Re: Inconsistent treatment of EBCDIC Newline (NL) character in codepage.c

According to this link : https://en.wikipedia.org/wiki/EBCDIC_1047

1047 x'15' translates to unicode 8859-1 (819) as X'85'.

Joe



On Mon, Apr 15, 2019 at 6:33 PM ikj1234i-/[email protected] [hercules-390] <
[email protected]> wrote:

>
>
> 1) Ref codepage.c, ebcdic_to_ascii[] translates x'15'
> to ASCII x'0A'.  However, cp_1047_to_819[] (among
> several others) translates it to x'85'.  In some other
> tables NL translates (again improperly, IMHO) to x'07'.
> Could I get an explanation for this, please?
>
> 2) In tk4- when using standard linux telnet client to
> connect to port 37054, there are line end problems
> when using CODEPAGE 819/1047 (due to the x'85').
> The problem disappears when using the default
> code page.  A PCAP trace was used to confirm the
> diagnosis.   This should be easily reproducible...
>
> 3) The convention of having x'0A' advance one line
> without moving the carriage to the left is only one
> possible interpretation.   All unix/linux systems since
> time immemorial have used x'0A' to mean "CR+LF"
>
> Max
>
> 
>