Re: Strange behavior of special characters in IFS source files

On 04-Jun-2017 11:16 -0600, Henrik Rützou wrote:

I have a IFS SBCS source file that has CCSID 1252 and the
characters:

// test @ æøå/ÆØÅ

2F2F2074 65737420 4020E6F8 E52FC6D8 C50D0A // test @ æøå/ÆØÅ

When I compile it on a CCSID 277 (danish) machine it becomes this
in the compiler list:

// test Ø {¦}/[@$

@ is in ASCII x'40', in ebcdic-37 x'7C'
that in ebcdic-277 displays as Ø
Ø is in ASCII x'D8', in ebcdic-37 x'80'
that in ebcdic-277 displays as @

Am I missing something or is it only a compiler print problem?

What is described, appears to me, to be a compiler print "problem"; apparently a side effect per EBCDIC CP00500 [coincidentally, a Code Page much like CP00037, for those characters] as choice for representing the data from the UCS2 data. Seemingly, that effect holds, irrespective the Character Identifier (CHRID) specification established for the printer file. FWiW, a CHGJOB CCSID(*HEX) /* or instead, to CCSID(500) or CCSID(37) */ will cause the spooled data to appear with the expected glyphs, for those characters, despite them not being the expected code points.

FWiW, the same effect might be seen using a job CCSID(37) compiling from a PF-SRC member with CCSID(277); the spooled data would be the expected CP00277 code points, but the /wrong/ glyphs would appear when viewing the spooled file (DSPSPLF), because apparently, the CHRID for the printing would have been established from the CCSID of the source.