[SPARK-4504] Minor bug fixes in bin/run-example

[witgo]

./sbt/sbt clean assembly =>
examples/target/scala-2.10/spark-examples_2-10-1.2.0-SNAPSHOT-hadoop1.0.4.jar

[SparkQA]

Test build #22804 has started for PR 3069 at commit e13fab7.

• This patch merges cleanly.

[srowen]

What is the purpose of this change? both patterns match the actual file name that gets generated, like spark-examples-1.2.0-SNAPSHOT-hadoop2.4.0.jar. The existing pattern looks clearer to me.

[witgo]

mvn package generated file is like this spark-examples-1.2.0-SNAPSHOT-* .
./sbt/sbt clean assembly generated file is like this spark-examples_2-10-1.2.0-SNAPSHOT-*( Here more such strings:_2-10).

spark-examples-*hadoop*.jar only matching spark-examples-1.2.0-SNAPSHOT-* .

[srowen]

Ah I see. But shouldn't the naming just be made consistent between SBT and Maven then?

[SparkQA]

Test build #22804 has finished for PR 3069 at commit e13fab7.

• This patch passes all tests.
• This patch merges cleanly.
• This patch adds no public classes.

[AmplabJenkins]

Test PASSed.
Refer to this link for build results (access rights to CI server needed):
https://amplab.cs.berkeley.edu/jenkins//job/SparkPullRequestBuilder/22804/
Test PASSed.

[witgo]

The current solution is simple to implement, and other source also used this solution. eg:
compute-classpath.cmd#L52

[SparkQA]

Test build #25055 has started for PR 3069 at commit 24905b7.

• This patch merges cleanly.

[SparkQA]

Test build #25055 has finished for PR 3069 at commit 24905b7.

• This patch passes all tests.
• This patch merges cleanly.
• This patch adds no public classes.

[AmplabJenkins]

Test PASSed.
Refer to this link for build results (access rights to CI server needed):
https://amplab.cs.berkeley.edu/jenkins//job/SparkPullRequestBuilder/25055/
Test PASSed.

[srowen]

I think this has been superseded by the discussion in #3377 ?