cassandra-commits mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Jim Witschey (JIRA)" <>
Subject [jira] [Commented] (CASSANDRA-10289) Fix cqlshlib tests
Date Tue, 29 Sep 2015 18:51:07 GMT


Jim Witschey commented on CASSANDRA-10289:

[~Stefania] Thank you, this is exactly the kind of feedback I was hoping for. Yes, the prompt
test was the failing one, and it's been fixed for CASSANDRA-10369; my mistake for not linking

bq. These tests are quite fast to run, I wonder if we should run them on every developer branch,
perhaps by adding them to the dtest suite. 

I agree that running them on every dev branch is a good idea. However, I'm not sure they belong
in dtests:

- One reason they're so fast is because they reuse the same cluster for all the tests. This
is possible in dtest but (I don't think) part it's of how we currently run them on CassCI.
I don't know if it would be a problem to run the dtests this way on CassCI.
- They also require access to cqlshlib internals, like the parser. I think this is probably
out of scope for the dtests.

We're talking in TE about the best way to add this to the tests that run when developers push

bq. We tend to test cqlsh via the dtests but some use cases would be easier to add to these
tests, e.g. completion.

I agree, and in general, there's a lot (for me in particular) to add to these tests.

I'll revert the changes you mentioned and file new tickets for them.

> Fix cqlshlib tests
> ------------------
>                 Key: CASSANDRA-10289
>                 URL:
>             Project: Cassandra
>          Issue Type: Bug
>          Components: Tests
>            Reporter: Jim Witschey
>            Assignee: Jim Witschey
>              Labels: cqlsh
>             Fix For: 3.0.0 rc2
>         Attachments: trunk-10289.txt
> The cqlsh tests in trunk haven't been running for a while:
> This looks like the driver errors that happened because of CASSANDRA-6717. Not sure why
it's happening now; the driver installation looks normal to me on those jobs. [~mshuler]?
> There were also some changes to cqlsh itself that also broke the test harness, but I
believe those are fixed here:
> Once the tests are running successfully on CassCI, I'll test my patch and mark as patch

This message was sent by Atlassian JIRA

View raw message