Processed: Re: Bug#886117: FTBFS: most of testsuite

[Debian Bug Tracking System]

Processing commands for cont...@bugs.debian.org:

> tags 886117 + unreproducible moreinfo
Bug #886117 [src:duplicity] FTBFS: most of testsuite
Added tag(s) moreinfo and unreproducible.
> severity 886117 normal
Bug #886117 [src:duplicity] FTBFS: most of testsuite
Severity set to 'normal' from 'serious'
> thanks
Stopping processing here.

Please contact me if you need assistance.
-- 
886117: https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=886117
Debian Bug Tracking System
Contact ow...@bugs.debian.org with problems

Bug#886117: FTBFS: most of testsuite

[Alexander Zangerl]

tags 886117 + unreproducible moreinfo
severity 886117 normal
thanks

On Wed, 03 Jan 2018 16:24:57 +0100, Adam Borowski writes:
>Working hypothesis so far is that the testsuite fails on btrfs; I won't be
>able to confirm until the evening.

sorry, but i can't reproduce your problem.

duplicity builds and tests fine on stretch with ext4 and and with btrfs,
kernel 4.9.69; on sid with ext4 and btrfs (same kernel), and on the
build daemons as per 
https://buildd.debian.org/status/package.php?p=duplicity=sid.

i've removed the output-silencing for the tar invocation that fails
in your environment (see the attached updated patch), but all that
changed is that i get a number of 'implausibly old timestamp' warnings;
these are legit as the test files do contain time stamps near 1.1.1970.

regards
az

Author: Michael Terry 
Subject: Disable some tests for being flaky

--- a/testing/functional/test_restart.py
+++ b/testing/functional/test_restart.py
@@ -111,6 +111,7 @@ class RestartTest(FunctionalTestCase):
 # there should be 2 differences found, one missing file, one mtime change
 # self.verify("testfiles/largefiles")
 
+@unittest.skip("Flaky test because it relies on knowing how many volumes the source files will be split into")
 def test_last_file_missing_at_end(self):
 """
 Test restart when the last file being backed up is missing on restart.
--- a/testing/unit/test_gpg.py
+++ b/testing/unit/test_gpg.py
@@ -129,6 +129,7 @@ class GPGTest(UnitTestCase):
 sig = decrypted_file.get_signature()
 assert sig == self.sign_key, sig[-8:]
 
+@unittest.skip("Flaky test because it relies on compressed size of random bytes")
 def test_GPGWriteFile(self):
 """Test GPGWriteFile"""
 size = 400 * 1000
@@ -144,6 +145,7 @@ class GPGTest(UnitTestCase):
  profile, size=size)
 # print os.stat("testfiles/output/gpgwrite.gpg").st_size
 
+@unittest.skip("Flaky test because it relies on compressed size of random bytes")
 def test_GzipWriteFile(self):
 """Test GzipWriteFile"""
 size = 400 * 1000
--- a/testing/unit/test_selection.py
+++ b/testing/unit/test_selection.py
@@ -173,6 +173,7 @@ class MatchingTest(UnitTestCase):
 assert select.glob_get_sf("**", 0)(root) == 0
 assert select.glob_get_sf("/foo/*", 0)(root) is None
 
+@unittest.skip("unreliable ass-U-me wrt / and /usr on one fs")
 def test_other_filesystems(self):
 """Test to see if --exclude-other-filesystems works correctly"""
 root = Path("/")
--- a/testing/unit/test_statistics.py
+++ b/testing/unit/test_statistics.py
@@ -59,12 +59,13 @@ class StatsObjTest(UnitTestCase):
 s1 = StatsDeltaProcess()
 assert s1.get_stat('SourceFiles') == 0
 
+@unittest.skip("TZ setting seems to fail under sbuild, #880251")
 def test_get_stats_string(self):
 """Test conversion of stat object into string"""
 s = StatsObj()
 stats_string = s.get_stats_string()
 assert stats_string == "", stats_string
-
+   
 self.set_obj(s)
 stats_string = s.get_stats_string()
 assert stats_string == """\
--- a/testing/__init__.py
+++ b/testing/__init__.py
@@ -85,7 +85,7 @@ class DuplicityTestCase(unittest.TestCas
 
 def unpack_testfiles(self):
 assert not os.system("rm -rf testfiles")
-assert not os.system("tar xzf testfiles.tar.gz > /dev/null 2>&1")
+assert not os.system("tar xzf testfiles.tar.gz")
 assert not os.system("mkdir testfiles/output testfiles/cache")
 
 def _update_env(self, key, value):
-- 
Alexander Zangerl + GPG Key 2FCCF66BB963BD5F + http://snafu.priv.at/
"I can't see any data coming out of this Tolkien Ring card."
"Well of course not, it confers invisibility." -- Anthony de Boer


signature.asc
Description: Digital Signature

Bug#886117: FTBFS: most of testsuite

[Alexander Zangerl]

On Wed, 03 Jan 2018 16:24:57 +0100, Adam Borowski writes:
>Working hypothesis so far is that the testsuite fails on btrfs; I won't be
>able to confirm until the evening.

i'm just building a kernel with btrfs to verify your very probable hypothesis;

upstream's test suite is a fairly yucktastic setup that ditches a lot of
very useful output for content-free assertions.

what does fail in your tests is the test priming in testing/__init__.py,
where testing/testfiles.tar.gz is unpacked.

it's running "tar xzf testfiles.tar.gz > /dev/null 2>&1" and that
tarball does contain a number of nastily named files (glancing at
hexdump tells me it's iso-8859-1 and not utf-8).
i suspect these file names are the root cause.

regards
az


-- 
Alexander Zangerl + GPG Key 2FCCF66BB963BD5F + http://snafu.priv.at/
"Ein Blitzableiter auf einem Kirchturm ist das denkbar stärkste 
Mißtrauensvotum gegen den lieben Gott." -- Karl Kraus


signature.asc
Description: Digital Signature

Bug#886117: FTBFS: most of testsuite

[Adam Borowski]

On Wed, Jan 03, 2018 at 06:29:06PM +1000, Alexander Zangerl wrote:
> On Tue, 02 Jan 2018 14:40:33 +0100, Adam Borowski writes:
> >I'm afraid that your package fails to build on current unstable, failing the
> >vast majority of tests:
> 
> looks like every one of the failed tests fails for the same reason,
> (which to me indicates rather a failure of the test suite
> than a failure of duplicity itself):
> 
>   File "/<>/testing/__init__.py", line 88, in unpack_testfiles
> assert not os.system("tar xzf testfiles.tar.gz > /dev/null 2>&1")
> AssertionError
> 
> i'll go dig deeper when i find the time, but i've got no clue how
> why that could be a platform-specific failure.

It's not: I tried on armhf (many times) and amd64 (once).

But, trying more:

machine   archresult

armhf armhf   ✗

arm64 arm64   ✗
arm64 armhf   ✗

amd64 amd64   ✗
amd64 i386✗
amd64 x32 ✗
amd64 armhf (qemu-user)   ✗
amd64 arm64 (qemu-user)   ✗
amd64 mips  (qemu-user)   ✗

i386  i386✓

The machine that succeeded differs by:
* (slightly) kernel: 4.15-rc3 vs 4.15-rc6 on the rest
* decade of manufacture (a dingy pre-NX Pentium4)
* host release: stretch vs unstable on the rest
* sbuild backend: dir vs btrfs on the rest
* underlying filesystem: ext4...

Working hypothesis so far is that the testsuite fails on btrfs; I won't be
able to confirm until the evening.


Meow!
-- 
// If you believe in so-called "intellectual property", please immediately
// cease using counterfeit alphabets.  Instead, contact the nearest temple
// of Amon, whose priests will provide you with scribal services for all
// your writing needs, for Reasonable And Non-Discriminatory prices.

Bug#886117: FTBFS: most of testsuite

[Alexander Zangerl]

On Tue, 02 Jan 2018 14:40:33 +0100, Adam Borowski writes:
>I'm afraid that your package fails to build on current unstable, failing the
>vast majority of tests:

looks like every one of the failed tests fails for the same reason,
(which to me indicates rather a failure of the test suite
than a failure of duplicity itself):

  File "/<>/testing/__init__.py", line 88, in unpack_testfiles
assert not os.system("tar xzf testfiles.tar.gz > /dev/null 2>&1")
AssertionError

i'll go dig deeper when i find the time, but i've got no clue how
why that could be a platform-specific failure.

regards
az



-- 
Alexander Zangerl + GPG Key 2FCCF66BB963BD5F + http://snafu.priv.at/
"Unterschätze nie die Macht dummer Leute, die einer Meinung sind."
 -- Kurt Tucholsky
% 
Which is worse: Ignorance or apathy? I don't know and I don't care.


signature.asc
Description: Digital Signature