[Top][All Lists]
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: [Freeipmi-devel] Major symbol exporting problem
From: |
Albert Chu |
Subject: |
Re: [Freeipmi-devel] Major symbol exporting problem |
Date: |
Fri, 02 Apr 2004 15:26:43 -0800 |
> I like the current approach you took. ipmi_ prefix. Simple and easy.
In the future, should we just remember to prefix every non-static
function with "ipmi_".
Al
--
Albert Chu
address@hidden
Lawrence Livermore National Laboratory
----- Original Message -----
From: Anand Babu <address@hidden>
Date: Friday, April 2, 2004 2:09 pm
Subject: Re: [Freeipmi-devel] Major symbol exporting problem
> ,----
> | > What library is the one that conflicts? Just libiberty or
> another?|
> | libprocps.
> |
> | > In fact I think this is The Right Way (TM)
> |
> | This issue came up due to the random luck that these two
> libraries named
> | two completely different functions identically. And
> unfortunately, they
> | were implemented in such a way that ipmi functions that required
> xmalloc| segfaulted.
> |
> | I think the correct way to do this is to make sure that all
> symbols that
> | we don't want exported are simply not exported. I'm don't know
> what the
> | most portable/correct way to do this though.
> |
> | Al
> `----
> I like the current approach you took. ipmi_ prefix. Simple and easy.
>
> --
> Albert Chu
> address@hidden
> Lawrence Livermore National Laboratory
>
> ----- Original Message -----
> From: Ian Zimmerman <address@hidden>
> Date: Friday, April 2, 2004 9:19 am
> Subject: Re: [Freeipmi-devel] Major symbol exporting problem
>
> > Albert Chu <address@hidden> writes:
> >
> > > Hmmm, off the top of my head, I guessing no. Wrapping the
> #ifdef
> > around> xmalloc will not stop the symbol from being exported.
> But
> > its worth a
> > > shot, I'll give it a shot later on today.
> > >
> > > There are ways to stop symbols from being exported via the
> > compiler.
> > > But I don't know if we want to do that ...
> > >
> >
> > What library is the one that conflicts? Just libiberty or another?
> >
> > We could put a check into configure.ac like
> >
> > LIBS="$LIBS -liberty"
> > AC_CHECK_FUNCS(xmalloc)
> >
> > this will define LIBOBJS=xmalloc.o if it is not found, then you can
> > include it conditionally like
> >
> > modules=foo.o bar.o $(LIBOBJS)
> >
> > In fact I think this is The Right Way (TM)
> >
>
>
> --
> _.|_
> (_||_)
> Free as in Freedom <www.gnu.org>
>