This is the mail archive of the 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: egcs 1.1b installation successful (except for some linker probs)

  In message <9809161019.AA11758@odin>you write:

  > When compiling egcs I used the sun linker again - for no particular 
  > reason (in fact, I didn't really think about it). When I recompiled the 
  > project I am working on, the compilation process was successful, 
  > but the (GNU) linker reported "Bad value" in the egcs libgcc.a file. I 
  > switched back to the sun linker for my project - I did a make, and I 
  > found some undefined symbols - these arose from the gcc 2.8.1 
  > object files - when I did a completely clean make and using the sun 
  > linker, everything worked fine. I did a completely clean make 
  > again, but this time I used the Gnu linker, and just got the same error 
  > as last time.
This may (or may not) be a bug in the gnu linker.  There is one known
case where code found in libgcc.a will tickle a bug in the gnu linker.

It it believed that the problem can only occur if you use the Sun
assembler.   If you use gas it is supposed to work correctly.

Recent snapshots of the gnu linker should not have ths bug.

  > obvious to me from the egcs website - how does the development 
  > of egcs relate to the development of gcc?
We develop at our own pace separately from gcc2.  On occasion we merge
in changes from the gcc2 development tree into our own.

  > Have the gcc development efforts been diverted to egcs,
Most gcc developers are working with egcs.  There are still a small
number of people working on gcc2 development though.


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