This is the mail archive of the gdb-patches@sourceware.org mailing list for the GDB 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]

Re: [PATCH] Bug 10645: small additions to existing testcase


On Tue, 03 Aug 2010 10:11:09 +0200, Dragos Tatulea wrote:
> ---
>  gdb/testsuite/gdb.base/watchpoint-hw-unreadable.c  |   45 +++++++
>  .../gdb.base/watchpoint-hw-unreadable.exp          |  127 ++++++++++++++++++++
>  2 files changed, 172 insertions(+), 0 deletions(-)
>  create mode 100644 gdb/testsuite/gdb.base/watchpoint-hw-unreadable.c
>  create mode 100644 gdb/testsuite/gdb.base/watchpoint-hw-unreadable.exp
> 

A GNU ChangeLog entry is missing.


> diff --git a/gdb/testsuite/gdb.base/watchpoint-hw-unreadable.c b/gdb/testsuite/gdb.base/watchpoint-hw-unreadable.c
> new file mode 100644
> index 0000000..887c293
> --- /dev/null
> +++ b/gdb/testsuite/gdb.base/watchpoint-hw-unreadable.c
> @@ -0,0 +1,45 @@
> +/* This testcase is part of GDB, the GNU debugger.
> +
> +   Copyright 2009 Free Software Foundation, Inc.
                ^^^^ The date should include also 2010.


[...]
> diff --git a/gdb/testsuite/gdb.base/watchpoint-hw-unreadable.exp b/gdb/testsuite/gdb.base/watchpoint-hw-unreadable.exp
> new file mode 100644
> index 0000000..034972f
> --- /dev/null
> +++ b/gdb/testsuite/gdb.base/watchpoint-hw-unreadable.exp
> @@ -0,0 +1,127 @@
> +# Copyright 2009 Free Software Foundation, Inc.
               ^^^^ The date should include also 2010.


[...]
> +# Tests watchpoints.
> +# Input: Array of list containing the folowing elements:
> +# 1: W_NAME - watched var name
> +# 2: W_ID - watched var id
> +# 3: OLD_VAL - expected old val
> +# 4: NEW_VAL - expected new val
> +proc gdb_watch_test {arg_array} {
> +    upvar $arg_array arg_arr
> +    global test
> +    global gdb_prompt
> +
> +    set expect_output "\r\ninfrun: stopped by watchpoint\r\n.*"
> +    foreach {k v} [array get arg_arr] {
> +        set w_name [lindex $v 0]
> +        set w_id [lindex $v 1]
> +        set old_val [lindex $v 2]
> +        set new_val [lindex $v 3]
> +        
> +        set expect_output "$expect_output\r\nHardware watchpoint $w_id: \\$w_name\[\r\n\]+Old value = $old_val\r\nNew value = $new_val"
> +    }
> +
> +    gdb_test_multiple "continue" $test {
> +        -re "$expect_output.*$gdb_prompt $" {
> +            #puts "PASS"
> +            pass $test
> +        }
> +        default {
> +            #puts "FAIL"
> +            fail $test
> +            return 0
> +        }

I do not see why you could not use simple gdb_test here instead of
gdb_test_multiple.


[...]
> +    || [target_info exists gdb,no_hardware_watchpoints]} then {
> +    verbose "Skipping watchpoint-hw test."
                         ^^^^^^^^^^^^^ This testfile is named differently.

[...]
> +gdb_test "set debug infrun 1"
   ^^^^^^^^ -> gdb_test_no_output nowadays.

Not an issue for this patch:
gdb_test could sanity check it must be given some expect string.


> +gdb_test "show debug infrun" "Inferior debugging is 1\\."
> +
> +gdb_test "watch *p" "ardware watchpoint 2: \\\*p"
> +gdb_test "watch **pp" "ardware watchpoint 3: \\\*\\\*pp"
> +set test "catch *p"

All the testcases generate:

PASS: gdb.base/watchpoint-hw-unreadable.exp: catch *p
PASS: gdb.base/watchpoint-hw-unreadable.exp: catch *p
PASS: gdb.base/watchpoint-hw-unreadable.exp: catch *p

The testcase names should be unique.


> +set c(0) [list *p 2 9 10]
> +set c(1) [list **pp 3 9 10]
> +gdb_watch_test c
> +unset c

Just such a qustion, maybe to simplify all the upvar, unset, set lines etc.
using just some?

gdb_watch_test {{*p 2 9 10} {**pp 3 9 10}}



Thanks,
Jan


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