Shivanad Hiremath wrote:

>
> I am facing the following  strange  behaviour of the pascal  'READLN()'  and
> 'READ()' functions on the MPE.
>
> Problem description:
> --------------------
>
> There is a file whose  record size is greater than 255 bytes.  This file can
> be of either fixed length or variable  length  record type.  Now a few lines
> of text is added to this  file.  Some of  these  lines  have  more  than 254
> characters in them.
>
> Also there is a Pascal  program  which I have  written,  that is supposed to
> open the above  mentioned  file for input and display  the  contents of that
> file line by line  (ie.  record by  record)  and also the  total  number  of
> characters in each line.
>
> Now I am observing some strange or unexpected behaviour of this small pascal
> program under the following instances of input file:
>
>     1.  If a line in the input file has less than 254  characters,  then the
>         behavior is correct and as expected.
>
>     2.  If a line in the input file has more than 254  characters,  then the
>         call to the pascal readln() function senses an end of line character
>         after 254th  character  and is ignoring  the rest of the  characters
>         from 255 till the end of that  record.  The  next  call to  readln()
>         reads (at the most 254 chars) FROM THE BEGINNING OF THE NEXT RECORD.
>
>         Even the behavior of the pascal function read() is the same as above.
>         If there  are more than 254  characters  in a line of  record,  then
>         after the 254th call to the read()  function  the eoln() is becoming
>         true.
>
> Any explaination for this strange behavior?
>
[snip]


There were some interesting comments from Stan, Jeff, Duane, ...
about the Pascal I/O performance, nobody however addressed the problem
described: is Pascal/XL 'readln()' and 'read()' NOT ABLE to read textfiles of
records longer than 254 bytes???

As far as I remember the same limitation also holds for 'write()' and
'writeln()'. I would be very interested to know a work-around!

There are many cases where performance is not an issue (could be a seldom
used program), but simplicity of programming is.

A small example:

I want to write into a text file a number of strings of DIFFERENT size
in a 'comma delimited format' (useful to be imported in some PC programs).

With Pascal I/O I can write:

Do
   stringA := ....; stringB := .... ; ....  StringZ := ... ;
   ....

   writeln(stringA, stringB, stringC, ..... ,stringZ) ;
until .... ;

To the performance guru's: how many lines of codes do you
need to program the equivalent of the writeln statement  using 'fwrite's?

How much MORE time do you need to program and to test your code?

How much is the increase in cost for your company? And, finally, ...

how many readln() can your CPU execute until the additional CPU cost
becomes higher than the additional programming costs?

... and what about program readabilty and maintenance, maybe years later,
  by a different programmer?

Kurt Sager


-----------------------------------------------------------------------
Address:                           E-Mail:              Phone:
Kurt Sager                         [log in to unmask]    +41 31 981 0666
SWS SoftWare Systems AG            [log in to unmask]
Freiburgstr. 634                   [log in to unmask]       Fax:
CH-3172 Niederwangen b. Bern  Switzerland               +41 31 981 3263
=======================================================================