Announcement

**Aldo Cavini** · 18 Sep 2003, 07:22 AM

Bob,

what I see at the first glance is you use LINE INPUT #2 to get the Rx stream: this means the statemnet will wait until a CR char is received. If for some reason the CR lacks, or the program doesn't read it correctly, the program will be locked. Better input the Rx data on a single char basis, and build the Rx record yourself. This gives you the possibility to discard wrong data and manage timeouts.

Aldo

------------------

**Aldo Cavini** · 18 Sep 2003, 07:36 AM

Another thought. You open the line FOR RANDOM. I never used this option on a serial comm - I use it on disk files to access fixed length record files. For your application is enough to open the com somethilng like

Code:

Open "COM1:1200,N,8,1" As #2

Be careful about the options for the status line handling. Since the line is slow (1200 baud) you can try

Code:

Open "com2:1200,n,8,1,cs,rs,ds,cd" as #2

------------------

[This message has been edited by Aldo Cavini (edited September 18, 2003).]

**Tom Hanlin** · 25 Sep 2003, 03:42 PM

OPEN FOR RANDOM is perfectly appropriate here.

------------------
Tom Hanlin
PowerBASIC Staff

**Aldo Cavini** · 25 Sep 2003, 11:50 PM

Ok, Tom... but, what does it mean? It is not described under the OPEN COM statement; on the other hand, under the normal OPEN statement I read the following:

record size is an integer expression ranging from 1 to 32767, specifying the length in bytes of each record in a random access file. The default record size is 128 bytes.

Hence, if I open a comm FOR RANDOM without declaring the record size, will the program work on a 128 byte length record basis? How can work LINE INPUT (CR terminated record) with a fixed length record? Can I use LINE INPUT on a RANDOM (not sequential) opened file?

I've carefully read the manual, and still have doubts about the strange options on COMM opening. As stated above, the FOR RAMDOM mode with the related RECORD SIZE. FOR OUTPUT: why can I open a comm FOR OUTPUT and also read the incoming chars? The WIDTH statement associated to a comm: how does it work? (the question is still unanswered from a previous forum).

Explanation needed!

Aldo

------------------

[This message has been edited by Aldo Cavini (edited September 26, 2003).]

**Davide Vecchi** · 26 Sep 2003, 01:18 AM

aldo,

if i open a comm for random without declaring the record size, will the program work
on a 128 byte length record basis?

in my understanding, the answer is yes if one uses get# and put#. otherwise, the file / port
should be considered as simply opened for input / output, so - for instance - a line input#
would read from the current position up to the first end-of-line (cr and/or lf), and a print#
would write from the current position.

can i use line input on a random (not sequential) opened file?

i think so; it certainly works on comm ports, as do input$() and input#.

why can i open a comm for output and also read the incoming chars?

i think the answer is the one from tom in http://www.powerbasic.com/support/pb...ead.php?t=1131 :
"...it's an undocumented and unexpected result of an oversight or bug, and can't be relied on
in any particular."

the width statement associated to a comm: how does it work

i join to this question . i'm sure that it works as expected for "lptn:":
if a line is longer than n chars, the line gets broken by inserting a cr (or cr+lf ?)
after the nth char. i was supposing it was working the same way for com, but the
above thread would suggest that this isn't necessarily so.

------------------
davide vecchi
[email protected]

**Aldo Cavini** · 26 Sep 2003, 01:49 AM

Davide,

my questions are bacause I prefer not to try undocumented syntaxes. Until now I always used a RANDOM file with only GET#/PUT#, BYNARY files with only GET$/PUT$, INPUT files with INPUTs and OUTPUT ones with PRINTs. All other combinations seem not to be documented, hence even trying them I can't be sure the behavior will remain unchanged on the next compiler versions.

For instance, if I GET# or PUT# a record to a random opened file, a subsequent PRINT will write the data to the current seek position or to the end of the file?

About COMMs, two possibilities.[list=1][*]FOR RANDOM is a dummy clause which doesn't matter at all[*]it does mean something. In this case, what? Can I GET# or PUT# to a COMM? Which LOCation will be Tx or Rx?[/list=a]

May be my questions are phylosophycal ones, but I'd like to know what does it mean Tom's declaration: "OPEN FOR RANDOM is perfectly appropriate here". In other words, what is the difference between opening the comm using the clause FOR RANDOM or not?

------------------

[This message has been edited by Aldo Cavini (edited September 26, 2003).]

**Criss French** · 26 Sep 2003, 08:52 AM

Sorry, too much of a rush to elaborate much. Here's what I use in an
app that reads a depth sounder at 6 pings (approx 12 bytes per ping)
every second. Runs for hours, never a hitch. BTW runs on 25Mhz 486,
and is actually ON COM / event driven.

I think that setting the ignore errors flags is a good idea. I check for
valid input content/sequences myself elsewhere.

Code:

hcomm$ = "COM2:9600,N,8,1,RS,CS,DS,CD,ME,FE"
OPEN hcomm$ FOR RANDOM AS #6

do...

IF NOT EOF(6) THEN hd$ = hd$ + INPUT$(LOC(6), 6)

...somthing with the data...

...loop.

------------------
What can go wrong will go wrong.
Anything can go wrong.
What hasn't?!?!

**Tom Hanlin** · 26 Sep 2003, 09:04 AM

Aldo, it's a legacy thing. BASIC comm support has been around longer
than BINARY file mode. The only other mode that supports both input
and output is RANDOM. So, you use FOR RANDOM if you need to open a
device for both input and output.

Record sizes and file seek positions are not meaningful when dealing
with comm devices. INPUT$ and PRINT# are fine here.

------------------
Tom Hanlin
PowerBASIC Staff

**Mel Bishop** · 26 Sep 2003, 12:48 PM

My own personal experience from ham radio data communications:
Enter one byte at a time and concanotate the string. i.e.:

Code:

open "COM2:9600,N,8,1,RS,CS,DS,CD" as #1 'ferget 'bout RANDOM, BINARY, or anything else
do
if not eof(1) then 
get$ #1,1,a$              '<-- One byte at a time
if a$ = chr$(13) or _     '<-- Carrage return or
   a$ = chr$(10) then     '<-- Line feed
call DoSomething
else
t$ = t$ + a$
end if
end if
loop

The above example gives you control absolute over incoming data.
If nothing is in the port, go on to something else. If data is
there, get it and process it.

[opinion]
I have found that when dealing with block input, there could be
a control character (Xon/Xoff, Cr$, Lf$, whatever) imbedded in
the string that would require additional code to parse out and
take whatever action is needed. The one-byte-at-a-time eliminates
this and makes for very simpler coding.
[/opinion]

------------------

**Aldo Cavini** · 28 Sep 2003, 10:58 PM

Tom,

thanks for the reply. A question more: I never used PRINT or INPUT$ on a FOR RANDOM opened disk file. Is there a by-design behavior? Will PRINT add chars to the end of the file (as FOR APPEND), or to the beginning (as FOR OUTPUT)? If to the beginning, will the file be truncated? Will the current seek position (or current record) affect the INPUT/OUTPUT statements?

All this is for curiosity; but I think it'd be useful to know what is by design and what is a side effect on this stuff, which is not declared on the manual.

Aldo

------------------

**Tom Hanlin** · 29 Sep 2003, 10:59 AM

Files and devices do not play by the same rules. I would not suggest
using PRINT# or INPUT$ with a file opened FOR RANDOM.

------------------
Tom Hanlin
PowerBASIC Staff

Announcement

Problem reading data from serial port

Problem reading data from serial port

Comment

Comment

Comment

Comment

Comment

Comment

Comment

Comment

Comment

Comment

Comment