Hui Zhu wrote:
Thanks Michael and Jilin,
I think the Michael's patch is more spend up. Jilin's patch is more
clear and can handle set_record_insn_max_num.
I suggest we choice speed up. Do you think it's OK?
If we choice it, Michael, could you please add some comments to
"record_list_release_first" or change it's name to
"record_list_release_first_without_update_xxx" to make it clear.
And please update set_record_insn_max_num.
Hah, sorry, I completely missed the implications for
set_record_insn_max_num. Thanks Jilin for catching it.
New diff:
2009-10-12 Michael Snyder <msnyder@vmware.com>
* record.c (record_list_release_first): Do not decrement
record_insn_num.
(set_insn_num_max): Remove printf.
Decrement record_insn_num in the loop.
Index: record.c
===================================================================
RCS file: /cvs/src/src/gdb/record.c,v
retrieving revision 1.20
diff -u -p -r1.20 record.c
--- record.c 27 Sep 2009 02:49:34 -0000 1.20
+++ record.c 13 Oct 2009 18:04:44 -0000
@@ -177,6 +177,11 @@ record_list_release_next (void)
}
}
+/* Delete the first instruction from the beginning of the log, to make
+ room for adding a new instruction at the end of the log.
+
+ Note -- this function does not modify record_insn_num. */
+
static void
record_list_release_first (void)
{
@@ -209,8 +214,6 @@ record_list_release_first (void)
if (type == record_end)
break;
}
-
- record_insn_num--;
}
/* Add a struct record_entry to record_arch_list. */
@@ -1260,12 +1263,12 @@ set_record_insn_max_num (char *args, int
{
if (record_insn_num > record_insn_max_num && record_insn_max_num)
{
- printf_unfiltered (_("Record instructions number is bigger than "
- "record instructions max number. Auto delete "
- "the first ones?\n"));
-
+ /* Count down record_insn_num while releasing records from list. */
while (record_insn_num > record_insn_max_num)
- record_list_release_first ();
+ {
+ record_list_release_first ();
+ record_insn_num--;
+ }
}
}