbig@astro.multiCLOTHESvax.de (Phillip Helbig---remove CLOTHES to reply) Subject: Re: LaTeX ---> PDF $ Message-ID: <dcb8k9$5q9$2@online.de>  F In article <opsumlybqrzgicya@hyrrokkin>, "Tom Linden" <tom@kednos.com> writes:   = > On Thu, 28 Jul 2005 13:17:29 +0000 (UTC), Michael Kraemer    > <m.kraemer@gsi.de> wrote:  > J > > In article <dcaf2d$10k$1@online.de>, helbig@astro.multiCLOTHESvax.de   > > (Phillip- > > Helbig---remove CLOTHES to reply) writes: I > >> However, I need a method to produce PDF files, either from the LaTeX H > >> source or from the .DVI or from the .PS produced by DVIPS.  I wouldE > >> prefer some sort of standalone solution, independent of my LaTeX M > >> installation (except of course that it would take some file generated by 5 > >> the LaTeX stuff and transform it to a PDF file).  > > - > > maybe there's a VMS version of "ps2pdf" ?  >   > What's wrong with Ghostscript?  D I have a (rather old) version of Ghostscript installed.  What's the  magic incantation?   ------------------------------   Date: 28 Jul 2005 19:32:46 GMT( From: rf@cl.cam.ac.uk (Robin Fairbairns) Subject: Re: LaTeX ---> PDF / Message-ID: <dcbbsu$96h$2@gemini.csx.cam.ac.uk>   S  helbig@astro.multiCLOTHESvax.de (Phillip Helbig---remove CLOTHES to reply) writes: ' >"Tom Linden" <tom@kednos.com> writes:  ! >> What's wrong with Ghostscript?  > E >I have a (rather old) version of Ghostscript installed.  What's the   >magic incantation?   C if you're running an old enough version of ghostscript, there's not * one.  report which version you're running.   ps2pdf does   Q  ghostscript -q -dNOPAUSE -sDEVICE=pdfwrite -SOutputFile=<output> <input> -c quit   $ which will be something about right. --  6 Robin (http://www.tex.ac.uk/faq) Fairbairns, Cambridge   ------------------------------  + Date: Thu, 28 Jul 2005 19:46:11 +0000 (UTC) P From: helbig@astro.multiCLOTHESvax.de (Phillip Helbig---remove CLOTHES to reply) Subject: Re: LaTeX ---> PDF $ Message-ID: <dcbcm3$ckk$3@online.de>  F In article <dcbbsu$96h$2@gemini.csx.cam.ac.uk>, rf@cl.cam.ac.uk (Robin Fairbairns) writes:   U >  helbig@astro.multiCLOTHESvax.de (Phillip Helbig---remove CLOTHES to reply) writes: ) > >"Tom Linden" <tom@kednos.com> writes:  # > >> What's wrong with Ghostscript?  > > G > >I have a (rather old) version of Ghostscript installed.  What's the   > >magic incantation?  > E > if you're running an old enough version of ghostscript, there's not , > one.  report which version you're running.   GS4_03  
 > ps2pdf does  > S >  ghostscript -q -dNOPAUSE -sDEVICE=pdfwrite -SOutputFile=<output> <input> -c quit   
 I tried this:   C $  gs -q -dNOPAUSE -sDEVICE=pdfwrite -sOutputFile='P2' 'P1' -c quit  $  EXIT   I I see the picture in the window, then it disappears and (as far as I can   tell), no file is created.   ------------------------------  % Date: Thu, 28 Jul 2005 13:55:37 -0400 - From: JF Mezei <jfmezei.spamnot@teksavvy.com>   Subject: Re: Open VMS TCP / Mail, Message-ID: <42E91C18.EF84B2DE@teksavvy.com>   andrew@floatingbear.ca wrote: F > - Pathway for TCP/IP and PONY EXPRESS for connecting VMS mail to the > outside universe via SMTP   E > Not being much of a systems guy (we have a contractor we call in as H > necessary) I am wondering if installing UCX would provide us with bothG > the TCP/IP connectivity and the ability for VMS mail to talk to other  > systems (MS-Exchange),      M Basically yes.  You may wish to contact Process software to see what they may J have to offer too. They provide Multinet and another stack for VMS. Didn'tL they end up buying Wollongong ? You may have some discount in getting a more; modern stack if you already have an old one they inherited.   N the SMTP software that comes with TCPIP Services works, but isn't state of theJ art. It has terrible (useless) logging for inbound emails. Spam protectionC excludes content scanning and is fairly basic by today's standards.    ------------------------------    Date: 28 Jul 2005 14:39:19 -0700$ From: "roger" <roger.tucker@mci.com>< Subject: Re: perfomance question about bool variables in cxxC Message-ID: <1122586759.734561.272070@o13g2000cwo.googlegroups.com>   F This should be simple.  The size of a bool in C++ is a single byte, soB referencing it on alpha, over an int64 will be slower, since alphaG loves aligned quadwords.  (older alpha CPUs don't even have single byte * load instructions - this was added later).   ------------------------------  % Date: Thu, 28 Jul 2005 21:29:03 -0400 ' From: Dave Froble <davef@tsoft-inc.com> < Subject: Re: perfomance question about bool variables in cxx0 Message-ID: <11ej1cp1nrt48d0@corp.supernews.com>   roger wrote:H > This should be simple.  The size of a bool in C++ is a single byte, so  > referencing it on alpha, over    Assuming you mean instead of  ' > an int64 will be slower,  since alpha I > loves aligned quadwords.  (older alpha CPUs don't even have single byte , > load instructions - this was added later).  C The method for quick access for all Alphas would be to insure that  H whatever data is quadword aligned.  If an int64 isn't quadword aligned,  then access will still suck.  I I guess that if the bytes, or even bits, are quadword aligned there will  @ be faster accessing.  One has to wonder about the value of such C nit-picking in most cases.  Do it several billion times, and there   should be room for improvement.    --  4 David Froble                       Tel: 724-529-04504 Dave Froble Enterprises, Inc.      Fax: 724-529-0596> DFE Ultralights, Inc.              E-Mail: davef@tsoft-inc.com 170 Grimplin Road  Vanderbilt, PA  15486    ------------------------------    Date: 28 Jul 2005 15:46:04 -0700$ From: "AEF" <spamsink2001@yahoo.com>A Subject: Re: Rounding v Truncation, was: Re: Platform Support vs. C Message-ID: <1122590764.379165.167990@g47g2000cwa.googlegroups.com>    Dave Froble wrote: > Lurker wrote: K > > "Simon Clubley" <clubley@remove_me.eisner.decus.org-Earth.UFP> wrote in 9 > > message news:V4rQSZ1gJSiV@eisner.encompasserve.org...  > >  > > L > >>BTW, the original comment by Dave Froble was that he could not find a VBM > >>language specification. Neither could I with a brief look, but I did find E > >>the VB.Net specification/manuals which he might find interesting.  > > 
 > > Location:  > >  > >>i > > http://msdn.microsoft.com/library/default.asp?url=/library/en-us/vblr7/html/vboriVBLangRefTopNode.asp  > > : > >>I found the following discussion on the CInt function: > >>@ > >>-2,147,483,648 through 2,147,483,647; fractions are rounded. > >> > >>...  > >>G > >>When the fractional part is exactly 0.5, CInt and CLng always round G > >>it to the nearest even number. For example, 0.5 rounds to 0 and 1.5  > >>rounds to 2. > >  > > H > > Looks like in this case at least VB is in agreement with IEEE specs.B > > See for example http://www.ip97.com/gcc/DEC-Alpha-Options.htmlA > > Since it also comes from the DEC-Alpha options it is at least ' > > somewhat relevant. It says in part:  > > 
 > > [snip] > > $ > > -mfp-rounding-mode=rounding-modeP > > Selects the IEEE rounding mode. Other Alpha compilers call this option -fprm3 > > rounding-mode. The rounding-mode can be one of:  > >   n O > >   Normal IEEE rounding mode. Floating point numbers are rounded towards the O > > nearest machine number or towards the even machine number in case of a tie.  > G > Not excusing myself for not being specific.  From a math perspective, I > the IEEE rounding mode, if it's as described above, is a piece of shit. J >   Consistancy is to be desired, and rounding .5 in both directions basedG > upon it's whole value is not consistant.  At least for any real world  > situations I'm aware of.     See    http://groups-beta.google.com/group/comp.lang.forth/browse_frm/thread/83e9b6468b8274ce/d6da79b3000313b7?lnk=st&q=ROUNDING+NEAREST+EVEN&rnum=10#d6da79b3000313b7   C and start from post number 11. It has to do with "double rounding".  (Sorry about any url wrap.)    ------------------------------    Date: 28 Jul 2005 19:57:50 -0700$ From: "AEF" <spamsink2001@yahoo.com>A Subject: Re: Rounding v Truncation, was: Re: Platform Support vs. C Message-ID: <1122605870.091470.310180@g43g2000cwa.googlegroups.com>    JF Mezei wrote:  > Dave Froble wrote:K > > Regardless, my past experiences are that in a computer language, moving K > > a real number into an integer variable causes any fractional part to be  > > truncated. > ' > Logically, I would tend to agree too.  > N > A real number is comprised of an integer followed by decimals. When you move9 > it to an integer, only the integer part should be used.  > H > truncating is a logical operation. You know exactly what you will get. > T > "rounding" is a subjective operation.  If you have 10.4444444444444444444444444446 > P > Is that machine going to see the 6, and then round each digit as it moves leftO > to come up with integer 11, or will it stop after n decimals and see only "4"  > and keep the integer at 10 ?  E There is no way it will round each digit in the fractional part. What  are you smoking?   ------------------------------  % Date: Thu, 28 Jul 2005 23:31:25 -0400 - From: JF Mezei <jfmezei.spamnot@teksavvy.com> A Subject: Re: Rounding v Truncation, was: Re: Platform Support vs. , Message-ID: <42E9A307.A3BF6949@teksavvy.com>  
 AEF wrote:G > There is no way it will round each digit in the fractional part. What  > are you smoking?   I don't smoke.    N At school, I was told that 7.399999999999999999999 would be rounded to 7.4 and then 7  : 3.499999999999999999999 would be rounded to 7.5 and then 8  N Now, considering how real numbers rarely exactly represent the intended value,M are are you so sure that you can compare exacly against .5000000000000 ??????    ------------------------------  % Date: Thu, 28 Jul 2005 21:41:57 -0700  From: Z <Z@no.spam> A Subject: Re: Rounding v Truncation, was: Re: Platform Support vs. , Message-ID: <osiGe.34790$Iv5.31505@fe02.lga>   JF Mezei wrote: P > At school, I was told that 7.399999999999999999999 would be rounded to 7.4 and > then 7 > < > 3.499999999999999999999 would be rounded to 7.5 and then 8  4 Errr ... it rounds to the nearest multiple of 7?  ;)    P > Now, considering how real numbers rarely exactly represent the intended value,O > are are you so sure that you can compare exacly against .5000000000000 ??????    Yes.  ; Do you understand how reals are represented in binary form?    ------------------------------  % Date: Thu, 28 Jul 2005 19:19:13 -0400 + From: Chip Coldwell <coldwell@gmail.nospam> - Subject: SYMVEC: a VMS linker options utility @ Message-ID: <Pine.LNX.4.61.0507281909410.9850@frank.harvard.edu>  E I wrote a tiny little utility called "SYMVEC" that I have found very  E helpful in porting Unix software that uses shareable images.  SYMVEC  E creates linker options files for shareable images from the component   object modules.   I To create a shareable image, the OpenVMS/Alpha linker utility requires a  G SYMBOL_VECTOR option that declares all of the universal symbols in the   shareable image. For example:   ,      $ LINK/SHAREABLE MY_MATH, SYS$INPUT/OPT      GSMATCH=lequal,1,1000%      SYMBOL_VECTOR=(myadd=PROCEDURE,- %                     mysub=PROCEDURE,- %                     mymul=PROCEDURE,- %                     mydiv=PROCEDURE,- #                     my_symbol=DATA)       Ctrl-Z   E will create a shareable image that exposes the symbols listed in the   SYMBOL_VECTOR option.   I The purpose of the SYMVEC utility is to automate the process of creating  J the linker options for shareable images, in particular the SYMBOL_VECTOR. I It does this by assuming that all global symbols in the object-file-spec  I are universal.  This assumption is not always apropriate; however, it is  3 usually apropriate when porting software from Unix.   < You invoke SYMVEC with a filespec for your object module[s]:        $ SYMVEC FILE.OBJ!      DISK$USER:[RANDOM]FILE.OBJ;1 '      SYMBOL_VECTOR=(OPENOUT=PROCEDURE,- &                     PREFIX=PROCEDURE,-&                     SYMBOL=PROCEDURE,-&                     SUFFIX=PROCEDURE,-'                     CLOSEOUT=PROCEDURE)     D You can either redefine SYS$OUTPUT or use the /OUTPUT option to the G utility to send the output to a file.  The input file-spec can contain  I wildcards (e.g. SYMVEC/OUTPUT=LINKER.OPTS [...]*.OBJ does what you want).   8 SYMVEC is downloadable as source code from my website at  2 http://frank.harvard.edu/~coldwell/vms/symvec.html  J Extract the archive and run "mms" in the resulting directory to build the E utility.  You can use it as a foreign command, or install it in your  I DCLTABLES with the .CLD file that is provided.  There's also a .HLP file   if you want to install that.  G Feel free to contact me with bug reports, suggestions for improvement,  H etc.  My email address is very similar to the one in the header of this  posting.  D Chip "Just Another Hobbyist Trying To Give Something Back" Coldwell.  I P.S. Although SYMVEC is written in C, just for kicks I avoided using the  D standard C libraries (e.g. printf, puts, etc) and made calls to RMS K directly instead.  That was just for practice; I really do know how to use  3 the standard C libraries when I want to.  Honestly.    --   Charles M. "Chip" Coldwell Turn on, log in, tune out    ------------------------------  # Date: Thu, 28 Jul 2005 22:52:25 GMT   From: John Santos <john@egh.com>F Subject: Re: [OT] Rounding v Truncation, was: Re: Platform Support vs.+ Message-ID: <JkdGe.15453$uy3.9114@trnddc05>    O'Brien Paddy wrote: >  >  >  > -----Original Message-----9 > From: Dave Weatherall [mailto:djw-nothere@nospam.nohow]  > Sent: Thu 7/28/2005 3:33 PM  > To: Info-VAX@Mvb.Saic.Com H > Subject: Re: [OT] Rounding v Truncation, was: Re: Platform Support vs. > H > On Wed, 27 Jul 2005 13:18:19 UTC, "Tom Linden" <tom@kednos.com> wrote: >  > I >  > >  Every programming language has it's own set of rules for implicit  >  > >> conversion.  > K >  > > Is the difference here not down to explicit (casted) versus implicit  >  > > conversion? > H >  > Indeed, that is why I said IMPLICIT conversion.  Each language has 
 > its own  >  > set of rules K >  > If you don't exploit EXPLICIT conversion then you had better know the   >  > rules.  > B > Indeed you did Tom. Sorry, I was thinking more about Simon's Ada? > Integer cast when I tapped the keyboard and your example code + > (obviously:-)) fitted in with my thought.  >  > -- > Cheers - Dave W. >  > 
 > ******** > K > I cannot argue with the concept of each language having it's own rules.   I > However, if we are doing integer arithmetic then I firmly believe that   > we should truncate.   F It's scrolled off my news server, and I don't know how to get it back,E but I think the original post wasn't about integer arithmetic per se, A but about implicit floating to int conversions.  If there isn't a F standard, or if the standard doesn't address the issue, I suppose eachA implementation is free to do whatever it wants.  Every BASIC I've A ever encountered truncates on implicit float to ints.  Looks like E VB is different.  There is an ANSI standard for Basic, but it is very B minimal.  I don't know if it even supports ints.  (IIRC, Dartmouth? BASIC only had floats.  I think there was an INT function which 4 returned the integer part of the float, as a float.)  > > In Fortran, I can do things like  IF (a/2*2.EQ.0) blah blah.  : I totally agree in the context of pure integer arithmetic.   > ! > This cannot work with rounding.  > E > Why do other languages need truncation or rounding rules?  Integer  C > arithmetic is Integer arithmetic and in mathematics, it works as  D > truncation.  I learnt trunction rules in integers before I learnt  > rounding rules in Reals.  C Someone else mentioned the IEEE rounding rules, which seem to match C the VB rounding rules.  This is interesting but irrelevant, because C the issue isn't "How do we round?", but "Do we round or truncate?", B and applies to float to int conversions (especially implicit ones,A since if you are doing explicit conversions, you have to tell the C compiler which function to use, so can select the method you want.)   ? I suppose the reason for the IEEE rules is to reduce systematic < errors you get by always rounding .5 up, but I think it justB introduces new systematic errors.  If you are in a situation whereB this matters, then you better account for it explicitly (or credit@ all the 1/2 pennies to my bank account... ;-).  Whole books have@ been written on the subject of numerical methods, and 90% of the> contents seem to be devoted to avoiding or removing systematic? errors that skew the results.  (I am far from an expert on this D subject, but it seems to me the IEEE method, while well-intentioned,A may by disguising the issue and not really solving it, cause more  problems than it cures.)   >  > Regards, Paddy   --   John Santos  Evans Griffiths & Hart, Inc. 781-861-0670 ext 539   ------------------------------   End of INFO-VAX 2005.419 ************************                                                                                                                                                                                                                                                                                                                                                                                                                                      