Contact to Metacomco (and Lattice!) because of their C Development Kit? (The Lattice C compiler)

[ql_freak]

Hi,

The source code to C68 is available, could that be enhanced to produce reentrant code?

Hi Derek,

look in the current C68 manual. C68 can produce reentrant code, with similar options as in Lattice C. (Code size with all static variables/data restricted to 32 or 64 KByten, not sure).

The question is: Has these (available) options from C68 been tested in the QDOS version of C68. They are not used in the libraries delivered with C68 (and so the code is not tested). Lattice was the most famous compiler at tits time, (Atari, Amiga, DOS, embedded, ...) and so it's reentrant options seem to work (Erling has even implemented a full K&R compatible library for the QL). But is this true for C68? Has anyone tried these options and has checked the produced code?

[bwinkel67]

...

"Pure" programs are programs that can be started through the Hotkey System multiple times and can share the same instance of code, but have a private data space - They have basically the same restrictions as ROMable code.

...

"Impure" programs are in no way any "worse" than "Pure" ones. Just different. There is actually not much point to desire only "pure" programs on the QL - You have to invest a large amount of effort if you aim for them in anything but assembly. And, I'd rather have a working, "impure" one, than a non-finished "pure" one.....

That was a very informative rundown. One question, if you EXEC the same program twice, I'm assuming the code isn't shared and each process runs its own copy.

[tofro]

The question is: Has these (available) options from C68 been tested in the QDOS version of C68. They are not used in the libraries delivered with C68 (and so the code is not tested). Lattice was the most famous compiler at tits time, (Atari, Amiga, DOS, embedded, ...) and so it's reentrant options seem to work (Erling has even implemented a full K&R compatible library for the QL). But is this true for C68? Has anyone tried these options and has checked the produced code?

I have and it works (with a few quirks). You need your own startup code and your own libraries (simply re-compiling is not enough, as some libs use static variables)

[tofro]

...

"Pure" programs are programs that can be started through the Hotkey System multiple times and can share the same instance of code, but have a private data space - They have basically the same restrictions as ROMable code.

...

"Impure" programs are in no way any "worse" than "Pure" ones. Just different. There is actually not much point to desire only "pure" programs on the QL - You have to invest a large amount of effort if you aim for them in anything but assembly. And, I'd rather have a working, "impure" one, than a non-finished "pure" one.....

That was a very informative rundown. One question, if you EXEC the same program twice, I'm assuming the code isn't shared and each process runs its own copy.

The code sharing between multiple instances of "pure" programs works with the HOTKEY system only. Simply EXECing the program twice creates two instances (EXEC can't now what you're running now is the same that you started ten seconds ago and has no correlation between the two).

[ql_freak]

The question is: Has these (available) options from C68 been tested in the QDOS version of C68. They are not used in the libraries delivered with C68 (and so the code is not tested). Lattice was the most famous compiler at tits time, (Atari, Amiga, DOS, embedded, ...) and so it's reentrant options seem to work (Erling has even implemented a full K&R compatible library for the QL). But is this true for C68? Has anyone tried these options and has checked the produced code?

I have and it works (with a few quirks). You need your own startup code and your own libraries (simply re-compiling is not enough, as some libs use static variables)

This is very nice to hear :-D Static variables are possible. There is an option for it in C68 (as in Lattice C), but then the code size (or size of static data? Not sure) is restricted to 32 or 64 KBytes (this is also the case with EJC if you want to create reentrant [pure] code).

Not really a problem for programs which should be pure (can be HOT_CHPed without the Igitit option). My rn program (created with EJC) is about 12.5 KBytes - and it does some complicated stuff, a "hello world" (i.e. with included <stdio.h>) is less than 9 KBytes :-)

[ql_freak]

The code sharing between multiple instances of "pure" programs works with the HOTKEY system only.

But you can make an executable Thing from any executable program (with e.g. HOT_CHP) and there is EXEP (which unfortunately doesn't support all the features of EX), but AFAIK at least command strings can be passed with EXEP. And EXEP can start executable Things(!). So it should be possible, to HOT_CHP a program, which must be started more than once and then call it with EXEP from your program.

IMPORTANT NOTE: With impure (Igitit option) programs you must use HOT_CHP1 or use the "I" option in HOT_CHP! All C programs (except those compiled with EJC in standard config) are impure programs. BTW: All PSION (office) programs are pure programs(!). I don't know if XCHANGE is also a pure program (but as it can load multiple documents, it doesn't need this feature IMHO).