[gccsdk] [Bug 240] fork() & prog$heap failure

Ron gettingchoppy at gmail.com
Mon Nov 30 21:36:54 PST 2015


In message <E1a2IBw-0001Bb-V5 at riscos.info>
          bugzilla-daemon at riscos.info wrote:

> http://www.riscos.info/bugzilla3/show_bug.cgi?id=240
>
> Duncan Moore <duncan.moore at gmx.com> changed:
>
>            What    |Removed                     |Added
> ----------------------------------------------------------------------------
>                  CC|                            |duncan.moore at gmx.com
>
> --- Comment #1 from Duncan Moore <duncan.moore at gmx.com> 2015-11-27 04:31:56 PST ---
> Update: This is still an issue with GCCSDK GCC 4.7.4 Release 2.
> However, I've since noticed that !GCC.docs.libunixlib.README specifically says
> that fork() and *$heap does not work:
> "If the program calls fork() then it has to use the wimpslot."
> So this is probably not a bug as such, and can be removed.
>

I don't know what relationship the heap has to tasks, but -
I noticed once that fork() will only work if the window manager is running,
this sort of implied (to me) that it is using the RISC OS task manager, for
a child process. I think this may be reflected by the RISC OS task id's returned
rather than unix style.
I'm guessing the alternative would be a replacement (or ported) task manager
from a unix source to simulate the unix system.
There is a lot of low level unix compatibility stuff I dont know, it's
almost as if I'm missing a book on the subject.
Possibily my lack of programming experience in native RISC OS/wimp doesn't help.
Ron M.




More information about the gcc mailing list