Mirror of the gdb-patches mailing list
 help / color / mirror / Atom feed
From: Tom de Vries <tdevries@suse.de>
To: Tom Tromey <tom@tromey.com>
Cc: gdb-patches@sourceware.org
Subject: Re: [PATCH 1/3][gdb/testsuite] Fix global array cp_class_table_history leak
Date: Tue, 2 Jun 2020 15:02:13 +0200	[thread overview]
Message-ID: <32ce36a4-1da6-7841-8e3b-fcb82b625e1a@suse.de> (raw)
In-Reply-To: <87tv073f1i.fsf@tromey.com>

[-- Attachment #1: Type: text/plain, Size: 783 bytes --]

On 22-05-2020 22:12, Tom Tromey wrote:
>>>>>> "Tom" == Tom de Vries <tdevries@suse.de> writes:
> 
> Tom> +# This test-case uses the cp_test_ptype_class cache (by means of the ibid
> Tom> +# argument), so initialize and finalize the cache.
> Tom> +init_cp_test_ptype_class_cache
> Tom> +
> Tom>  do_tests
> Tom> +
> Tom> +finish_cp_test_ptype_class_cache
> 
> If the users always have to do this init/test/finish thing, then maybe
> some kind of wrapper proc would be more appropriate.
> 

Done.

> Also, it's more robust to use 'catch' to ensure that the cleanup is
> always run.
> 

Done.

> However, is the finish bit needed?  If the init proc cleans up all the
> state, then it seems unnecessary.

Agreed, strictly speaking it's unnecessary, it's just a cleanup.

Thanks,
- Tom



[-- Attachment #2: 0001-gdb-testsuite-Fix-global-array-cp_class_table_history-leak.patch --]
[-- Type: text/x-patch, Size: 4943 bytes --]

[gdb/testsuite] Fix global array cp_class_table_history leak

Test-cases using proc cp_test_ptype_class set a global array
cp_class_table_history, which remains set after the test-case has run.  This
has the potential to:
- cause tcl errors, as well as
- reuse results from one test-case in another test-case when that is not
  appropriate

Fix this by:
- moving the variable into a namespace, and
- only using the variable in the test-cases that need it, and
- resetting the variable at the end of those test-cases.

gdb/testsuite/ChangeLog:

2020-05-19  Tom de Vries  <tdevries@suse.de>

	PR testsuite/25996
	* lib/cp-support.exp (init_cp_test_ptype_class_cache)
	(finish_cp_test_ptype_class_cache, use_cp_test_ptype_class_cache):
	New proc.
	(cp_test_ptype_class): Only use use_cp_class_table_history if
	init_cp_test_ptype_class_cache was called.
	* gdb.cp/inherit.exp: Call use_cp_test_ptype_class_cache.
	* gdb.cp/virtfunc.exp: Same.

---
 gdb/testsuite/gdb.cp/inherit.exp  |  6 +++-
 gdb/testsuite/gdb.cp/virtfunc.exp |  6 +++-
 gdb/testsuite/lib/cp-support.exp  | 68 +++++++++++++++++++++++++++++++++------
 3 files changed, 68 insertions(+), 12 deletions(-)

diff --git a/gdb/testsuite/gdb.cp/inherit.exp b/gdb/testsuite/gdb.cp/inherit.exp
index 2d4635c96a..67a11533dd 100644
--- a/gdb/testsuite/gdb.cp/inherit.exp
+++ b/gdb/testsuite/gdb.cp/inherit.exp
@@ -719,4 +719,8 @@ proc do_tests { } {
     test_print_mvi_classes
 }
 
-do_tests
+# This test-case uses the cp_test_ptype_class cache (by means of the ibid
+# argument), so mark it as such.
+use_cp_test_ptype_class_cache {
+    do_tests
+}
diff --git a/gdb/testsuite/gdb.cp/virtfunc.exp b/gdb/testsuite/gdb.cp/virtfunc.exp
index 2b046c5759..dc83091692 100644
--- a/gdb/testsuite/gdb.cp/virtfunc.exp
+++ b/gdb/testsuite/gdb.cp/virtfunc.exp
@@ -289,4 +289,8 @@ proc do_tests {} {
     gdb_test "step" ".*E::vg.*" "step through thunk into E::vg"
 }
 
-do_tests
+# This test-case uses the cp_test_ptype_class cache (by means of the ibid
+# argument), so mark it as such.
+use_cp_test_ptype_class_cache {
+    do_tests
+}
diff --git a/gdb/testsuite/lib/cp-support.exp b/gdb/testsuite/lib/cp-support.exp
index b4a5582742..ba97fbbd87 100644
--- a/gdb/testsuite/lib/cp-support.exp
+++ b/gdb/testsuite/lib/cp-support.exp
@@ -255,17 +255,22 @@ proc cp_test_ptype_class { in_exp in_testname in_key in_tag in_class_table
 	set in_command "ptype${in_ptype_arg} $in_exp"
     }
 
-    # Save class tables in a history array for reuse.
-
-    global cp_class_table_history
-    if { $in_class_table == "ibid" } then {
-	if { ! [info exists cp_class_table_history("$in_key,$in_tag") ] } then {
-	    fail "$in_testname // bad ibid"
-	    return false
+    namespace upvar ::cp_support_internal:: use_cp_class_table_history \
+	use_cp_class_table_history
+    if { [info exists use_cp_class_table_history ] } {
+	# Save class tables in a history array for reuse.
+
+	namespace upvar ::cp_support_internal:: cp_class_table_history \
+	    cp_class_table_history
+	if { $in_class_table == "ibid" } then {
+	    if { ! [info exists cp_class_table_history("$in_key,$in_tag") ] } then {
+		fail "$in_testname // bad ibid"
+		return false
+	    }
+	    set in_class_table $cp_class_table_history("$in_key,$in_tag")
+	} else {
+	    set cp_class_table_history("$in_key,$in_tag") $in_class_table
 	}
-	set in_class_table $cp_class_table_history("$in_key,$in_tag")
-    } else {
-	set cp_class_table_history("$in_key,$in_tag") $in_class_table
     }
 
     # Split the class table into separate tables.
@@ -764,3 +769,46 @@ proc cp_test_ptype_class { in_exp in_testname in_key in_tag in_class_table
 
     return true
 }
+
+# Initialize the cp_test_ptype_class cache.
+
+proc init_cp_test_ptype_class_cache { } {
+    namespace upvar ::cp_support_internal:: use_cp_class_table_history \
+	use_cp_class_table_history
+    set use_cp_class_table_history 1
+
+    namespace upvar ::cp_support_internal:: cp_class_table_history \
+	cp_class_table_history
+    if { [info exists cp_class_table_history] } {
+	unset cp_class_table_history
+    }
+}
+
+# Finalize the cp_test_ptype_class cache.
+
+proc finish_cp_test_ptype_class_cache { } {
+    namespace upvar ::cp_support_internal:: use_cp_class_table_history \
+	use_cp_class_table_history
+    unset use_cp_class_table_history
+
+    namespace upvar ::cp_support_internal:: cp_class_table_history \
+	cp_class_table_history
+    if { [info exists cp_class_table_history] } {
+	unset cp_class_table_history
+    }
+}
+
+# Run WRAPPED_BODY using the cp_test_ptype_class_cache.
+
+proc use_cp_test_ptype_class_cache { wrapped_body } {
+    init_cp_test_ptype_class_cache
+    set code [catch {uplevel 1 $wrapped_body} result]
+    finish_cp_test_ptype_class_cache
+
+    if {$code == 1} {
+	global errorInfo errorCode
+	return -code error -errorinfo $errorInfo -errorcode $errorCode $result
+    } elseif {$code > 1} {
+	return -code $code $result
+    }
+}

      reply	other threads:[~2020-06-02 13:02 UTC|newest]

Thread overview: 3+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2020-05-19 16:29 Tom de Vries
2020-05-22 20:12 ` Tom Tromey
2020-06-02 13:02   ` Tom de Vries [this message]

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=32ce36a4-1da6-7841-8e3b-fcb82b625e1a@suse.de \
    --to=tdevries@suse.de \
    --cc=gdb-patches@sourceware.org \
    --cc=tom@tromey.com \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox