Item    9714325                 89/01/30        10:01

From:   STENO.POOL                      RTC Stenographer Pool

To:     NMORGENSTERN                    N. Leonard Morgenstern

Sub: Minutes of your RT Conference.

>From Steno: Jennifer Marlowe

Date: 01/29/89   Time: 21:45EST
Attendees:
[NMORGENSTERN]

Minutes:
<NMORGENSTERN> How's Chicago?
<TA> hit is going pretty good so far.
<NMORGENSTERN> What kind of work are you doing?
<TA> programming and I am getting a headach with f-pc.
<NMORGENSTERN> I just got it a couple of weeks ago & am looking at it now....
<NMORGENSTERN> What's your problem?
<TA> I am having disk writing problems (hwrite).
<NMORGENSTERN> Let me look it up.
<NMORGENSTERN> "Handle write", is that it?
<TA> I am having problem with it writing a space at the end of the ...
<TA> yes
<TA> of the file.
<NMORGENSTERN> Hmmm
<NMORGENSTERN> Why can't you set up a 1-byte string with a space (BL)...
<TA> when I try to write one byte before that and continue on it stops
<NMORGENSTERN> and send it using adr 1 handle HWRITE?
<TA> at the end of 300 bytes or so saying it will not write any more.
<TA> I am using it as the second file and I send it handle 2
<NMORGENSTERN> BRB I will get my DOS book.
<TA> ok, I only write 120 bytes at a time with crlf at the end.
<TA> and then I want to append another 120 bytes ect..
<NMORGENSTERN> When you say handle 2...
<NMORGENSTERN> do you mean standard output device, which is normally CON ??
<NMORGENSTERN> .
<TA> hold on for a second while I look at something.  .]
<TA> I sent the pointer to the end of file and then ...
<TA> I give buffer 120 file2 hwrite   ....file2 is constant for adr of file2.
<TA> .
<DHWEINSTEIN> is here.
<NMORGENSTERN> Hi Dave.
<[Dave&Friends] DHWEINSTEIN> Hullo
<NMORGENSTERN> Have you had any experience with DOS file handles....
<NMORGENSTERN> We have a problem that needs solution.
<NMORGENSTERN> .
<TA> not too much.
<TA> .
<[Dave&Friends] DHWEINSTEIN> Yes. I like FPC. I like F-PC. I like F-PC.
<NMORGENSTERN> It is on F-PC!
<[Dave&Friends] DHWEINSTEIN> What is the problem with DOS handles in F-PC?
<NMORGENSTERN> TA (I don't remember his first name)
	has a problem appending output...
<NMORGENSTERN> to the end of the file. He will be back in a minute.
<[Dave&Friends] DHWEINSTEIN> Hmmnn...that I haven't done.
<TA> I try to write 120 bytes to a file and it sticks
	an space byte (32) at the end.
<TA> .
<NMORGENSTERN> I know the general concept, but haven't actually used F-PC
<NMORGENSTERN> Why 120 bytes?
<[mitch] TA> it is for a formula and It is 120 bytes
	long and the next one has ...
<[mitch] TA> to be on the next line .
<[mitch] TA> .
<NMORGENSTERN> Do you have any ideas, Dave?
<NMORGENSTERN> (my way of saying, I don't know. :-)
<NMORGENSTERN> I guess that what you want is to send 120 bytes, then a CRLF...
<NMORGENSTERN> but you are getting a blank in addition or instead.
<NMORGENSTERN> .
<[mitch] TA> I may break down and save everything in
	memory and then write a complete file...
<[mitch] TA> .... no the 120 byte array includes crlf
	and it adds the sp on it.
<[mitch] TA> .
<NMORGENSTERN> In other words after the CRLF
<NMORGENSTERN> .
<[mitch] TA> yes      .
<NMORGENSTERN> Let me check something.
<NMORGENSTERN> .
<[mitch] TA> I am using version 2 of f-pc.
<[mitch] TA> .
<NMORGENSTERN> How do you issue the CRLF?
<[Dave&Friends] DHWEINSTEIN> has left.
<NMORGENSTERN> Do you use the word CR?
<[mitch] TA> no I did a C! of 0D 0A at the end 2 bytes of the buff.
<[mitch] TA> .
<NMORGENSTERN> How do you mean C! -- that moves into memory.
<NMORGENSTERN> .
<NMORGENSTERN> I get it, you move those bytes into the buffer! :-)
<NMORGENSTERN> then send the whole package.
<[mitch] TA> 13 buff 118 + c! ...
<[mitch] TA> 10 buff 119 + c! ... then buff 120 handle2 hwrite.
<[mitch] TA> .
<NMORGENSTERN> That should work. Are you sending the
	output to a printer or to a file?
<NMORGENSTERN> .
<[mitch] TA> to a disk file .
<[mitch] TA> .
<NMORGENSTERN> Very puzzling. Let me play with it ...
<NMORGENSTERN> If I solve it, I will send you a message.
<[mitch] TA> ok.  I am probably just store every line in mem then ...
<[mitch] TA> .. write to the file the whole works.
<[mitch] TA> .
<NMORGENSTERN> That would be more efficient, but it shouldn't make any...
<NMORGENSTERN> difference to the output.
<APPERT> is here.
<NMORGENSTERN> Kevin! Mitch here has a problem with F-PC...
<NMORGENSTERN> Weren't you one of those who worked on its develpment?
<NMORGENSTERN> .
<[Kevin] APPERT> nope.
<[Kevin] APPERT> I know a little about it
<NMORGENSTERN> Are you familiar with DOS I/O?
<[Kevin] APPERT> whats problem?
<NMORGENSTERN> Let's see if I have it right....
<NMORGENSTERN> He sets up lines of 120 bytes, the last 2 of which are...
<NMORGENSTERN> CR and LF, that is 13 10...
<NMORGENSTERN> Then he sends 120 bytes to a disk file...
<NMORGENSTERN> and gets an extra space right afterwards...
<NMORGENSTERN> Mitch, does that show up in the file as ASCII 32?
<[mitch] TA> correct.
<[mitch] TA> .
<[Kevin] APPERT> reduce the count by one (size of write) and what happens
<[Kevin] APPERT>    ?
<[mitch] TA> I haven't tried that yet.
<[mitch] TA> .
<[Kevin] APPERT> are you sure you are counting off
	the location of the cr/lf correctly?
<[Kevin] APPERT> the first location  of your buffer is offset 0...
<[mitch] TA> yes since the buff starts at 0 and I write to the ...
<[mitch] TA> 120th byte.
<[Kevin] APPERT> and the size of the write counts that as location 1
<[Kevin] APPERT> I think that would produce the opp. effect, though
<[mitch] TA> no it starts writing at the 0 offset.
<[Kevin] APPERT> which do you write at offset 120 (decimal?)
<[mitch] TA> there is something mentioned in the hlp files that...
<[mitch] TA> mentions about segments... 13 in decimal.
<[mitch] TA> .
<NMORGENSTERN> If the buffer is 120 bytes...
<NMORGENSTERN> then the start is byte 0 and the end is byte 119.
<NMORGENSTERN> .
<[mitch] TA> it is allotted 121 bytes
<NMORGENSTERN> Is byte 0 a length byte?
<[mitch] TA> and I tell it to write 120 bytes but crlf is in the file.
<[mitch] TA> no.
<[Kevin] APPERT> no, the forth word takes the length on the stack
<[mitch] TA> is there any other seq. write words?
<[Kevin] APPERT> I'd have to explore a bit to answer that
<NMORGENSTERN> Try this...
<NMORGENSTERN> buffer 130 dump
<NMORGENSTERN> and see whether there is a blank right after the CR LF...
<[Kevin] APPERT> have you tried numbering the characters,
	writing out the numbered buffer, and seeing exactly how many you have written?
<NMORGENSTERN> that you are sending.
<[mitch] TA> I will when I get a chance don't have access at moment.
<[Kevin] APPERT> good luck
<NMORGENSTERN> I find I/O very puzzling!
<NMORGENSTERN> especially printers.
<NMORGENSTERN> But also file I/O etc. We are burdened...
<[Kevin] APPERT> I like block.  I understand block.
	 I can explain it in about 5 min.
<[mitch] TA> you cannot write to much seq. if you start before...
<NMORGENSTERN> with the history of obsolete machines such as the teletype.
<[mitch] TA> the end of the file.
<NMORGENSTERN> A question:
<NMORGENSTERN> When DOS positions itself at the end of a file...
<NMORGENSTERN> does it go the the current end or 1 past the end?
<NMORGENSTERN> that shoule be "to the"
<[mitch] TA> I dont have an ibm so .....
<[mitch] TA> I cannot answer that.
<[mitch] TA> .
<NMORGENSTERN> All DOS machines should be compatible at that level, I think.
<NMORGENSTERN> .
<[mitch] TA> I only wish.  I use two different forths and sometimes...
<[mitch] TA> it dirves me up the wall.  I implemented $" for f-pc...
<[mitch] TA> because I had it on the other and no way of putting a...
<[mitch] TA> string addr and count on the stack.
<[Kevin] APPERT> I usually diddle this sort of thing
	untill it works.  That's what forth is for.
<[Kevin] APPERT> I gotta scoot
<NMORGENSTERN> What is $"
<NMORGENSTERN> Bye Kevin
<[Kevin] APPERT> nice meeting you , Mitch
<[mitch] TA> bye
<[Kevin] APPERT> has left.
<[mitch] TA> it is...
<[mitch] TA> ex:   $" hello there" would leave an adr and count....
<[mitch] TA> on stack that you can do from console.
<[mitch] TA> it comes in handy.
<[mitch] TA> .
<NMORGENSTERN> How does it differ from " "
<NMORGENSTERN> .
<[mitch] TA> " works in colon definition only.
<NMORGENSTERN> Got it.
<[mitch] TA> at least that is what i think I found out.
<NMORGENSTERN> You are right. " "  ." " etc. are only within colon defs.
<[mitch] TA> And when you do test with strings it comes in really handy.
<NMORGENSTERN> F83 has .(   ) which will print a contained string.
<[mitch] TA> thanks....it drove me up the wall for a while.
<[mitch] TA> do want to print just store in a temp buff and put adr c on stack
<[mitch] TA> .
<[mitch] TA> but that is nothing like this disk file problem though.
<NMORGENSTERN> FF-PC also has .( )  They output immediate
<[mitch] TA> I had spent 5 hours on it.
<NMORGENSTERN> Right, very puzzling. As I say, I will play with it...
<NMORGENSTERN> & see what I can come up with.
<[mitch] TA> the formulas cannot have a space on the first char.
<NMORGENSTERN> What operating system do you use?
<[mitch] TA> I need to speed it up anyway so I will write to a 32k buff.
<[mitch] TA> at work it is 3.? and 3.3 for when I am emulating.
<NMORGENSTERN> I have 3.2
<[mitch] TA> here at home.
<[mitch] TA> woops it is dos 3.2 here at home and probably at work to.\
<[mitch] TA> I dont pay too much attention to that.
<NMORGENSTERN> It shouldn't make any difference. The handle system started...
<[mitch] TA> if this keeps up I may go back to c.
<NMORGENSTERN> with DOS 2.0 and I don't think it has changed much.
<NMORGENSTERN> Try dumping your buffer and counting bytes.
<NMORGENSTERN> I'll bet a miscount is sneaking in somewhere.
<[mitch] TA> I will.  what the diff between v3 and v2 of f-pc.
<[mitch] TA> I have to have an exact count or the formula is no good!!
<NMORGENSTERN> The latest version is 2.25, which is
	what I have. I got it from Ting 2 weeks ago.
<NMORGENSTERN> They have decided that will be IT! for a while.
<[mitch] TA> what is the improvements?
<NMORGENSTERN> I doubt there will Don't know -- I suspect
	none in the area of I/O...
<NMORGENSTERN> which tend to relate pretty directly...
<NMORGENSTERN> to the DOS system.
<NMORGENSTERN> .
<[mitch] TA> ok.  since I have to pay for d/l it I didn't bother yet.
<[mitch] TA> .
<NMORGENSTERN> Wait, if you can. There are always minor bug fixes, etc. etc...
<NMORGENSTERN> in even the best programs.
<NMORGENSTERN> .
<[mitch] TA> I know but I dint feel it worth it yet since I am paying for it..
<[mitch] TA> my company wants me to use C but are letting me us f-pc.
<[mitch] TA> .
<[mitch] TA> boy I am mispelling left and write....ment use not us.
<[mitch] TA> .
<NMORGENSTERN> I'll upload E-Mail to report my progress
<NMORGENSTERN> .
<[mitch] TA> ok.  I might sit down tonight and write a little code.
<NMORGENSTERN> I have to be leaving soon. The
<[mitch] TA> I am having a little problem with abort. but I don't need it yet.
<NMORGENSTERN> Nice thing about Forth is that it links easily to assembler.
<NMORGENSTERN> I'd like to talk about that, because error handling is...
<[mitch] TA> yup.  things just have not gone right today.
<NMORGENSTERN> probably the most difficult problem in any computer language...
<NMORGENSTERN> ABORT" is primive. It just starts over....
<NMORGENSTERN> primitive, that is...
<NMORGENSTERN> More sophisticated handling rapidly gets into advanced areas...
<NMORGENSTERN> What is your specific problem?
<[mitch] TA> it works everywhere else execpt at one place .
<NMORGENSTERN> Are you using ABORT" or ABORT  ??
<[mitch] TA> abort"
<NMORGENSTERN> Remember it requires a flag on the stack.
<[mitch] TA> yes I put one on there and it still didn't work.
<NMORGENSTERN> Strange.
<[mitch] TA> I am tring to find it now I think I switch to quit.
<NMORGENSTERN> Try putting .S in just before you call it to see what the stack 
was at that moment.
<NMORGENSTERN> .
<[mitch] TA> I will
<NMORGENSTERN> QUIT doesn't reset the stack, but ABORT" does.
<[mitch] TA> just a temp fix.
<[mitch] TA> it was in a begin while loop and if statement.
<[mitch] TA> it was only for if you selected the wrong file to convert
<NMORGENSTERN> DOS distinguishes between CON the console and start over
<NMORGENSTERN> Watch out for ABORT" in an IF statement!. Normally...
<NMORGENSTERN> it doesn't need one....
<[mitch] TA> I may take it out then.
<NMORGENSTERN> That is, if you write IF ABORT" Goofup " ..... THEN...
<NMORGENSTERN> wyou will not get the desired effect.
<[mitch] TA> the if that is.  I used if -1 abort"    jflsjf" then ....
<[mitch] TA> .
<NMORGENSTERN> You could use ?DUP tHAT SHOULd work!
<[mitch] TA> I will. do you have the assembly commands for f-pc?
<NMORGENSTERN> No but I have F-PC technical reference manual in addition...
<NMORGENSTERN> to the User's manual.
<NMORGENSTERN> The assembler is on disk, I believe.
<[mitch] TA> I dont have either manual.  well I need to get going.
<NMORGENSTERN> Me, too. ...
<[mitch] TA> I want to thank you for your time and trouble.  send
<NMORGENSTERN> Hope I have been helpful & I will...
<NMORGENSTERN> get bactto you.
<[mitch] TA> some email to ta sometime.
<[mitch] TA> ok, thankssssssssssssss!!!!!!!!!!!!!!!!!!!!!!
<NMORGENSTERN> Right. You leave first. I hve to close down.
<NMORGENSTERN> .
<[mitch] TA> bye for now.
<[mitch] TA> has left.
=== End of Steno notes. ===
Note sent Monday Jan 30
TA
FIGSYSOPS$
F-PC I/O
  I actually used F-PC for the first time today to try to duplicate your 
problem. I tried your method of using C! to insert 13 10 in positions 118 and 
119 of the buffer, then used
    adr 120 handle HWRITE to send it to the file. 
   (It is not necessary to reposition the pointer, as this is done 
automatically by DOS.) After writing a couple of lines, I finished off with a 
control-Z and closed the file. Next, I exited, and examined the file with 
Norton Utilities. It all looked ok.
 I wonder if your problem might be in reading the file. Lines terminated with 
CRLF are read with LINEREAD ( -- adr len). This uses an internal buffer, and 
returns the address. The documentation reads, slightly edited, "Read one line 
from the current file, buffered by INBUF ... When switching to a new file, use 
     0.0 handle MOVEPOINTER
to reset the file pointer to the beginning of the file, and execute IBRESET so 
the next LINEREAD will cause a read from the disk file. The read line length 
is limited to 250 bytes. Lines read with LINEREAD are terminated with a LF 
byte (10)'
 	The assembler is called PASM. It is an integral part of F-PC and is well 
documented in the manual. The latest version is "frozen" at 2.25. The manual 
can be obtained from the Forth Interest Group. I saw it mentioned in my 
Forth Dimensions that just arrived today. The cost is $75 for the Users' 
manual, the technical manual, and all the source files on 4 disks filled 
almost solid with .ARC files. If you are doing serious programming in F-PC I 
strongly advise you to purchase it. A bargain for power!
*S