File ended with 0A instead of 0D 0A

[Jan Vos]

I have problems with files ended with the hexadecimal code 0A.

How do I get files ended with the code 0D 0A with powerbasic 3.5

regards,

Jan

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

[Mel Bishop]

I'm not exactly sure what the question is but you might EOF(x).

open "somefile.txt" for input as #1
do until eof(1)
...process
...process
loop
close #1

I'm pretty sure PB/DOS will process up to the physical end of
the file.


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

[Michael Mattias]

Open for INPUT, LINE INPUT will not work with these files.

You have to open the file in BINARY or RANDOM mode and parse up to the delimiter (x'0A') yourself.

(This behavior was carried into the Windows compilers, so same rules apply there).

MCM

[John Hackl]

That means it is a Unix file. Unix files end with linefeeds (lf CHR$(10)), whereas DOS files end with a carriage return and a linefeed (crlf CHR$(13) CHR$(10)).

Typically people get Unix files through FTP and FTP can be set to do the conversion during the transfer by typing ASCII at the FTP> prompt. That is by far the easiest way to get the file so that it has crlf's.

It is kind of surprising that I wasn't able to find a conversion program in the PB forums, but maybe someone else can. What I used to do was to use a modified version of BUFIN.BAS. BUFIN.BAS can be found here:
http://www.ethanwiner.com/WINER.ZIP

[This message has been edited by John Hackl (edited March 20, 2004).]

[Jan Vos]

this works for me...

thanks for advice.


clean:
open "tmp" for output as #1
open file$ for binary as #2
while not eof(2)
get$ #2, 192, c$
a$ = a$ + c$
q$ = a$ : gosub cleanhead : a$ = q$
again:
p = instr(a$,chr$(10))
if p > 0 and p <= len(a$) then
q$ = left$(a$,p-1)
gosub cleanhead
gosub cleantail
if len(q$) > 0 then print #1, q$
a$ = mid$(a$,p+1, len(a$)-p)
goto again
end if
wend
close 2
close 1
shell "copy tmp "+file$+">nul
kill "tmp"
?"oke file "+file$+" cleaned


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

[Jan Vos]

sorry I forgot something.

cleanhead:
while left$(q$,1) = chr$(13) or left$(q$,1) = chr$(10) or left$(q$,1) = chr$(26) or left$(q$,1) = chr$(0)
q$ = mid$(q$,2,len(q$)-1)
wend
return

cleantail:
while right$(q$,1) = chr$(13) or right$(q$,1) = chr$(10) or right$(q$,1) = chr$(32) or right$(q$,1) = chr$(26) or left$(q$,1) = chr$(0)
q$ = left$(q$,len(q$)-1)
wend
return


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

[Michael Mattias]

>It is kind of surprising that I wasn't able to find a conversion program in the PB forums

If you can find a suitable Win program here (and you can), and get someone to compile it for you, no reason your MS-DOS program can't..

   SHELL "lf2crlf " & myfile$

MCM

[Tom Hanlin]

LF2CRLF is part of the PB/CC samples, which you can find here:

http://www.powerbasic.com/files/pub/demos/

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

[Michael Mattias]

>LF2CRLF is part of the PB/CC samples, which you can find here:

As written, fails if file is CRLF delimited. Uses:

REPLACE $LF WITH $CRLF IN sText

So if your file is CRLF-delimited (PC), it ends up CRCRLF delimited.

Can change easily enough...

REPLACE $CRLF WITH LF IN sText   ' change PC-delimited to Unix Delimted
                                 ' does nothing if file not PC-delimited
REPLACE $LF WITH $CRLF IN sText

MCM

[Tom Hanlin]

I expect you want

REPLACE $CRLF WITH [b]$[/b]LF IN sText

there.

It's true, LF2CRLF was not intended to convert CR/LF files to CR/LF files.
For that, use the COPY command.

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

[Michael Mattias]

What I meant (with typo corrected on $LF) was, a using program could SHELL the LF2CRLF program without knowing or caring how it's delimited: because what comes back is <U>always</U> CRLF-delimited.

i.e., no need to check the delimiters at all.