On Tue, 26 Aug 2008 21:28:33 +1000, Shane wrote: > >I know we keep harping on to IBM (et al) to provide us with >(documented !!!) APIs, but I have to consider this one a less than >outstanding success. And like the CSI, the documentation is (was) >average at best. Playing with it is about the only way to figure it out. >I seem to recall a year or two back I tried the assembler interface, and >even had a lash at the C sample when I was trying to get my head around >that. Don't ever recall having tried Rexx, but there is going to be a >fair slog involved in getting it to work generically methinks. > I've experimented successfully with the C sample. Looking at it, IIRC, the control blocks/reply buffers GIMAPI uses have internal pointers to each other -- a construct particularly hostile to Rexx. Alas.
>Good luck. > Should be a Requirement. As a model, see ICSF, with a relatively Rexx-friendly API. Possible sample in SYS1.SAMPLIB(CSFTEST). All control blocks are passed as arguments by "address LINKPGM"; no internal pointers. And there should be a Requirement for a Rexx facility to support LOAD of repetitively called API type programs to avoid the overhead of LINK within a tight loop, which I suspect the ICSF interface does. Ah, well, there's always LPA. -- gil ---------------------------------------------------------------------- For IBM-MAIN subscribe / signoff / archive access instructions, send email to [EMAIL PROTECTED] with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html
