Re: pseudo-maintainer's package unhappy on m86k
On Mon, 15 Nov 1999, Christian T. Steigies wrote:
> If this bug were easy to fix, I would have done so, but it seems it requires
> some c++ knowlegde...
> plus I the error msg did not say much to me, except
> that it does not compile...
The error msg says a non-cost pointer is not a const pointer, so it can't
be passed via reference in place of one. Two things would solve this
error:
1) The pointer in question is changed to const. (i.e.
intel.cpp:1600 is changed from "char *dummy;" to "const char *dummy;")
or
2) The function definition (i.e. physical.h:551) could either have
the const or '&' removed (must change in declaration as well)
Either would fix it. Changing the defining of *dummy is less likely to
kill other code.
> Somehow I have the impression that the code used is arch specific, when I
> see this
> *** [../x11/release/intel.o] Error 1
It was initially written at MIT some years ago on non-x86 machines. Among
others it runs on now are decmips, athena-sun5, alpha, hp[700, 9000...],
powerpc, and rs6000.
> you have an idea what to fix, send me a patch and Ill try it at home. If
Changing line 1600 of intel.cpp as explained above will fix the compile
error in the bug report, but the bug feels too random (why only m68k?).
My gut says it could be a symptom of something else, and the build might
die somewhere else.
> you think you can fix it yourself, we might dig out a temporarily user
> account for you, but Xevil is in non-free, so maybe its illegal
> accoriding to debians new law+order to do xevil development on debian
> machines...
It used to be under the GPL. The author changed this because he quit his
job to work on it full time. I suspect when he gets another job it will
become GPL'ed again.
> Being in non-free its also not important for the debian
> release, its just another bug.
If you guys don't care, I don't care - as you said, I am an "x86 folk". I
just feel it is my obligation as the maintainer to fix it for the m68k
users.
> And Im not going to learn c++ now to fix
> this package.
When I suggested it might be faster for someone with an m68k machine to
fix it than to create me an account, I meant to imply someone who
already knew C++.
I am OK with whatever you want to do. I won't object to making me
a temp account, trying my guess-a-fix, or doing nothing.
Custaba,
Ben
Reply to: