Re: GCIC Fullscreen applications
"kerravon86" <kerravon86-/[email protected]>
| Newsgroups | gmane.comp.emulators.hercules390.vm |
|---|---|
| Message-ID | <[email protected]> |
--- In [email protected], "mfnoel" <mikenoel37@...> wrote: > > OK, understand that. But I hadn't seen that > kind of thing going on in pdpclib. Maybe I > need to check the source more carefully! stdout is a global variable, for starters. All open files are kept in a static array in fact. No way that will survive. Some of the assembler is non-reentrant too, although it looks like it has been removed from the MVS version. The CMS version has this: MVC ATTNOD,0(R5) ... ATTNOD DC CL4'XXXX' WHERE ORDER MAY BE LIFO OR FIFO. Although even that doesn't appear to actually be used. So I guess it's pretty good. But unless things like stdout are reimplemented to reference a CRAB, with a dedicated register, you're not going to get very far. Note that Robert O'Hara has done some of this reimplementation already in his GCCLIB, and he did it with intrusive code rather than changing the compiler (and losing a register). The intrusive code would appear to be the superior option for the exact situation he has though (PDPCLIB as a shared DLL, basically). BFN. Paul.