On 2014-01-15, at 01:43, Jonathan Scott wrote: > > Constants with type CA are currently translated from EBCDIC > codepage 037 to the 7-bit displayable ASCII codes hex 20 through > 7E, not to code page 819. Anything which does not translate to a > valid ASCII character in that range is left untranslated. > Does this use ASMALTAS, or somthing else, internal?
> (The 7E tilde character was missing until quite recently, which I > spotted and fixed when testing the changes for PM75317, which was > mainly to ensure that the HLASM source code could be stored in a > UTF-8 repository if necessary). > I'm puzzled. Does HLASM itself participate in storing code in a UTF-8 repository, or in extracting such code? Can HLASM directly assemble code in UTF-8 (perhaps on Linux; almost certainly not on traditional systems. > I do not know why this scheme was chosen historically. It does > seem rather US-centric to me. > Me too. Is it described in the Reference? > The ASMALTAS table used for TRANSLATE(AS) converts code page 037 > to code page 819 (ISO 8859-1), using a full 256-byte mapping. > Is this described in the Reference? Does this apply only to C'...', and not to CA'...', as my experiment with a user-defined translate table seems to show. Thanks, gil
