Skip to contentSkip to conversation

btrfs-progs: misc-test/034: Avoid debug log populating stdout

PATCH In review Qu Wenruo 3 participants 4 messages Jul 24, 2019 Btrfs file system

lore
Share & tools
lore

Qu Wenruo

author
#0 lore raw

PATCH 1/1 btrfs-progs: misc-test/034: Avoid debug log populating stdout

When running misc-test/034, we got unexpected log output:

      [TEST/misc]   033-filename-length-limit
      [TEST/misc]   034-metadata-uuid

Checking btrfstune logic
Checking dump-super output
Checking output after fsid change
Checking for incompat textual representation
Checking setting fsid back to original
Testing btrfs-image restore

This is caused by commit 2570cff076b1 ("btrfs-progs: test: cleanup misc-tests/034") which uses _log facility which also populates stdout.

Revert to echo "$*" >> "$RESULTS" to remove the noise.

---
 tests/misc-tests/034-metadata-uuid/test.sh | 12 ++++++------
 1 file changed, 6 insertions(+), 6 deletions(-)
tests/misc-tests/034-metadata-uuid/test.sh +6 -6
index 3ef110cda823..3a6771a85a61 100755
@@ -31,7 +31,7 @@ read_metadata_uuid() {
 check_btrfstune() {
 	local fsid
 
-	_log "Checking btrfstune logic"
+	echo "Checking btrfstune logic" >> "$RESULTS"
 	# test with random uuid
 	run_check $SUDO_HELPER "$TOP/btrfstune" -m "$TEST_DEV"
 
@@ -66,7 +66,7 @@ check_dump_super_output() {
 	local dev_item_match
 	local old_metadata_uuid
 
-	_log "Checking dump-super output"
+	echo "Checking dump-super output" >> "$RESULTS"
 	# assert that metadata/fsid match on non-changed fs
 	fsid=$(read_fsid "$TEST_DEV")
 	metadata_uuid=$(read_metadata_uuid "$TEST_DEV")
@@ -78,7 +78,7 @@ check_dump_super_output() {
 	[ $dev_item_match = "[match]" ] || _fail "dev_item.fsid doesn't match on non-metadata uuid fs"
 
 
-	_log "Checking output after fsid change"
+	echo "Checking output after fsid change" >> "$RESULTS"
 	# change metadatauuid and ensure everything in the output is still correct
 	old_metadata_uuid=$metadata_uuid
 	run_check $SUDO_HELPER "$TOP/btrfstune" -M d88c8333-a652-4476-b225-2e9284eb59f1 "$TEST_DEV"
@@ -91,13 +91,13 @@ check_dump_super_output() {
 	[ "$fsid" = "d88c8333-a652-4476-b225-2e9284eb59f1" ] || _fail "btrfstune metadata_uuid change failed"
 	[ "$old_metadata_uuid" = "$metadata_uuid" ] || _fail "metadata_uuid changed unexpectedly"
 
-	_log "Checking for incompat textual representation"
+	echo "Checking for incompat textual representation" >> "$RESULTS"
 	# check for textual output of the new incompat feature
 	run_check_stdout $SUDO_HELPER "$TOP/btrfs" inspect-internal dump-super \
 		"$TEST_DEV" | grep -q METADATA_UUID
 	[ $? -eq 0 ] || _fail "Didn't find textual representation of METADATA_UUID feature"
 
-	_log "Checking setting fsid back to original"
+	echo "Checking setting fsid back to original" >> "$RESULTS"
 	# ensure that  setting the fsid back to the original works
 	run_check $SUDO_HELPER "$TOP/btrfstune" -M "$old_metadata_uuid" "$TEST_DEV"
 
@@ -116,7 +116,7 @@ check_image_restore() {
 	local fsid_restored
 	local metadata_uuid_restored
 
-	_log "Testing btrfs-image restore"
+	echo "Testing btrfs-image restore" >> "$RESULTS"
 	run_check_mkfs_test_dev
 	run_check $SUDO_HELPER "$TOP/btrfstune" -m "$TEST_DEV"
 	fsid=$(read_fsid "$TEST_DEV")
Signature · 1 line
2.22.0
#1 lore raw
Quoted 10 lines · Qu Wenruo

When running misc-test/034, we got unexpected log output:

      [TEST/misc]   033-filename-length-limit
      [TEST/misc]   034-metadata-uuid

Checking btrfstune logic
Checking dump-super output
Checking output after fsid change
Checking for incompat textual representation
Checking setting fsid back to original
Testing btrfs-image restore

Quoted 4 lines

This is caused by commit 2570cff076b1 ("btrfs-progs: test: cleanup misc-tests/034") which uses _log facility which also populates stdout.

Revert to echo "$*" >> "$RESULTS" to remove the noise.

Yes to remove the noise but the idea was to avoid manual redirections to $RESULTS, so we need a new helper or adjust _log.

#2 lore raw
Quoted 11 lines · David Sterba
Quoted 10 lines · Qu Wenruo

When running misc-test/034, we got unexpected log output:

      [TEST/misc]   033-filename-length-limit
      [TEST/misc]   034-metadata-uuid

Checking btrfstune logic
Checking dump-super output
Checking output after fsid change
Checking for incompat textual representation
Checking setting fsid back to original
Testing btrfs-image restore

Quoted 8 lines
Quoted 4 lines

This is caused by commit 2570cff076b1 ("btrfs-progs: test: cleanup misc-tests/034") which uses _log facility which also populates stdout.

Revert to echo "$*" >> "$RESULTS" to remove the noise.

Yes to remove the noise but the idea was to avoid manual redirections to $RESULTS, so we need a new helper or adjust _log.

OK, I'll change _log to avoid populating stdout.

BTW, the reason I didn't change _log is at commit e547fdb16667 ("btrfs-progs: tests, add more common helpers") it's using _log in _run_mayfail(), but now it's not.

So some commits are doing conflicting modification during that cleanup.

Thanks,
Qu

#3 lore raw
Quoted 17 lines · Qu Wenruo
Quoted 14 lines · David Sterba
Quoted 13 lines · Qu Wenruo

When running misc-test/034, we got unexpected log output:

      [TEST/misc]   033-filename-length-limit
      [TEST/misc]   034-metadata-uuid

Checking btrfstune logic
Checking dump-super output
Checking output after fsid change
Checking for incompat textual representation
Checking setting fsid back to original
Testing btrfs-image restore

This is caused by commit 2570cff076b1 ("btrfs-progs: test: cleanup misc-tests/034") which uses _log facility which also populates stdout.

Quoted 7 lines
Quoted 5 lines
Quoted 1 line

Revert to echo "$*" >> "$RESULTS" to remove the noise.

Yes to remove the noise but the idea was to avoid manual redirections to $RESULTS, so we need a new helper or adjust _log.

OK, I'll change _log to avoid populating stdout.

There are more calls to 'echo ... $RESULTS' in several tests, so that would be good to unify.

Quoted 3 lines

BTW, the reason I didn't change _log is at commit e547fdb16667 ("btrfs-progs: tests, add more common helpers") it's using _log in _run_mayfail(), but now it's not.

I see, the code has evolved. The helpers like run_mayfail are allowed to use the raw calls (or can use own internal helpers if that makes sense), and we can make _log exclusively used by tests, once it does print to stdout.

Quoted 1 line

So some commits are doing conflicting modification during that cleanup.

Yeah.