Merge ~liushuyu-011/ubuntu/+source/davmail:ubuntu/devel into ubuntu/+source/davmail:ubuntu/devel

Proposed by Zixing Liu
Status: Rejected
Rejected by: Robie Basak
Proposed branch: ~liushuyu-011/ubuntu/+source/davmail:ubuntu/devel
Merge into: ubuntu/+source/davmail:ubuntu/devel
Diff against target: 11 lines (+1/-1)
1 file modified
debian/tests/binary-starts (+1/-1)
Reviewer Review Type Date Requested Status
Athos Ribeiro (community) Abstain
Graham Inggs (community) Abstain
git-ubuntu import Pending
Review via email: mp+448108@code.launchpad.net

Description of the change

This fixes the autopkgtest failure where nc printed out a "connection successful" message to stderr and failing the test.

To post a comment you must log in.
Revision history for this message
Graham Inggs (ginggs) wrote :

The usual way to do this is to add:

Restrictions: allow-stderr

to debian/tests/control.

review: Needs Fixing
Revision history for this message
Zixing Liu (liushuyu-011) wrote :

> The usual way to do this is to add:
>
> Restrictions: allow-stderr
>
> to debian/tests/control.

However, the script relies on stderr to fail the test when nc fails to connect to the test server.

Is that still warrant an `allow-stderr`?

Revision history for this message
Graham Inggs (ginggs) wrote :

OK, I see what you mean.

I'm not convinced we want to carry this as a delta in Ubuntu, but I'll leave my review neutral in case another sponsor disagrees.

Do we know why this regressed in Ubuntu and not in Debian? I think it's worth forwarding this fix there, to make their test more robust in case they are also affected in future.

review: Abstain
Revision history for this message
Zixing Liu (liushuyu-011) wrote :

I think the reason might be Ubuntu and Debian are using different nc implementations

Revision history for this message
Michael Hudson-Doyle (mwhudson) wrote :

What about this? Either way we should probably see if Debian will take a patch, it seems an unfortunate reason to introduce a delta.

diff --git a/debian/tests/binary-starts b/debian/tests/binary-starts
index 5aabca3..e469b7e 100644
--- a/debian/tests/binary-starts
+++ b/debian/tests/binary-starts
@@ -8,5 +8,5 @@ davmail-server /tmp/davmail.properties &

 pid=$!
 sleep 10
-nc -z localhost 1025 || echo "Daemon does not seem to listen" >&2
+nc.traditional -z localhost 1025 || echo "Daemon does not seem to listen" >&2
 kill $pid
diff --git a/debian/tests/control b/debian/tests/control
index a546e70..b352ebc 100644
--- a/debian/tests/control
+++ b/debian/tests/control
@@ -1,2 +1,2 @@
 Tests: binary-starts
-Depends: davmail-server, netcat
+Depends: davmail-server, netcat-traditional

Revision history for this message
Graham Inggs (ginggs) wrote :

I filed https://bugs.debian.org/1043143 about this.

I think it should be policy for autopkgtests not to depend on virtual packages (similar to how real packages must be listed before virtual alternatives in Build-Depends).

I'll try to get this policy accepted in Debian.

Revision history for this message
Vladimir Petko (vpa1977) wrote :

The bug is fixed in Debian and openjdk-* migrated =)
Probably the MP can be closed.

Revision history for this message
Athos Ribeiro (athos-ribeiro) wrote :

I am taking the sponsors slot here to clear up the sponsoring queue at http://reports.qa.ubuntu.com/reports/sponsoring/general.html since this seems to be fixed now. Unfortunately, I cannot properly set the status of this MP.

Revision history for this message
Athos Ribeiro (athos-ribeiro) :
review: Abstain

Unmerged commits

51ab18c... by Zixing Liu

d/tests/binary-starts: redirect nc's success message to stdout

Preview Diff

[H/L] Next/Prev Comment, [J/K] Next/Prev File, [N/P] Next/Prev Hunk
1diff --git a/debian/tests/binary-starts b/debian/tests/binary-starts
2index 5aabca3..65f6770 100644
3--- a/debian/tests/binary-starts
4+++ b/debian/tests/binary-starts
5@@ -8,5 +8,5 @@ davmail-server /tmp/davmail.properties &
6
7 pid=$!
8 sleep 10
9-nc -z localhost 1025 || echo "Daemon does not seem to listen" >&2
10+nc -z localhost 1025 2>&1 || echo "Daemon does not seem to listen" >&2
11 kill $pid

Subscribers

People subscribed via source and target branches

to all changes: