Conference noted::hackers_v1

>    $ search netserver.log; "DUA1:[SMITH"

	Use the DCL SEARCH command:

	$ SEARCH netserver.log "DUA1:[SMITH]" /output=search.txt


	(try $ HELP SEARCH for more info

>    has all of the f$do_neat_things documented?


	$ HELP LEXICAL

    
    Max Headroom sixel files? Was that a hint that you have them available?
    
    /pjh

    Oh, yes, I posted a note in the TERMINALS notes file a while back,
    there's also mumble.SIXEL.STAR_TREK] if you like.  The *.SVT files
    are video sixels for VT-125s and VT240s (and other closely related
    things) and the *.SPR files are printable sixels for LA-100s and
    LN-03s and such.  Terminal devices other than those listed here
    may give widely varying results, but you get what you pay for...
    
    Um, please don't everyone hit on poor SNDBOX all at once, it's only
    a Vaxstation I and if it gets too many network connection tasks
    at once it tends to become unconscious and need rebooting.  Since
    I have the full network database (lets not discuss the wisdon of
    doing that), it takes about 1/2 hour to reboot.  Late afternoons
    tend to be especially bad times...  One day 6 people decided that
    4:30 (exactly) was a good time to look at sixel files.
    
    In a related note, does anyone have an easy way to limit the number
    of incoming network connections to some small number so that people
    will get "remote node out of resources" error messages?  I don't
    want to limit outgoing links or LAT (we do actually use this machine
    for work related tasks), just to keep SNDBOX from breaking under
    the strain.
                                                          
    Willie

    Adding a little more to what .1 said:
    
    There is no lexical function that performs the same function as
    the DCL SEARCH command.  As suggested, use the SEARCH command
    in your command procedure. 
    
    You can get more information on the SEARCH command and on lexical 
    functions from DCL HELP, or in the DCL Dictionary (Volume 2 of the 
    VMS Document Set).                      
    
    					andrew

    re: .1, .4
    
    	Thanks, I guess I have to search/out and then see if it generated
    the output file, yes?
    
    re: .0, .3
    
    	YOU ARE NOT LISTENING!!!!!!!!!!!!!!!!!!
    
    WITHIN 5 MINUTES OF POSTING THAT NOTE I FOUND MYSELF UNABLE TO LOG
    IN AND ALL MY USERS WERE HUNG.  I FINALLY GOT DOWN TO DCL AND A
    SHO SYS TOLD ME THAT I HAD 9 COUNT THEM 9 FAL_mumble and SERVER_mumble
    TASKS CLUTTERING UP MY SYSTEM, HALF OF WHICH WERE SWAPPED OUT!
    
    They are gone now, but when memory is exhausted that's all she wrote!
    I reserve the right to arbitrarily delete (isn't stop/id wonderful?)
    network FAL requests if the load gets too great....
    
    Willie
    

    Well, it seems to work (even if it does take forever...), but is
    there any way to eliminate the:
    
    %SEARCH-I-NULLFILE, file DUA1:[SCRATCH]NETSERVER.LOG;79 contains
    no records
    %SEARCH-I-NOMATCHES, no strings matched
    
    info messages from going to the terminal?  I've done search/nolog
    but that doesn't seem to have helped.  I suspect I iave to assign
    sys$error or something to null:, does anyone know?
    
    Willie

    To limit links to your node, use NCP.  MCR NCP SET EXEC MAX LINKS nn
    will cut down on unwanted traffic.

re .6

$ SET MESSAGE/NOFACILITY/NOSEVERITY/NOIDENTIFICATION/NOTEXT
$ SEARCH/OUT=...
$ SET MESSAGE/FACILITY/SEVERITY/IDENTIFICATION/TEXT
 

$ SAVE_MESSAGE = F$ENVIRONMENT("MESSAGE")
$ SET MESSAGE/NOFACILITY/NOSEVERITY/NOIDENTIFICATION/NOTEXT
$ SEARCH/OUT=...
$ SET MESSAGE 'SAVE_MESSAGE'

    Thanks much guys, sorting things out by hand was really getting
    to be a drag, especially when I had to wade thru 60+ files a day.
    
    Willie

    You could also define/user both sys$error and sys$output to nl:
    
    p.s. thanks for the sixels.

    Any time!  I've always wanted a frame grabber, and since I had taken
    quite a few neat pictures off the net, I thought I'd return a few.
    I'll post updates here when I get more downloaded (and the 2400
    baud dialins to LTN work again).
    
    Am I correct in assuming that a network copy takes two links?  So
    if I've told NCP to set exec max links to 8 I've limited SNDBOX
    to 4 copiers at a time, yes?  I don't have to set alias max links
    or anything like that do I?
    
    Willie

    If you simply want to limit incoming FAL connects, give the object it's
    own username/password, and use the UAF parameter MAXJOB or MAXDETACH
    (I forget which) to limit the number of active processes. This will
    control the number of processes created using the default access control,
    and leave some of the EXEC MAX LINKS available for your own use.
    
    It's been a while since I played in this area, but I seem to recall
    that if you want to do this, you must define username/password pairs
    for the objects you wish to allow access to, and remove the EXEC
    NONPRIV USER and PASS. I did this on a system a while back for a
    completely differant reason, (SET AUDIT/ALARM/ENAB gives a meaningful
    report when the username part tells you what object is being requested)
    but some of the other benefits are useful as well.

    Dave

Going back to the original question...

You can also do $SEARCH/NOUTPUT <file> <mumble>

and then check $status to see if it found anything. I can't remember what 
numbers come back now, you'll have to do show sym $status after a few 
experiments to find out.

Cheers,

Graham

    Look in VCQUAL::[CRT_PICTURES]
    
    Mike
    

< Note 557.15 by PNO::HEISER "Our ROCK is not as their ROCK !!" >
                               -< More *.PIC's >-

Hi Mike,

	Need the device that the *.pic files are on?

Thanks,

Al.
    

    Try:
    
    	VCQUAL::work1:[crt_pictures]
    
    or
    
    	VCQUAL::crt_pictures:
    
    (just found that leafing thru the old netserver.log files on vcqual...)
    
    Willie
    

< Note 557.17 by 30206::SMITH "William P.N. (WOOKIE::) Smith" >


Hi Willie,

	Much obliged, thanks for the help....got'em...


Al.
   

I don't suggest things like

	$ set message/nothis/nothat/none-of-that-either
	$ do mumble
	$ set message/restore-all-that-stuff

The problem is that if the user ^Y's out of mumble, messages are turned
off without user knowing it.

I suggest

	$ define/user sys$output file
	$ define/user sys$error file		! file can be "nl:" if you like
	$ do mumble

but of course that only works if the "do mumble" runs an image.  Some SHOW
commands don't.  For ones that don't, the following is often available:

	$ show/output=file whatever

another thing that works is

	$ spawn/nolog/in=nl:/out=file do mumble

/Eric

    Sorry gang, SNDBOX is down for a little while for a face lift, so
    you can't do much with the files on the (currently unmounted) drive
    DUA1:, and you can't send me nastygrams cuz MIAL isn't up yet, and
    ....  Well, you get the idea.   I'll post another note when SNDBOX
    is back on the air.
    
    Willie

    OK, SNDBOX has had an upgrade from VS-I to uVax I, and should have
    better response time, and less tendency to try to run everyone off
    of disk.  Please let me know at WOOKIE::SMITH if you have any problems
    getting to any of the files.  I have put 4 new files in
    
    	sndbox::dua1:[smith.imagewise.sixel.max_headroom]
    
    they are:	blankreg.*
    		cheviot.*
    		theora4.*
    		tower.*
    
    I've had a request to get some frames from the new generation of
    star trek, which I will probably do if I have time in the near future.
    One problem that I'm having with all this is that LTN is a toll
    call from where I live, and Kermit (even at 2400 baud) is not a
    very efficient user of bandwidth.  Does anyone out there have any
    routines working that would read:
    
    	1)	8 inch floppies from CP/M in Compupro #3
    
    	2)	5 inch 96TPI CP/M floppies from Avatar TC-I
    		[this has the same physical format as SNDBOX
    			does on it's 5 inch disks, but I don't
    			have anything to read CP/M file formats]
    
    	3)	DC-600A tape cartridge in QIC-120 format?
    
    Give me a hand and I can really churn out the pictures (and even
    take requests), but if I'm on my own, things might go a bit slower...
    
    Well, anyway, enjoy the files, and send me your name if you are
    interested in being on my sixels interest list for updates and such.
    
    Willie

    Hmm, I've got a bunch of Star Trek, the Next Generation and Dr.
    Who and other files, but all told they add up to a bit over 3.5
    megabytes, and even with Bay State, that's gonna cost me $15-20,
    so it might be a little while...
    
    Willie

I think I remember seeing a CP/M format utility in the toolshed.  You may
want to check that out.  I remember that because I discovered it after cracking
a CP/M floppy the "hard" way to get a file off it. (grrr...)

-Mike

    Umm, yes I did once way back when get the CP/M utility from the
    toolshed, but it didn't make much sense to me at the time and was
    for 8 inch drives.  I'm looking for something analogous for the
    5 inch drives...  Any ideas how to convert (subvert?) that utility
    to work on a microvax with RX-50s?
    
    Willie

>Any ideas how to convert (subvert?) that utility to work on a microvax with
>RX-50s?

Find a CP/M machine with both 8" and 5 1/4" drives and use it! :-)

I suppose one could get the source and change the various parameters differing
between 8" and 5 1/4" drives...  Question, if a CP/M disk is mounted foreign on
a VMS system, will VMS accept it at all?  Probably the best way to test this is
fill a CP/M disk with text, mounting it on a VMS machine foreign, then DUMP the
disk and see if you can recognize the text in the dump.

My method of cracking the (8") floppy won't really help you, what I did was
copu the entire disk to a file (which created a file whose records were fixed-
length 512byte blocks), CONVERTing the file to fixed-length 128byte blocks,
then editing the file (!) deleting the lines (blocks) that weren't part of the
program I was trying to get, then putting the blocks in proper order, then
CONVERTing the resulting file to normal text.  Since I was looking for
recognizable text, this wouldn't work so hot for binary files.  I knew nothing
about CP/M file format at the time except it used 128 byte blocks.

-Mike

    Well, I have a CP/M machine with 8 inch 3740 drives, but there are
    a few minor problems.
    
    	1)	I have to boot another operating system to talk to the
    8 inch drives in 3740 format, so it involves another copy step (at
    the very least.
    
    	2)	8 inch drives (in that format) don't hold a lot, and
    I've got a few megabytes worth of sixel files.
    
    	3)	I don't have access to a VAX with 8 inch floppies, though
    I do have access to an RSX machine with 8 inch floppies and a DECUS
    program that is read_only, very slow, and currently hiding somewhere.
    
    	4)	I have access to a microvax with RX-50s that has the
    same physical format as one of my CP/M machines [96TPI, 512 byte
    sectors, 10 sectors per track].  I have dumped a microvax formatted
    disk on my home machine and there's stuff there, though I have no
    idea what it all means.  I'd really like to be able to read and
    write CP/M disks in RX-50 physical format.....
    
    Oh, well, some one of these weeks, I'll get the files over to SNDBOX,
    I added a better color filter to the frame grabber and the pictures
    are noticably sharper.
    
    Willie

I guess I know less about CP/M than I thought I did.  Thought CP/M disks were
always 128 byte sectors.  I do understand ODS-2 somewhat.

Perhaps this is a project for someone.  Either write a VMS program that
understands CP/M format disks or a CP/M program that can write ODS-2 (or ODS-1).

If you have some document describing the format of the CP/M floppies, I'll look
into kludging something together. 

-Mike

    A couple of more files arrived via Kermit (they got caught up in
    some higher priority stuff):
    
            SNDBOX::DUA1:[smith.imagewise.sixel]kermit.*
                                                alf.*
    
    as before, *.SPR is for LN03+ and LA-100 and *.SVT is for VT-125
    and VT-240.
    
    I've got (in addition to those already mentioned) some fairly nice
    shots from a space shuttle tape, coming soon to a node near you.
    Some progress has been made on reading floppies (I have around 20
    disks awaiting), so I'll probably post them gradually when I do
    get to the point where I can read them in.  Actually, there's no
    way I can post 15,000 blocks all at once, so I may start archiving
    some of the older ones, unless someone wants to provide another
    posting location for them on a larger machine.....  SNDBOX is only
    a microvax I with a couple of RD-52s.
    
    Willie
    
    

    re: CPM to DEC
    
    Wasn't there a board that you could plug into a PRO that allowed
    CP/M and POS coexistance? I assume that environment had some sort
    of ability for xfering stuff back and forth. Some where in there
    might be an answer for reading the 5.25" CPM floppies and writing
    ODS1 floppies. Any old PRO hackers dabbled in this area???
    
    Dave

    A couple more files, printable sixels only for now:
    
    	SNDBOX::DUA1:[smith.imagewise.sixel.space.shuttle]*.spr
    
    One shot of the shuttle in landing configuration and one on the
    transport plane.  More later when I can...
    
    Willie

    re:
    < Note 557.14 by IOSG::PYE "Graham - ALL-IN-1 Sorcerer's Apprentice">
                                                                        
    >You can also do $SEARCH/NOUTPUT <file> <mumble>
    >and then check $status to see if it found anything.
     
    Well, I tried that manually and the status seems to return a 1 when
    the string I'm looking for is found, but when I check $status I
    seem to get bizarre results.  Can someone tell me what I'm doing
    wrong?  Reproduced below is my 'manual' method:
    
    $ sho sym $status
      $STATUS == "%X10000001"
    $ sea/nooutput *.*; pfoo
    $ sho sym $status
      $STATUS == "%X08D78053"
    $ search/nolog netserver.log; dua1:[smith /nooutput
    $ sho sym $status
      $STATUS == "%X00000001"
    $ if $status .equ. 1 then write sys$output "gotone!"
    $ if $status .equ. 1 then write sys$output "gotone!"
    $ sho sym $status
      $STATUS == "%X00038068"
    $ search/nolog netserver.log; dua1:[smith /nooutput
    $ sho sym $status
      $STATUS == "%X00000001"
    $ if $status .equ. 1 then write sys$output "gotone!"
    $ sho sym $status
      $STATUS == "%X00038068"
    $
    
    
    Many thanks in advance!
    
    Willie
    

    BTW:  The reason this is coming back to haunt me is that I'm getting
    an increasing number of FAL logs, and my old method of sea/out=mumble
    and then seeing if mumble existed is generating quite a number of
    page faults:
    
    in 00:47:12.21   
        330473 page faults on a uV-I
    
    This makes sorting up to 200 FAL log files a day rather time
    consuming...
    
    Willie
    
    
    

Here is some code that I use to do this sort of thing, which seems to work 
alright.

One of the gotchas that I have been done by is that if you put SH SYM $STATUS
in to check what is going on, $STATUS is overwritten by the output of the SH
SYM! This can get you verrry confused... In that case copy $STATUS into another
symbol, and then test the copy later. I have avoided that in this example by
arranging the logic so only one simple test is needed. 

Good Luck,

Graham Pye.

PS This code is actually not in use anymore, for reasons discussed in the 
COMMAND_PROCEDURES conference...
$!
$! Work out whether I am at home, by looking for the process rights identifier 
$! DIALUP. The only way I can find of doing this is as follows....
$!
$ At_home = "FALSE"
$ show process/privilege/output=sys$scratch:login_dialup_check.tmp
$ on error then goto Dial_search_fail
$ define/user_mode sys$output NL:
$ define/user_mode sys$error NL:
$ search/nooutput sys$scratch:login_dialup_check.tmp dialup
$ if $status .eq. 1 then At_home = "TRUE"
$Dial_search_fail:
$ on error then $
$ if f$search( "sys$scratch:login_dialup_check.tmp" ) .nes. "" then -
      delete/nolog/noconf sys$scratch:login_dialup_check.tmp.

    RE: .31
    
    	I think I can help clear up some of the "mysterious" status
    	values.
    
    	I find status %X00038068 to be:
    
    %CLI-W-OPER, unrecognized operator or expression - check spelling...
    
    	definitely a reference to ".equ." which should be ".EQ." [numeric]
    	or ".EQS." [string].
    
    	It is often helpful to translate $status into a message with
    the f$message lexical, but observe that many (most) vms commands
    (as .33 notes) change $status.  I find it helpful to save $status
    in a temporary symbol if more than one operation against a particular
    $status value is to be done...
    
    				-- Brian
    

    > definitely a reference to ".equ." which should be ".EQ." [numeric]
    > or ".EQS." [string].
     
    Yup, that was it, my fal log sorting routine runs much faster now,
    many thanks for all the assistance!
    
    Willie

    Well, I've finally moved to where work is a local call, and I can
    let Kermit work long nights while I'm asleep, so I've posted the
    following:
    
    	SNDBOX::DUA1:[smith.imagewise.sixel.stng]
    
    A bunch of Star Trek : The Next Generation files where (as usual)
    *.svt are sixels for video terminals (VT-125, VT-240, etc) and *.spr
    are sixels for printers (LA-100, LN03+, etc).  Enjoy!
    
    Willie
    
    Dr Who and more space pictures on the way, some of the older stuff
    may be archived to make space.  Please note that these are not to
    be used for customer demos!

    Rather than storing these images as seperate files in subdirectories,
    you might try creating text libraries, and inserting the images
    as modules therin. You could then use the librarian's DCX compression
    facility to reduce the total space requirements. This might also
    reduce your network load, as (1) a single copy would get an entire
    set of images, and (2) the compressed file means less volume. At
    the receiving end, people could simply extract to their terminal
    for display, or burst the library, as desired.

    Good idea, I'm also considering using the LZW compression routines
    from the toolbox to crunch the files.  The only problem with these
    solutions is that I'll be getting a lot more mail from people having
    problems getting the pictures.  Geez, I posted a note in the Star_trek
    conference and had my note quickly mailed back to me by the moderator
    saying 1) it was in the wrong place, I should have searched the
    entire conference of some 4000 unseen and figured out the 'best'
    place to put it, and 2) it needed better directions for the less
    computer literate among us.  Some one of these years I'm sure I'll
    get back to him on it. :+[  Oh, well, I still get lots of positive
    fan mail too!  I also get about 100 netserver log files per day
    for several days after I post new files.....
    
    Willie

    I'm aware there are a number of other, possibly better compression
    packages out there. The major advantage of the librarian mechanism
    is that it exists on *every* vax, and the decompression is automatic
    on extract. The user instructions (a README.1ST file?) then become...
    
    Copy node::file.tlb to your directory
    LIBR/LIS file.tlb			to see names of pictures
    LIBR/EXT=name/OUT=TT: file.tlb	to display pictures.
    
    Dave

    Hmm, not bad, I'll have to look into building the liabraries in
    the first place, but I suppose it's somewhere in the manuals or
    the help files.....
    
    		THANKS!
    
    Willie

    Re: .39
    
    What is this bias I see hear?  What do you mean by "*every* vax"?
    What you really mean is "*every* VMS system".
    
    If you put them into VMS libraries, then you are going to get alot
    of mail saying "how do I burst the library under ULTRIX, DOS, RSTS..."
    
    The Liszp (sp?) compression utilies would be ok, because at least
    the decompress utility exists under ULTRIX.
    
    How maney VMS blocks total do the files take up?

    re: sizes - 
    
    [.sixel]			1627 blocks
    [.sixel.max_headroom]       2602
    [.sixel.space.shuttle]	253 (will be many more RSN)
    [.sixel.star_trek]		2237
    [.sixel.stng]		4525
    
    I've only got about 8000 free blocks on that drive, so before I
    get too many more files transferred, I'm going to have to do something.
    It may be easier to just dump the older ones to tape and recover them
    on demand.
    
    Willie
    

    re: .41
    mea culpa - I did mean/should have said VMS systems
    
    FWIW - I didn't intend to get in a discussion about alternative
    compression packages. Each have their own set of advantages and
    disadvantages. The major advantage of the VMS Librarian is that
    I think there are probably more VMS systems on the Easynet than
    other opsys's, and that most people have access to one someplace.
    
    I'm sure some hacker out there could put together a com file that
    would burst the library on the VMS system, allowing those so inclined
    to move the individual files to their non-VMS home systems.
    
    Enough on this tangent,
    		Dave

Use Martin Minow's LZ [de]compression routines, and provide a copy of the decom-
pression routine in the same place as the compressed files.

It's simple to transfer such compressed files with Kermit, and lots shorter to
send a 100-block file rather than a 190-block one!  When I'm concerned about
file characteristics, I use VMS BACKUP to build a saveset, then compress the
saveset.  It gets decompressed and broken out at the other end.  Simple, yet
extremely nutritious.  (Note: all this works properly under VMS only with the
current version of BLISS Kermit, 3.3.111.  Earlier versions require a kludge to
transfer compressed files -- which I can provide.)

Librarian compression isn't as effective.

---Pete

    I threw some users off of SNDBOX [how dare they take up my precious
    disk space with their files when I need it for sixels? :+]  ] and
    recovered a bit of space, so the crunch has been averted for a while.
    I'd really hate to be a VMS bigot and blow off any alternate operating
    systems (heck, my favorite node, WOOKIE, is RSX), so for now I'll
    try to keep things on line, and back them up to tape when I get
    low on space.
    
    While I have all of your attentions, is there any fairly easy way
    to sort out netserver.log files to find out which of my sixels seems
    to be most popular?  I've shrunk them a bit with SEARCH/OUT to get
    down to the requestor and file names, but there is still too much
    to sort thru manually.
    
    Willie

Take a look at SCAN to do that, (probably take about 30 lines to get
some totals)

Robb

    OK, I'll byte, what is SCAN?  HELP doesn't know, and there are some
    things in the toolshed, but I'm a bit confused.  Maybe the FILESCAN
    system service?
    
    Willie

SCAN is a string manipulation language that is sorta like pascal, and allows
you to set up patterns that it search's for an executes a `macro' when it 
comes them.  It is a supported DEC prodcuct, and the NET-KIT is available 
somewhere.

If I can find a NetServer.log around, I'll whip up a sample program to show
you.

Robb

    Would you like a squished concatenation of netserver.log's to play
    with?  I've got plenty!
    
    Willie

    Yet Another Batch of sixels are up, space shuttle pics can now be
    found at:
    
    	SNDBOX::DUA1:[smith.imagewise.sixel.space.shuttle]
    
    Where, as always, *.SVT is Sixels for Video Terminals and *.SPR
    is Sixels for PRinters.  Only 2 of these came from my normal setup
    (using Beta tape), all the rest were from a VHS tape I rented, so
    the lower picture quality isn't my fault....  :+)
    
    Enjoy!
    
    Willie