bug in bfd/dwarf2.c recently

[Nick Clifton]

Hi Marty,

> I used the October and 12/04 cvs snapshots and came across this problem:

> This seemed to fix the coredumping issue:

> +              && each_func->name 
>               && addr >= arange->low
>               && addr < arange->high
>               && strcmp (name, each_func->name) == 0

Thanks - I have checked this patch in together with this ChangeLog entry:

bfd/ChangeLog
2005-12-27  Marty Leisner  <leisner@rochester.rr.com>

	* dwarf2.c (lookup_symbol_in_function_table): Check for a function
	name before passing it to strcmp.


If you have the time, we would very much appreciate a test case that can 
reproduce this bug, so that we can add an entry to the linker's testsuite.

Cheers
   Nick