Re: [hatari-devel] A question about the debugger |
[ Thread Index |
Date Index
| More lists.tuxfamily.org/hatari-devel Archives
]
Hi,
On tiistai 15 syyskuu 2015, Roger Burrows wrote:
> On 14 Sep 2015 at 22:14, Roger Burrows wrote:
> > On 15 Sep 2015 at 0:08, Eero Tamminen wrote:
> > > Then you need also to tell Hatari to invoke debugger on
> > > the specified exceptions. Easiest way to do that is to
> > >
> > > add "autostart" option:
> > > hatari --debug-except bus,autostart ./test.prg
> > >
> > > This will enable exception debugging after program
> > > is autostarted, to avoid TOS bootup exceptions invoking
> > > the debugger.
> >
> > The program doesn't get autostarted, perhaps because my C drive is an
> > emulated ACSI drive, not the GEMDOS drive. But I'll see if the
> > debugger still gets invoked on the crash.
>
> It apparently doesn't.
Yes, autostarting relies on hooking TOS *.INF file
loading, which can be done only with GEMDOS drive
interception.
(And the above --debug-except autostart option hooks
into autostarting functionality.)
> > > (Alternatively you can use "setopt -D" command in debugger
> > > to toggle at run-time whether previously specified exceptions
> > > invoke the debugger or not.)
> >
> > Will it work if I add "setopt -D" to the debugger initialisation file?
> > I guess
> > I'll try if all else fails.
>
> I tried this. The debugger is entered when the first bus exception
> happens during EmuTOS initialisation.
And every other matching exception TOS bootup does...
> After typing 'c' to continue, the
> logfile is closed and all trace output goes to the console.
I've fixed this along with symbol trace not going to trace file:
http://hg.tuxfamily.org/mercurialroot/hatari/hatari/
Should be part of next Windows snapshot.
> > > After scripting the debugging, you don't need to interact with
> > > debugger, so you can just start Hatari directly from DOS prompt
> > > and redirect all Hatari output to a file.
> >
> > Ah, now we're back to my original problem: attempting to redirect
> > debugging output to a file did not seem to work for me in Windows.
> > But I don't care as long as I can get a workaround ...
>
> Unfortunately none of the above worked, and trying to redirect stderr via
> 2>debug.log doesn't want to work in the Windows implementation.
I've used stderr and stdout redirecting from Windows DOS
prompt (at work) and it has worked fine for the programs
I used.
What exactly means "doesn't want to work"?
- That only part goes to stderr and part to stdout
- That if you redirect both, nothing is in either of
the files
?
- Eero