Re: Sorry, does not tell much

Hi,

I just tested Borland C 3.1 in IDE mode and it runs, but
with the new kernel "2027rc" ( it was working with "2027test")
I am using the newest fdxms (18-9-02) and lbacahce.

I remember that I had problems with some later version, so
I supose that now TC30 should work too :)

Maybe it is just time for a new release of Kernel+FreeCom+fdxms

Alain

> Hi Dos gurus,
> 
>     I have a stubborn yet perhaps interesting problem trying to get the 
> Borland Turbo C++3.0 for DOS( it actually says "for DOS" on the 3.5" 
> installation diskettes) executable "TC.EXE" to run.  The command line 
> tools, "TCC.EXE" and "TLINK.EXE", however, run fine.  Boreland Turbo 
> C++3.0 is entered in the compatibility list, so I was hopeful about 
> getting it to run in FreeDOS.  But I have not had any luck yet gettng 
> TC.EXE to run(maybe they were referring to the command line tools?). 
>  After some days tinkering with this thing, I tried running it within 
> the xdosemu in Linux where the program still failed as before, but this 
> time I ran the dosdebug utility.   I sent the errors output by dosdebug 
> to Eric, and his comments  and  the errors follow thusly:
> 
> >Hi, here my analysis:
> >
> >
> >>General Protection Fault, system state: stopped
> >>AX=36b0  BX=00ed  CX=ffff  DX=0000  SI=0806  DI=0abe  SP=07b6  BP=07c4
> >>DS=00ed  ES=0040  FS=0000  GS=0000  FL=0897
> >>CS:IP=0000:ffff       SS:SP=00ed:07b6
> >>
> >
> >This means that your program started to jump to a position where it is
> >not supposed to jump at all (CS: always ends at FFFF, so you cannot go
> >on from there - in real mode, you may be able to reach the 64k+1 byte
> >point, and on an 8086, you would wrap around to 0 here).
> >
> >Probably something else went wrong longer before the G.P.F. happened,
> >for example a stack overflow. The crash seems to happen while still in
> >v86 mode, so protected mode was not even reached. It seems that the
> >program runs in ED:xxx and tried to access data at 40:xxx, maybe you
> >can find anything interesting by reading the data and code there, or
> >the content of the stack (data from ED:7B6 ... ED:xxx). 
> >
> TC.EXE is the largest of the three >~  1.2 Mb and should present the ide 
> version of tcc. All three are, according to the included documentation, 
> supposed to run in protected mode.  This protected mode can be provided 
> if necessary by a dpmi host in the directory called "dpmires".  The docs 
> recommend adding programs to the dpmihost by running another included 
> program called "dpmiinst" if you run into certain errors.    I've tried 
> that as well with only the same  results:  totally crashed sytsem - 
> keyboard inoperable and an interesting yet otherwise nonfuncional 
> "psychedelic" pattern on the console.  
> 
> Any hints or suggestions by anyone, especially if they have had success 
> with this particular package, would be greatly appreciated.
> 
> B.T.W. Sorry to interrupt the great " Pascal Interrupts debate. " ;-)
> 
> George
> 
> ----------
> list options/archives/etc.: http://www.topica.com/lists/fd-dev
> unsubscribe: send blank email to: [email protected]
> 
> 
>