feat(eval): one tagged 88-task dataset with runtime tasks, plus bashkit_generate eval by chaliy · Pull Request #2664 · everruns/bashkit

refactor(eval): merge bash, repo and hard tasks into one tagged dataset

chaliy

committed

4705ea2

feat(eval): add bashkit_generate one-shot script generation eval

chaliy

committed

4fad91b

fix(sqlite): persist databases in rollback-journal mode

chaliy

committed

7021b1a

fix(interpreter): run a script's shebang interpreter when it is a builtin

chaliy

committed

be96673

feat(eval): add mode=runtime tasks with python3 and sqlite3

chaliy

committed

0a9d0f8

chore(eval): bashkit_generate results for five models (OpenAI pending)

chaliy

committed

650d781

chore(eval): bashkit_bash tagged run, five models (partial: provider credit)

chaliy

committed

1af412b

chore(eval): results for tagged 88-task and generate evals, site lineup by task count

chaliy

committed

27ab0c7

test(eval): generate rt_py_large_log input with awk, not python3

chaliy

committed

4f018fc

test(eval): widen task timeouts under coverage instrumentation

chaliy

committed

ac458aa

ci(coverage): raise tarpaulin per-binary timeout to 900s

chaliy

committed

e72bc40

Revert "ci(coverage): raise tarpaulin per-binary timeout to 900s"

chaliy

committed

68b3759

Revert "test(eval): widen task timeouts under coverage instrumentation"

chaliy

committed

b343dcb

ci(coverage): exclude bashkit-eval from the tarpaulin workspace

chaliy

committed

40956d8

Navigation Menu

Repository navigation

feat(eval): one tagged 88-task dataset with runtime tasks, plus bashkit_generate eval - #2664

feat(eval): one tagged 88-task dataset with runtime tasks, plus bashkit_generate eval#2664
chaliy merged 14 commits into
mainfrom
claude/project-thread-hndwnc

Commits on Oct 9, 2026

refactor(eval): merge bash, repo and hard tasks into one tagged dataset

feat(eval): add bashkit_generate one-shot script generation eval

fix(sqlite): persist databases in rollback-journal mode

fix(interpreter): run a script's shebang interpreter when it is a builtin

feat(eval): add mode=runtime tasks with python3 and sqlite3

chore(eval): bashkit_generate results for five models (OpenAI pending)

chore(eval): bashkit_bash tagged run, five models (partial: provider credit)

chore(eval): results for tagged 88-task and generate evals, site lineup by task count

test(eval): generate rt_py_large_log input with awk, not python3

test(eval): widen task timeouts under coverage instrumentation

ci(coverage): raise tarpaulin per-binary timeout to 900s

Revert "ci(coverage): raise tarpaulin per-binary timeout to 900s"

Revert "test(eval): widen task timeouts under coverage instrumentation"

ci(coverage): exclude bashkit-eval from the tarpaulin workspace