This is the mail archive of the mailing list for the Archer 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]

[delayed-symfile] [commit] Fix a testcase regression due to: (nodebugging symbols found)

commit 25f4478edb6f332ecfb34672f62537b3cd7d97c2

FSF GDB code prints `(no debugging symbols found)' when at none of the loaded
objfiles has any debug info.

Former archer-tromey-delayed-symfile branch did print `(no debugging symbols
found)' on each objfile that has no debug info.  At least on
(modified)kernel- with no kernel-debuginfo
installed its vDSO is missing the debug info.  Thus gdb.base/annota1.exp
regressed on it as it does not expect `(no debugging symbols found)'.

This change is disputable as patched Archer is not completely compatible with
former FSF GDB behavior.
* It could be made compatible but it would be more compatible code.
* I find the behavior with this patch more meaningful than the FSF GDB one.
* It has no regression for the current FSF GDB testsuite.

	* symfile.c (symbol_file_add_with_addrs_or_offsets): Permit the

	optional printing of `(no debugging symbols found)' only for MAINLINE.
 gdb/symfile.c |    4 +++-
 1 files changed, 3 insertions(+), 1 deletions(-)

diff --git a/gdb/symfile.c b/gdb/symfile.c
index f86c216..cd34a7b 100644
--- a/gdb/symfile.c
+++ b/gdb/symfile.c
@@ -1064,7 +1064,9 @@ symbol_file_add_with_addrs_or_offsets (bfd *abfd, int from_tty,
       xfree (debugfile);
-  if (print_symbol_loading && !has_any_debug_symbols (objfile))
+  /* has_any_debug_symbols is not fully compatible with the former calls which
+     would just be needlessly expensive here.  */
+  if (print_symbol_loading && !has_any_debug_symbols (objfile) && mainline)
       wrap_here ("");
       printf_unfiltered (_("(no debugging symbols found)"));

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