[SPARK-8725][PROJECT-INFRA] Test modules in topologically-sorted order in dev/run-tests#10885
[SPARK-8725][PROJECT-INFRA] Test modules in topologically-sorted order in dev/run-tests#10885JoshRosen wants to merge 7 commits intoapache:masterfrom
Conversation
| @@ -0,0 +1,85 @@ | |||
| ####################################################################### | |||
| # Implements a topological sort algorithm. | |||
There was a problem hiding this comment.
Oops, I missed this. It looks like it's Apache licensed and you have correctly preserved the header. https://bitbucket.org/ericvsmith/toposort/src/25b5894c4229cb888f77cf0c077c05e2464446ac/LICENSE.txt?fileviewer=file-view-default
You'll need to put the NOTICE contents in our NOTICE:
https://bitbucket.org/ericvsmith/toposort/src/25b5894c4229cb888f77cf0c077c05e2464446ac/NOTICE?fileviewer=file-view-default
That should be all.
There was a problem hiding this comment.
Do I just append the entire contents of the notice verbatim? Is there a shorthand way of doing this (e.g. just including just the copyright)? I got a bit confused looking at the existing NOTICE file and since it hasn't been updated in a while I didn't spot any smaller patches / diffs to model my change on.
There was a problem hiding this comment.
Yes just append. I don't know if we've got a strong format going, except to try to precede the text from Foo with a line like "Foo" or "For Foo:". Anything sane should be OK.
Technically you reproduce all relevant parts of the NOTICE and only relevant parts. All is relevant here. Really, not sure the project needed this file but hey. Practically you could argue you can abbreviate it, but given it's short, I'd just copy it.
There was a problem hiding this comment.
Alright, updated the NOTICE to append it verbatim.
|
Test build #49943 has finished for PR 10885 at commit
|
|
Test build #49944 has finished for PR 10885 at commit
|
|
Whoops: |
|
Test build #49955 has finished for PR 10885 at commit
|
|
Test build #49956 has finished for PR 10885 at commit
|
|
Unless anyone has objections, I'm planning to merge this today. If anyone complains about the loss of Python 2.6 support in this script, they can submit a PR to fix compatibility themselves (again, this is only relevant to developers of Spark / maintainers of its test infrastructure, not end users, and I expect that they're using non-ancient Python versions or can easily upgrade). |
|
Test build #50113 has finished for PR 10885 at commit
|
|
Going to merge this now; will address any problems in followups. |
This patch improves our
dev/run-testsscript to test modules in a topologically-sorted order based on modules' dependencies. This will help to ensure that bugs in upstream projects are not misattributed to downstream projects because those projects' tests were the first ones to exhibit the failureTopological sorting is also useful for shortening the feedback loop when testing pull requests: if I make a change in SQL then the SQL tests should run before MLlib, not after.
In addition, this patch also updates our test module definitions to split
sqlintocatalyst,sql, andhivein order to allow more tests to be skipped when changing onlyhive/files.