FineEnvs/repo2rlenv-r2e-gym
Repo2RLEnv R2E-Gym / SWEGEN: Harbor dataset on Hugging Face with 100 tasks. Mine supported bug-fix commits, measure failing-to-passing and regression tests across revisions, and export the buggy starting state and deterministic verifier. Contains 100 Harbor tasks generated with the owned r2e gym…
Tasks
- IndexedSet silently wraps out-of-range negative indices instead of raising IndexError
- join with does not support a strict parameter
- in range returns wrong result for reversed ranges (start end)
- product index raises TypeError when given iterator inputs
- median raises IndexError on empty collections instead of returning NaN
- get real name() returns wrong component for multi-part dotted identifiers
- all equal makes unnecessary extra groupby. next call
- drop while and take while raise TypeError when passed an iterator
- IndexedSet.update() adds iterables as items instead of their elements when called with multiple arguments
- InvalidMethod should subclass AttributeError for compatibility with typing.Protocol and runtime checkable
- pluralize() produces wrong plural for words ending in "x"
- every().hour.at('MM:SS') parses minutes/seconds incorrectly
- @cache decorator's invalidate() raises KeyError when called twice on the same argument
- value chain silently swallows TypeError raised while iterating an argument
- powerset of sets() does not support a baseset parameter
- sep argument ignored in args2sh and args2cmd
- Bits conversion methods don't preserve empty instances
- Title: reraise does not support callable into argument
- chunk raises ZeroDivisionError when size is 0 or negative
- idle seconds() raises TypeError when no jobs are scheduled
- interleave evenly raises IndexError when called with no iterables
- sort together does not support a strict parameter
- The WHERE clause parser in sqlparse/sql.py does not treat the INTO keyword as a clause terminator. As a result, when parsing a query like…
- last() raises incorrect ValueError when reversed is set to None
- iequals returns True when comparing iterables of different lengths if one contains mock.ANY
- gray product and partial product fail when iterators are passed with repeat argument
- IndexedSet slicing returns wrong results after element removal
- to list returns a dict values view instead of a list when given a dict
- ComplementSet <= and = comparisons raise AttributeError due to typo
- DoS protection limits should raise SQLParseError instead of silently returning
- Bug: reversed() on empty numeric range raises IndexError
- isiterable returns wrong result for objects with getitem but no iter
- log durations doesn't expose a patchable timer in funcy.debug
- dissoc does not support the factory keyword argument
- Bug: GUIDerator raises ValueError with incorrect bounds message
- @throttle() raises TypeError when called with a timedelta argument
- numeric range slicing with negative step returns wrong result
- decimal relative time raises TypeError when given a timezone-aware datetime
- BarrelList slicing raises ValueError for negative slice indices
- @decorator -made decorators with only keyword args don't support being used without parentheses
- fold crashes when used with multiprocessing.Pool().map due to unpicklable lambda
- AtomicSaver and iter find files crash when given a pathlib.Path argument
- pearson type raises RuntimeError and ZeroDivisionError for valid inputs
- Bits class accepts values that exceed the declared bit length
- Bug: BarrelList.insert raises IndexError for very negative indices instead of inserting at the beginning
- Bits negative indexing returns wrong result
- The TLRUCache.expire() method currently returns None . It should instead return an iterable of (key, value) pairs for all cache entries…
- constrained batches silently accepts non-positive max count values
- polynomial from roots raises RecursionError for large inputs
- ParsedException drops anchor lines from Python 3.11+ tracebacks
- zip object and zip object deep raise ValueError when called with an empty list
- last index of returns -1 when the value only exists at index 0
- mean and mean by raise ZeroDivisionError on empty collections
- combination with replacement index gives wrong results and fails to raise errors when None is in the iterable or element
- partition all silently produces wrong output when a sequence has an incorrect len
- parse qsl ignores the encoding parameter when decoding percent-encoded query strings
- Bug: exactly n raises ValueError for negative n values
- ZeroDivisionError when calling get histogram counts() on data with zero interquartile range
- once per raises TypeError when called with unhashable arguments
- str() on a scheduled functools.partial job raises AttributeError
- TTLCache.expire() returns None instead of expired (key, value) pairs
- get parameters() returns empty list for SQL functions with typed literal arguments
- peekable is not subscriptable for generic type annotations
- split before , split after , and split when yield an empty list when given an empty iterable and maxsplit=0
- autocurry decorator does not preserve function metadata (docstring, name, etc.)
- Bug: Requirement('demo===x,y') raises InvalidSpecifier instead of InvalidRequirement
- sliced() silently accepts negative slice sizes instead of raising ValueError
- invoke() does not raise an error when using restricted keys in the path
- BarrelList slice deletion raises IndexError for out-of-bounds stop and multi-barrel spans
- funcutils.wraps forwards defaulted arguments positionally instead of as keywords
- seekable object does not support subscript (index) access
- Function object has no attribute get window
- The raiser() function in funcy/flow.py does not handle the case where a plain string is passed as the first argument. When a string like…
- iter index raises ValueError for negative start / stop on general iterables
- The @mru cache decorator should emit a DeprecationWarning when called, but currently it does not issue any warning.
- ecoutils.get profile(scrub=True) still performs identifying lookups before masking them
- wraps fails when the wrapped function has keyword-only arguments
- pull at raises IndexError for out-of-range or duplicate indexes
- numeric range equality is wrong for single-element ranges with different steps
- cachetools.cached with lock and info=True acquires lock too many times on cache miss
- Bug: .at() with timezone schedules job one day too late in certain scenarios
- OrderedMultiDict.addlist() silently drops values when passed a one-shot iterator
- MRUCache does not emit a DeprecationWarning on instantiation
- @decorator decorators lose their name when called with arguments
- singularize() corrupts words that end in double 's'
- get real name() returns alias instead of actual table name when alias is present
- is sorted fails with objects that only implement lt
- merge with iteratee is not applied to nested objects
- pprint tree produces incorrect tree formatting
- tail() does not raise ValueError for negative n with sized iterables
- Bug: LowerBound(None, True) and UpperBound(None, True) don't normalize inclusive and sort above their False counterparts
- @cachedmethod does not pass self to the key function
- repr() crashes on a partially constructed Job (before .do() is called)
- TIMESTAMP typed literal not grouped correctly in sqlparse
- memoize ignores key func when the function is passed directly
- Long-running jobs skip periods when they finish past their scheduled time
- Bug: Assigning an immediately-expired value to an existing TLRUCache key silently retains the stale old value
- chunked() raises a confusing error when called with a negative n
- rpartial does not accept keyword arguments
- Empty type name or field name raises IndexError instead of ValueError in namedtuple / namedlist