Rollup patch is attached, incorporating all feedback so far (I
think!).  As usual, you should be able to use interdiff to view the
changes.  Let me know if that doesn't work and I can make git spit it
out.

Unfortunately the tests are randomly failing on me.  I'm not sure
what's going on, but it looks like we're not always waiting long
enough after runto_main, somehow.

The relevant expect code (in gdb.base/skip.exp) is:

if ![runto_main] { fail "skip tests suppressed" }
send_gdb "step\n"
# XXX This gdb_test sometimes fails, due to an apparent race.
gdb_test "bt" "\\s*\\#0\\s+main.*" "step after all ignored"

Here's a diff of gdb.log in a successful run (-) and an unsuccessful run (+).

@@ -112,20 +112,20 @@
 Start it from the beginning? (y or n) y
 Starting program: /home/jlebar/code/gdb/debug/gdb/testsuite/gdb.base/skip

 Breakpoint 2, main () at ../../../src/gdb/testsuite/gdb.base/skip.c:7
 7	  return baz(foo(), bar());
 (gdb) step
 bt
 8	}
-(gdb) bt
+(gdb) FAIL: gdb.base/skip.exp: step after all ignored
+bt
 #0  main () at ../../../src/gdb/testsuite/gdb.base/skip.c:8
-(gdb) PASS: gdb.base/skip.exp: step after all ignored
-skip delete 1
+(gdb) skip delete 1
 (gdb) info skip
 Num     Type           Enb Address            What
 2       function       y   0x00000000004004e4 main at
../../../src/gdb/testsuite/gdb.base/skip.c:5
 3       file           y
../../../src/gdb/testsuite/gdb.base/skip1.c
 4       function       y   0x000000000040052b baz at
../../../src/gdb/testsuite/gdb.base/skip1.c:6
 (gdb) PASS: gdb.base/skip.exp: info skip
 delete breakpoints
 Delete all breakpoints? (y or n) y
@@ -226,17 +226,16 @@

Any ideas what's going on here?

-Justin

On Tue, Oct 25, 2011 at 8:26 PM, Stan Shebs <stan_shebs@mentor.com> wrote:
> On 10/25/2011 01:16 PM, Justin Lebar wrote:
>>
>> In the unlikely case that I actually opened the manual (not because
>> it's no good, but because most people I know who use GDB aren't aware
>> of the manual -- it's not well-linked on the web), I think the fact
>> that there are |skip file| and |skip function| commands would clear
>> things up pretty quickly.
>>
>> There's a cost to a comment like this; it's not useful to most
>> readers, but they have to parse it anyway.  (What the heck does "skip
>> function file" mean?)
>
> Another purpose for having niggling details in the manual is that in the
> absence of a formal specification, the manual is our promise to users of
> what they should expect to work, and how it will work.  So users find
> themselves referring to the manual when they try something that seems
> obvious, and are surprised that it doesn't work as expected.
>
>> But again, I don't really have a stake in this either way.  I just
>> want to get my patch in.  :)
>
> I'd like to see a new rollup with all the feedback incorporated, I've lost
> track of the state... :-)
>
> Stan
>
>