Closed tdurieux closed 6 years ago
Arja patches: yyxhdy/defects4j-patches
Capgen patches: https://github.com/justinwm/CapGen/tree/master/Patches (Context-Aware Patch Generation for Better Automated Program Repair)
S3 patches: https://github.com/anonymousfserepair2017/FSE2017-S3-SyntaxSemanticRepairData (S3: Syntax- and Semantic-Guided Repair Synthesis via Programming by Examples)
HDRepair: hdrepair.zip
previous dead link: https://drive.google.com/folderview?id=0B8X_gXjMuN4xeTV2UmtCRU9rdk0&usp=sharing Link is dead, just asked to the authors by email
ssfix patches: https://github.com/qixin5/ssFix/blob/master/expt0/patch/
Sketchfix patches: https://github.com/anonymoususer18/sketchFix_evaluation/
SoFix patches: SoFix_Patch_Data.zip (got them by email)
(Mining StackOverflow for Program Repair)
For Elixir patches, the link seems only contains patches for bug.jar instead of defects4J. Anyone has defects4J patches?
Elixir-Generated-Patch-Defect4J.XLSX Exilir for defects4j
For S3 patches: https://github.com/anonymousfserepair2017/FSE2017-S3-SyntaxSemanticRepairData (S3: Syntax- and Semantic-Guided Repair Synthesis via Programming by Examples)
From their paper, they did not evaluate their approach on defects4J, they evaluated on IntroClassJava and 100 real-world bugs (Closure OrientDB Math Molgenis Heritrix), maybe some of them were collected from defects4J, but not the defects4J benchmark. So should we ignore the patches from S3?
I think we should not consider them
Hi @martinezmatias In this issue, we collect all the correct patches (only the correct patches claimed in the paper, not plausible patches). As patches generated by different tools are in different format, so we need to unify them in a unified format (Interpret the patch file as a unified diff).
I have collected all correct patches from these tools: ACS, CapGen, Elixir, Jaid, SketchFix, ssFix. you can take a look in this file : https://github.com/kth-tcs/defects4-repair-reloaded/blob/master/main-table.tsv All patches are available in folder patches. Just to make sure we have a unified folder structure and name structure for all patches, so we can run our script _patched_programs.sh_ to generate corresponding patches program based on our buggy defects4j programs.
For Dynamoth, I have some files not compatible with our defects4j buggy programs, but I will submit them soon and record them in the main-table.tsv.
For the rest patches from the tools: Nopol, Cardumen, NPEFix, jGenProg 2017, jGebProg-Tibra-mode 2017, HDRepair and SoFix we can do it together. Would you prefre to collect the patches from Cardumen, jGenProg 2017, jGebProg-Tibra-mode 2017 ?
Hi @SophieHYe
Would you prefre to collect the patches from Cardumen, jGenProg 2017, jGebProg-Tibra-mode 2017 ?
Perfect. I will collect, format and push those patches.
Patches of "Using a Probabilistic Model to Predict Bug Fixes" (SANER 18)
HDrepair github: https://github.com/xuanbachle/data-bugfixes
SimFix patches (ISSTA'18): https://github.com/xgdsmileboy/SimFix/
Elixir
Elixir-Generated-Patch-BugsDotJar.xlsx
Dynamoth
https://github.com/tdurieux/dynamoth-experiments/tree/master/results/2017-november
Nopol
https://github.com/Spirals-Team/defects4j-repair/tree/master/results/2017-march
Cardumen
https://github.com/SpoonLabs/astor-experiments/tree/master/cardumen-patches
NPEFix
https://github.com/Spirals-Team/itzal-experiments/tree/master/results/generation
ACS
https://github.com/Adobee/ACS/tree/master/patch/ACS
jGenProg 2017
https://github.com/SpoonLabs/astor-experiments/tree/master/jgenprog-statement-patches-2017/
jGebProg-Tibra-mode 2017
https://github.com/SpoonLabs/astor-experiments/tree/master/tibra-patches-2017