Re: error-tokens and locations?

[Ian Badcoe]

At 13:13 17/11/2003 -0800, you wrote:
> Ian Badcoe <address@hidden> writes:
>
> >       The bug is still present under 1.875c, if I build my test
> > program with the enclosed makefile (which differs only in using an
> > extra bash shell to invoke bison) then it builds happily and runs with
> > the same output as 1.875b.
>
> Thanks for looking into it, and for reporting a workaround.  I don't
> know Cygwin myself, but it appears that the problem is something
> fairly basic with invoking Bison, and we'll need a Cygwin expert to
> debug it.  It could be a Bison bug, or it could be a Cygwin bug.
> Perhaps someone else on this list can volunteer to look into it
> further.

Hi,
        I'm glad to help, but have I confused you between the original
locations bug and the coincidental crash bug that I got when I
started building 1.875c?

e.g. My work-around is for the crash, but when I do that, what I am
succeeding at is demonstrating that the original problem still exists
in 1.975c.

Or do you mean that you tried my test-program on another system
and the problem did not recur?  That would surprise me if it was the
case, since it seems to be fairly clearly something to do with the
logic of the error algorithm, and not a detail of file-systems or such-
like.

Not that I would be surprised to be surprised... if you know what I
mean.

Ian B

Free transport into the future, while you wait.