This is the mail archive of the gcc@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]

Re: glibc2.1 [offtopic]




  In message <m10EzM4-000390C@ocean.lucon.org>you write:
  > > 
  > >   > It looks to me that "#prama weak" is much more consistent than
  > >   > attribute in egcs 1.1.2. Those gthr-*h files are not old codes. They
  > >   > were added after we started egcs. I just copied what we had. Why now
  > >   > suddenly do you want something different? I hate to see we do things
  > >   > one way in one place and another way in a different place.
  > > No, it means that the gthr files are wrong and need to be fixed.
  > > 
  > 
  > Thanks for clearing it up. I am gald I asked. Now, where should I
  > define ATTRIBUTE_WEAK? Since it will be used in more than once
  > place, it should be in somethere more general.
For the mainline it belongs in gansidecl.h with the rest of the attribute
definitions.

  Note that these attribute defines are not target specific defines and
  defining them in gansidecl.h is clean and the right thing to do.  If you
  look at the mainline you'll see that tm.h already includes gansidecl.h.

However, egcs-1.1 pre-dates the conversion to using gansidecl in a more
consistent manner.  So I would recommend that you place the define for
ATTRIBUTE_WEAK within crtstuff.c for the egcs-1.1 branch.

We can/should deal with fixing gthr-* as an independent patch.

jeff


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