Closed wtwhite closed 4 months ago
Resolves #96 insofar as it brings the build failure count down from 601 to 445, though not all the way to the previous runs (run 34)'s 406.
Comes at the cost of making java-compilers.json "uneven" for ecj compilations in 2 ways:
java-compilers.json
ecj
ecj-3.11.1.v201509021
ecj-3.12.3
ecj-3.15.1
openjdk-11.0.19
-Dmaven.compiler.source=$JDK_MAJOR_VERSION
Will rectify both in a future, fresh run -- right now there's no time.
Added an issue to the dataset repo for posterity: https://github.com/binaryeq/dataset/issues/27
Resolves #96 insofar as it brings the build failure count down from 601 to 445, though not all the way to the previous runs (run 34)'s 406.
Comes at the cost of making
java-compilers.json
"uneven" forecj
compilations in 2 ways:ecj-3.11.1.v201509021
,ecj-3.12.3
andecj-3.15.1
are all (still) built using the originalopenjdk-11.0.19
docker imageecj-3.11.1.v201509021
,ecj-3.12.3
lack-Dmaven.compiler.source=$JDK_MAJOR_VERSION
, since its absence didn't cause many build failuresWill rectify both in a future, fresh run -- right now there's no time.