🚀 go-pugleaf

  I have been trying to get gdb-4.4 up on an RS/6000 550, AIX 3.1.6, using
gcc-2.1. (by the way, thanks for the debugging support :) I get a lot of 
errors like this:

/usr/include/linenum.h:48: warning: `LINESZ' redefined
./../include/coff/rs6000.h:112: warning: this is the location of the previous
definition

and it fail with this:

coff-rs6000.c: In function `rs6000coff_archive_p':
coff-rs6000.c:237: incompatible types in assignment
coff-rs6000.c: In function `rs6000coff_core_p':
coff-rs6000.c:361: warning: passing arg 1 of `open' discards `const' from
  pointer target type
coff-rs6000.c:368: warning: comparison between pointer and integer
coff-rs6000.c:369: warning: comparison between pointer and integer
coff-rs6000.c:381: incompatible types in assignment
coff-rs6000.c:384: cannot convert to a pointer type

  So, do I need to run fixincludes to fix up my include files? I seem to hear of
people using gdb-4.4 on the RS/6000, and there are certainly files for it, so
what have I done now?

  Using the native xde on the RS/6000 with gcc-2.1 generated code, I get "funny"
sorts of errors. Quite often I get:

      internal error: assertion failed at line 534 in file frame.c

and when I try using optimization ("-O"), I get additional errors:

      "LC" is not defined

The variable window shows a bunch of "external" variables like:

      LC..1
      LC..12

that I assume are due to gcc-2.1. Is gcc-2.1 using the correct debug format?

  Worse, the value of the variables do not accurately get reported in the
variables window (at least "register int" doesn't). In fact xde doesn't seem to
like gcc-2.1 code at all.

  HHHHHHHeeeeeelllllpppppp..............

    Cheers,

    Tom McConnell
--
 Tom McConnell          |     Internet: tmcconne@sedona.intel.com
 Intel, Corp. C3-21     |     Phone: (602)-554-8229
 5000 W. Chandler Blvd. | The opinions expressed are my own. No one in 
 Chandler, AZ  85226    | their right mind would claim them.