This is the mail archive of the
gdb-patches@sourceware.org
mailing list for the GDB project.
Re: [PATCH] Fix MI dprintf-insert not printing when a location is pending.
- From: Antoine Tremblay <antoine dot tremblay at ericsson dot com>
- To: Keith Seitz <keiths at redhat dot com>, <gdb-patches at sourceware dot org>
- Date: Fri, 27 Mar 2015 08:34:43 -0400
- Subject: Re: [PATCH] Fix MI dprintf-insert not printing when a location is pending.
- Authentication-results: sourceware.org; auth=none
- References: <1427388426-26548-1-git-send-email-antoine dot tremblay at ericsson dot com> <551459BF dot 9090406 at redhat dot com>
On 03/26/2015 03:10 PM, Keith Seitz wrote:
On 03/26/2015 09:47 AM, Antoine Tremblay wrote:
gdb/ChangeLog:
PR breakpoints/16465
* breakpoint.c (create_breakpoint): Fix missing extra_string.
This is a real nit, so please don't go making any changes here unless a
maintainer requests it, but this changelog entry doesn't really explain
the change you've made. ["Save `extra_string' for pending breakpoints."
is much more descriptive/helpful.]
Indeed why not :)
@@ -9783,10 +9786,16 @@ create_breakpoint (struct gdbarch *gdbarch,
cond_string = xstrdup (cond_string);
make_cleanup (xfree, cond_string);
}
+ /* Create a private copy of any extra string. */
+ if (extra_string)
We explicitly test against NULL for pointers. [I know there are quite a
few violations of this in this function. All are awaiting an
easy/obvious separate cleanup. :-)]
Yes, it will look weird to just change the one I added though...
That's what I did anyway so that the others can be part of a cleanup patch.
+ {
+ extra_string = xstrdup (extra_string);
+ make_cleanup (xfree, extra_string);
+ }
b->cond_string = cond_string;
+ b->extra_string = extra_string;
b->thread = thread;
}
- b->extra_string = NULL;
b->ignore_count = ignore_count;
b->disposition = tempflag ? disp_del : disp_donttouch;
b->condition_not_parsed = 1;
FWIW, I have pretty much an identical change in my locations API
refactor, where I ran across this problem (and more) during testing.
:) Glad it can remove unrelated stuff from your patch
diff --git a/gdb/testsuite/gdb.mi/mi-dprintf-pending.exp
b/gdb/testsuite/gdb.mi/mi-dprintf-pending.exp
new file mode 100644
index 0000000..6832f1d
--- /dev/null
+++ b/gdb/testsuite/gdb.mi/mi-dprintf-pending.exp
@@ -0,0 +1,77 @@
+if {[skip_shlib_tests]} {
+if [get_compiler_info] {
+if { [gdb_compile_shlib $libsrc1 $lib_sl1 $lib_opts] != "" } {
+if { [gdb_compile $srcdir/$subdir/$srcfile $binfile executable
$exec_opts] != ""} {
This is a /big/ nitpick, but it's something that consistently irritates
me: compare the coding style of the four statements above. While there
is no "rule" governing which is most correct/desired, I always use the
first. I would ask you to choose one style and be consistent, but I am
not asking you to make any changes right now. Just please keep this in
mind in future patches.
I really had not noticed that, thanks for making me notice! I must
confess I do a lot of copy & paste when writing a test. The if {[func]}
seems indeed the one used in all the tcl docs so I'll use that.
+# Set pending dprintf via MI.
+mi_gdb_test "-dprintf-insert -f pendfunc1 \"hello\"" \
+
".*\\^done,bkpt={number=\"1\",type=\"dprintf\",disp=\"keep\",enabled=\"y\",addr=\"<PENDING>\",pending=\"pendfunc1\",times=\"0\",original-location=\"pendfunc1\"}"
\
+ "mi set dprintf"
+
+mi_gdb_test "-break-insert $bp_location1" ".*" "mi insert breakpoint
bp_location1"
Is it possible to use mi_make_breakpoint for these tests?
Unfortunately for the dprintf one mi_make_breakpoint doesn't support
pending breakpoints, it creates something like :
bkpt={number="2",type=".*",disp=".*",enabled=".*",addr=".*",func=".*",
file=".*/myfile.c",fullname=".*",line="3",thread-groups=\[.*\],
times="0".*original-location=".*"}
But with pending funcs it should be pending= ...
It could be the subject of another patch to add that support.
I used mi_create_breakpoint for the other breakpoint now
+
+mi_run_cmd
+
+set msg "mi dprintf"
+gdb_expect {
+ -re ".*~\"hello\"" {
+ pass $msg
+ }
+ -re ".*$mi_gdb_prompt$" {
+ fail $msg
+ }
+ timeout {
+ fail $msg
+ }
+}
This a pretty common test suite idiom, I think. Can mi_gdb_test be used
instead of gdb_expect?
That I can't since mi_gdb_test requires a command and in this case I'm
just doing expect on that comes after mi_run_cmd, there's no command
associated with it..
+mi_expect_stop ".*" ".*" ".*" ".*" ".*" "" "$msg stop"
diff --git a/gdb/testsuite/gdb.mi/mi-dprintf-pendshr.c
b/gdb/testsuite/gdb.mi/mi-dprintf-pendshr.c
new file mode 100644
index 0000000..fe49a8d
--- /dev/null
+++ b/gdb/testsuite/gdb.mi/mi-dprintf-pendshr.c
@@ -0,0 +1,27 @@
+/* This testcase is part of GDB, the GNU debugger.
+
+ Copyright 2015 Free Software Foundation, Inc.
+
+ This program is free software; you can redistribute it and/or modify
+ it under the terms of the GNU General Public License as published by
+ the Free Software Foundation; either version 3 of the License, or
+ (at your option) any later version.
+
+ This program is distributed in the hope that it will be useful,
+ but WITHOUT ANY WARRANTY; without even the implied warranty of
+ MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
+ GNU General Public License for more details.
+
+ You should have received a copy of the GNU General Public License
+ along with this program. If not, see
<http://www.gnu.org/licenses/>. */
+
+void
+pendfunc1 ()
+{
+}
+
+void
+pendfunc ()
+{
+ pendfunc1();
+}
IIRC, we are now requiring test case conformance to the coding standard.
[At least, that's what I've been told in the past.] So, "(void)" in the
function decls and spaces between function names and '('. [I don't think
we're requiring function comments for trivial stuff like this, though.]
Indeed that's really my old habits dying hard... fixed sorry about that.
Thanks for the review ! , Patch v2 is coming up in a minute as a
separate email...
Antoine