Date: Fri, 25 Jul 1997 08:52:15 +1100 From: Bill Currie Subject: Re: BCSIO trouble In-reply-to: <33D72571.4DD4A21C@LSTM.Ruhr-UNI-Bochum.De> To: Thomas Demmer Cc: djgpp AT delorie DOT com Message-id: <199707242056.IAA07252@teleng1.tait.co.nz gatekeeper.tait.co.nz> Organization: Tait Electronics Limited MIME-version: 1.0 Content-type: text/plain; charset=US-ASCII Content-transfer-encoding: 7BIT Comments: Authenticated sender is Precedence: bulk > Hi all > what is wrong with this piece of code: > > /* Initializing stuff like in bcsio20/test.c */ > while( EOF != (c=fgetc(f))){ > sio_put(comPort,c); > while( sio_senddone( comPort ) ); > } This looks fine, though I would have used `while (sio_put(comPort,c));' (better with a timeout of course) to make sure the `c' got into the buffer. > The program immediately crashes with an SIGSEGV > in sio_senddone(). Umm... did you reassemble serio.o? There's a bug in gas (including 2.8.1) that causes problems with serio.s. I have the patch (should have been in the zip) to fix this bug, but it's for the gas source from binutils 2.8.1. > The supplied terminal program also crashes, but mostly > with a GPF in RMCB. > Any idea what might be wrong? Did you lock the memory? serio.doc says how at the bottom. Unfortuanatly, there's an error in both the doc file and term.c: the line is supposed to read __attribute__((constructor)), I forgot the 's' in `constructor'. > BTW, I had to update to NetScape 4.0, so if this post > contains HTML, I am sorry, send me a personal flame, or > even better, how to switch this idiotic feature off. I coped, (cut, paste, save, ....). Bill -- Leave others their otherness.