Github user ScrapCodes commented on the pull request:
https://github.com/apache/incubator-spark/pull/585#issuecomment-34850248
Well in this commit itself, I have configured it for 0.9.0-incubating
(which is not correct, as we don't intend version compatibility across 1.0.0
and 0.9 ). It generates a whole lot of errors. Should It be posted here ?
It can be run like this.
`sbt mima-report-binary-issues`
A sample(first few lines) of errors...
```
[info] spark-mllib: found 6 potential binary incompatibilities
[error] * object org.apache.spark.mllib.recommendation.MFDataGenerator
does not have a correspondent in new version
[error] filter with:
ProblemFilters.exclude[MissingClassProblem]("org.apache.spark.mllib.recommendation.MFDataGenerator$")
[error] * class org.apache.spark.mllib.recommendation.MFDataGenerator does
not have a correspondent in new version
[error] filter with:
ProblemFilters.exclude[MissingClassProblem]("org.apache.spark.mllib.recommendation.MFDataGenerator")
[error] * class org.apache.spark.mllib.optimization.SquaredGradient does
not have a correspondent in new version
[error] filter with:
ProblemFilters.exclude[MissingClassProblem]("org.apache.spark.mllib.optimization.SquaredGradient")
[error] * method
gradient()org.apache.spark.mllib.optimization.SquaredGradient in class
org.apache.spark.mllib.regression.LinearRegressionWithSGD has now a different
result type; was: org.apache.spark.mllib.optimization.SquaredGradient, is now:
org.apache.spark.mllib.optimization.LeastSquaresGradient
[error] filter with:
ProblemFilters.exclude[IncompatibleResultTypeProblem]("org.apache.spark.mllib.regression.LinearRegressionWithSGD.gradient")
[error] * method
gradient()org.apache.spark.mllib.optimization.SquaredGradient in class
org.apache.spark.mllib.regression.RidgeRegressionWithSGD has now a different
result type; was: org.apache.spark.mllib.optimization.SquaredGradient, is now:
org.apache.spark.mllib.optimization.LeastSquaresGradient
[error] filter with:
ProblemFilters.exclude[IncompatibleResultTypeProblem]("org.apache.spark.mllib.regression.RidgeRegressionWithSGD.gradient")
[error] * method
gradient()org.apache.spark.mllib.optimization.SquaredGradient in class
org.apache.spark.mllib.regression.LassoWithSGD has now a different result type;
was: org.apache.spark.mllib.optimization.SquaredGradient, is now:
org.apache.spark.mllib.optimization.LeastSquaresGradient
[error] filter with:
ProblemFilters.exclude[IncompatibleResultTypeProblem]("org.apache.spark.mllib.regression.LassoWithSGD.gradient")
[info] spark-streaming: found 0 potential binary incompatibilities
...(elided)
```
