Re: LF translation in QSH

[CRPence <CRPbottle@xxxxxxxxx>]

On 3/9/11 12:32 PM, Dennis wrote:
> <<SNIP>>
> The reason I noticed this is that I sent a file like this to a
> Windows system using ASCII mode, and the Windows system shows no LF
> (of course), and instead views it as one streamed line.
>
> What would be the appropriate CCSID for me to use for such an
> operation?  Following on to that, should I (not) be able to specify
> one CCSID for most purposes and be done with it (assuming I'm only
> using IBM i, Windows and Linux systems in US English)?  Or is CCSID
> even the problem?!?!
> Maybe I should be using something other than "...\n" in my
> printf() ?

  While *nix [and emulation] might be happy with the effects as coded, 
I think for output sent to Windoze, using the "\r\n" in the printf() is 
more desirable; i.e. "\r\n" used to effect CRLF which an application in 
Windoze might prefer [by default].

  However I would have expected 0x0A for ASCII LF as translated from 
the EBCDIC 0x25 LF [Line Feed] which I expected from "\n", not the 0x85 
which appears to be the ASCII NEL [NExt Line] corresponding to the 
EBCDIC 0x15 NL [New Line] control character.  Hmmm, too long since I 
have used a printf() to recall accurately the effects for "\n" in 
EBCDIC... so maybe that will not resolve, even if 0x0D for "\r" should 
make a Windoze [text viewer\editor] program happier or at least the 
output prettier :-)

  FWiW: the choices in ENDLINFMT() as with CYPTOSTMF, reflect the 
difference between what each OS might prefer [by default], and a user 
application might perform or offer something similar for when the target 
of the text data is known.

Regards, Chuck