I'm not sure if it is related, but the IBM C library has a wierd feature (bug) - search the IBM C/C++ Programmers Guide for "_EDC_ZERO_RECLEN". Warning: the documentation is pretty confusing.
On Fri, Jan 11, 2013 at 10:02 AM, Paul Gilmartin <[email protected]>wrote: > On Fri, 11 Jan 2013 08:18:37 -0500, Steven St.Jean wrote: > > >Gil, > > > >> It would seem so, but as I reported earlier in this thread, it doesn't > >> work. > >> Empty records are corrupted; I don't get "the file back exactly as it > >> started out". > > > >Sorry, I missed that. > > > >> I suspect this is purely a z/OS flaw, > > > >I agree. If it does indeed work (or fail to work) as you describe, I > would open a PMR with z/OS Comm Server. > > > WAD is highly likey for this. It's easy to suspect that there's > a common subroutine to add record to target file, called for > all transfer modes, that converts empty records to spaces. > And that behavior may have been inherited from VM/CMS > where empty records are strictly prohibited by the file system. > > And they might consider it an incompatible behavior. It might > break end user code that does LH; BCTR; EX; MVC to copy > variable length records. Same reason that ISPF/PDF changes > empty records even with SET PRESERVE ON. > > -- gil > > ---------------------------------------------------------------------- > For IBM-MAIN subscribe / signoff / archive access instructions, > send email to [email protected] with the message: INFO IBM-MAIN > ---------------------------------------------------------------------- For IBM-MAIN subscribe / signoff / archive access instructions, send email to [email protected] with the message: INFO IBM-MAIN
