People,
Some clarifications: * I have never worked at IBM, and have never worked on the assembler or binder. However, I have done many things in my 47-year IBM mainframe career (mostly MVS -> z/OS, but with a smattering of VSE, VM, Fujitsu MSP, and even Hitachi VOS3). Currently I am working with a well-known debugging product. I am a voracious reader, and often test things out to fill in any gaps in my knowledge. * I am a 'lurker' on this (and other) list(s). I rarely post, but when I do it often starts something, which is a good thing. * I made a mistake in my previous posting about RMODE64. The ESD object records (and CESD load module records) do not have an 'add 4 to the length flag' In fact, object ESD entries do not have a length field at all. Sorry. Interesting, my testing has shown that you cannot tell the difference in the ESD between ' DC V(FREDV)' and ' EXTRN FREDX / DC A(FREDX)' or ' DC VD(FREDVD)' and ' EXTRN FREDX8 / DC AD(FREDX8)'. This would have ramifications for a disassembler. The ESD record has the assigned ESDID for those entries. * However, the 'add 4 to the length-1' flag *IS* in the object RLD (and, presumably, in the RLD for the load module). My testing shows that the 'add 4' flag is set for AD or VD constants. AD/VD constant RLDS have the assigned ESDID in them (so you can tie the RLD and ESD records together). * Yes, Paul, VD(XYZZY) is valid in the latest versions of HLA, and, as documented above, generates RLD entries with the additional length flag set. Peter Morrison ---------------------------------------------------------------------- For IBM-MAIN subscribe / signoff / archive access instructions, send email to [email protected] with the message: INFO IBM-MAIN
