[Pkg-javascript-devel] Bug#812714: node-redis: FTBFS - tests fail, redis-server not killed if tests fail

[Michael Tautschnig]

Control: severity -1 wishlist

Hi Andreas,

On Fri, Jan 29, 2016 at 12:51:57 +0100, Andreas Beckmann wrote:
[...]
> I could not reproduce this either.
> This looks a bit fishy, mixing different server versions.
> Did you have stale process running before the test?
> 
> $ grep -E 'redis-server|Connected' node-redis-build-log.txt
> dpkg-source: info: applying 0001-launch_redis-server_during_test.patch
>  -> Considering build-dep redis-server
>    -> Trying redis-server
>  -> Installing  debhelper dh-buildinfo nodejs redis-server
>   libxml2 man-db nodejs po-debconf redis-server redis-tools
> Selecting previously unselected package redis-server.
> Preparing to unpack .../redis-server_2%3a3.0.6-2_amd64.deb ...
> Unpacking redis-server (2:3.0.6-2) ...
> Setting up redis-server (2:3.0.6-2) ...
> [redis-server]                 _._
> 
> Connected to 127.0.0.1:6379, Redis server version 2.6.9
> - ipv4:Connected to 127.0.0.1:6379, Redis server version 2.6.9
> - unix_socket:Connected to /tmp/redis.sock, Redis server version 3.0.6
> 

A very good catch indeed. There may have been an old zombie server running,
started by whatever package. I am a bit surprised that the Redis server still
claimed a successful start-up, despite another daemon already listening on port
6379 (and /tmp/redis.sock), but that's not node-redis' fault.

Downgrading to wishlist: Would it be possible to randomize the port number that
the redis-server is listening on for the purpose of these tests?

Feel free to close the bug if you think the above suggestion is overkill.

Best,
Michael

-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 841 bytes
Desc: not available
URL: <http://lists.alioth.debian.org/pipermail/pkg-javascript-devel/attachments/20160218/9383813d/attachment.sig>