This is the mail archive of the gcc-patches@gcc.gnu.org mailing list for the GCC project.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

Re: 4.0.2/4.1 PATCH: Enforce PIC code for mips libffi (PR libgcj/21943)


On Fri, 10 Jun 2005, Richard Sandiford wrote:

> >  The underlying tool defaults to non-PIC, but if invoked through the GCC 
> > driver it does not.
> 
> Exactly.  That's what's silly!  Contrary to what you said later
> in your reply, I see gas as a user-level tool in its own right,
> independent of gcc.  Therefore...

 Hmm, I can see your point, but you are a toolchain developer.  I don't 
recall having a reason to call gas manually for anything but debugging the 
toolchain itself.  For me calling gas directly is like calling ld this 
way -- it's possible, but you really need to know what you are doing.  But 
perhaps my use of the toolchain is not typical enough, either.

> > Your reason for doing it differently for IRIX is not immediately clear
> > to me, but if you find it important, I won't argue.
> 
> This is just flamebait.  I'd tried to make clear that I didn't choose

 Sorry about that -- I haven't meant to annoy you.  I just stick to 
technical arguments.  Cross-platform consistency of tools is a technical 
property for me.

> the current IRIX behaviour, have no attachment to it, and certainly
> don't find it important.  I just don't see the Linux behaviour to be
> superior enough to reject Rainer's patch and ask him to change the
> IRIX default instead.  Your original reply raised that as a possibilty,

 Does the current IRIX default works e.g. for gmp?  Or openssl?

> and I was trying to justify why I wasn't doing that, and was accepting
> Rainer's patch more-or-less as-is.

 Well, initially I suggested rejection, but not after your response, which 
is why I left it with no further comments; I just accepted your point to 
use ".abicalls" as obvious.  Perhaps I should have been more explicit.  
Please assume all my following discussion to have been meant to be an 
addition to the patch.

  Maciej


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]