Reported Bugs

A list of 323 bugs and issues reported across 11 open-source applications spanning deep learning frameworks (PyTorch, TensorFlow, JAX), LLM serving engines (vLLM, SGLang), databases (ArangoDB, PostgreSQL), compilers (Clang/LLVM, GCC), and system utilities (cronie, OpenSSH). 73.1% have been confirmed or fixed upstream.

Last updated: 2026-11-14 00:18:33 UTC 11 Applications Spreadsheet: Google Sheets ↗
Total Reported
323
11 applications 73.1% confirmed or fixed
Confirmed (Awaiting Fix)
154
47.7% Triaged by maintainers
Fixed Upstream
82
25.4% Patched in releases
Under Review
54
16.7% Awaiting maintainer triage
Rejected
33
10.2% Closed as invalid or duplicate

Reported Bugs by Application

Confirmed (Awaiting Fix) Fixed Pending Rejected
PyTorch Deep Learning 91 bugs
TensorFlow Deep Learning 61 bugs
JAX JAX Autodiff / ML 38 bugs
vLLM LLM Serving 34 bugs
SGL SGLang LLM Serving 33 bugs
ArangoDB Graph Database 24 bugs
cronie System Daemon 13 bugs
PostgreSQL Database Engine 10 bugs
Clang / LLVM Compiler / Toolchain 9 bugs
GCC GCC Compiler / Toolchain 5 bugs
OpenSSH Security / Network 5 bugs

Resolution Status

73.1% Confirmed + Fixed
236
Confirmed & Fixed
Awaiting Fix 154 (47.7%)
Fixed 82 (25.4%)
Pending 54 (16.7%)
Rejected 33 (10.2%)
Classification by Bug Type
Functional 158 Crash 88 Inconsistent 46 False positive 22 Overflow 3 Silent 2 Nan 2 Internal exception 2
Showing 323 of 323
No reported bugs matched your filter or search query.
PyTorch Deep Learning 91 bugs
41 Confirmed (Awaiting Fix) 28 Fixed 2 Pending 20 Rejected
Confirmed torch.logcumsumexp Inconsistent Centaur
torch.logcumsumexp runtime error inconsistency on CPU and CUDA (also documentation missing constraint)
Rejected torch.nn.functional.conv_transpose2d Inconsistent Not a bug Centaur
torch.nn.functional.conv_transpose2d produces inconsistent output on CPU and CUDA
Rejected torch.nn.functional.conv_transpose2d Overflow Duplicate Centaur
`torch.nn.functional.conv_transpose2d` has inconsistent handling of `float16` overflow on CPU
Confirmed torch.add Inconsistent Flashfuzz
Inconsistent float16 overflow behavior between CPU and CUDA devices
Confirmed torch.lcm_ Crash Flashfuzz
FPE when using `torch.lcm_` with int32 tensor and int16 scalar
Confirmed torch.clamp_min Inconsistent Flashfuzz
Inconsistent overflow handling in `torch.clamp_min` between CPU and CUDA for float16 tensors
Confirmed torch.fmod Crash Flashfuzz
FPE Crash in torch.fmod with large negative integer modulo value
Fixed torch.cosine_similarity Inconsistent Flashfuzz
Inconsistent Error Handling Between CPU and CUDA for `cosine_similarity`
Confirmed torch.addcmul Inconsistent Flashfuzz
Inconsistent Error Handling in `torch.addcmul` with `int16` Tensors
Confirmed torch.native_channel_shuffle Crash Flashfuzz
`torch.native_channel_shuffle` crashes with Floating Point Exception when given large integer parameter
Confirmed torch.pixel_shuffle() Crash Flashfuzz
FPE when calling `torch.pixel_shuffle()` with empty tensors and large upscale_factor
Confirmed torch.to_dense() Crash Flashfuzz
Segmentation fault when converting sparse COO tensor with complex values to dense
Confirmed torch.combinations Crash Flashfuzz
`torch.combinations` exhibits excessive memory usage and hangs for moderate `n` and `r` due to `n^r`
Confirmed torch.dequantize Inconsistent Flashfuzz
torch.dequantize result inconsistent on CPU and GPU
Confirmed torch.aminmax Inconsistent Flashfuzz
torch.aminmax reducing over all dims should not accept keepdim
Confirmed torch.batch_norm_update_stats Crash Flashfuzz
Memory Corruption in `torch.batch_norm_update_stats`
Confirmed torch.nn.InstanceNorm3d Inconsistent Greyfox
torch.nn.InstanceNorm3d producing inconsistent output for float16 tensors on CPU and GPU
Confirmed torch.dot Crash Greyfox
torch.dot has inconsistent support for int64 (long)
Rejected torch.nn.InstanceNorm2d Crash Not a bug Greyfox
torch.nn.InstanceNorm2d throws "mixed dtype" error with track_running_stats set to True
Confirmed torch.nn.ReflectionPad2d Crash Greyfox
torch.nn.ReflectionPad2d throws CUDA error on large tensors
Confirmed torch.logcumsumexp Crash Greyfox
torch.logcumsumexp runtime error inconsistency on CPU and CUDA (also documentation missing constraint)
Confirmed torch.Tensor.resize_ Crash RBF
`resize_` updates tensor shape metadata even when storage resize fails, creating corrupted "Zombie" tensors
Confirmed torch.nn.functional.pixel_shuffle Crash RBF
`torch.nn.functional.pixel_shuffle` throws SIGFPE (Floating Point Exception) with large upscale_factor
Fixed torch.nn.NLLLoss Inconsistent RBF
`torch.nn.NLLLoss` has inconsistent error handling: CUDA skips type check for empty inputs vs CPU
Confirmed torch.pixel_shuffle Overflow RBF
torch.pixel_shuffle CPU-GPU crash with massive upscale factor
Confirmed torch.fbgemm_linear_quantize_weight, torch.fbgemm_pack_gemm_matrix_fp16 Crash RBF
`torch.fbgemm_linear_quantize_weight` crashes with segmentation fault (SIGSEGV) with a cuda tensor
Rejected torch._stack Crash Not a bug RBF
`torch._stack` crashes with segmentation fault (`SIGSEGV`) on empty tensors
Fixed torch.native_channel_shuffle Crash RBF
`torch.native_channel_shuffle` crashes with FPE on GPU when `groups=0`
Confirmed torch.arange Crash RBF
`torch.arange` triggers `SIGFPE` (Floating Point Exception) instead of RuntimeError when writing float results to `int64` out tensor
Fixed torch.special.zeta, torch.special.spherical_bessel_j0, torch.special.i1, torch.special.modified_bessel_k0, torch.special.sinc, torch.special.psi, torch.special.modified_bessel_i1, torch.special.entr, torch.special.i1e, torch.special.log_ndtr, torch.special.i0, torch.special.hermite_polynomial_he, torch.special.i0e, torch.special.gammaln, torch.lgamma Inconsistent RBF
`torch.special.zeta` returns incorrect output with `uint16` on CUDA, along with many other `torch.special` APIs
Confirmed torch.linalg.slogdet, torch.linalg.inv, torch.inverse Inconsistent RBF
`torch.linalg.slogdet` does not propagate `NaN` in CUDA
Confirmed torch.logdet Inconsistent RBF
torch.logdet produces incorrect results for singular matrices on CUDA vs CPU
Confirmed torch._remove_batch_dim Crash RBF
Segmentation fault (core dumped) in `torch._remove_batch_dim`
Confirmed torch.lcm Crash RBF
Floating point exception on torch.lcm
Fixed torch.floor_divide Crash RBF
`torch.floor_divide` crashes with Floating point exception (`SIGFPE`) when dividing `INT64_MIN`
Fixed torch._C._nn.slow_conv3d Crash RBF
Segmentation Fault Due to Empty Stride Argument in `torch._C._nn.slow_conv3d`
Confirmed torch.addcmul Silent RBF
Feedback about torch.addcdiv
Fixed torch.nn.functional.softshrink Overflow Centaur
torch.nn.functional.softshrink throws overflow error on CUDA but not on CPU
Fixed torch.nn.functional.rrelu Inconsistent Centaur
`torch.nn.functional.rrelu` crashes on CPU with `training=True` when `lower` or `upper` is set to `inf`
Fixed torch.native_channel_shuffle Crash Centaur
`torch.native_channel_shuffle` throws FPE when `groups` > size of the second dimension of the input tensor
Fixed torch.nn.LPPool2d Inconsistent Centaur
`torch.nn.LPPool2d` throws inconsistent error on CPU and GPU
Fixed torch.nn.functional.normalize Nan Centaur
torch.nn.functional.normalize producing nan values with a large p value and tensor of complex numbers
Fixed torch.nanmean Inconsistent Flashfuzz
Inconsistent behavior and misleading error message for `torch.nanmean()` with complex dtypes
Fixed torch.arange Inconsistent Flashfuzz
Inconsistent behavior between CPU and GPU implementations of `torch.arange`
Fixed torch.batch_norm Inconsistent Flashfuzz
`torch.batch_norm` shows inconsistent error behavior between CPU and GPU
Fixed torch.fused_moving_avg_obs_fake_quant Inconsistent Flashfuzz
IInconsistent Error Handling in `torch.fused_moving_avg_obs_fake_quant` Between CPU and GPU Implementations
Fixed lp_pool1d Inconsistent Flashfuzz
`lp_pool1d` behavior inconsistency between CPU and GPU
Fixed torch.choose_qparams_optimized Crash Flashfuzz
Segmentation fault when calling `torch.choose_qparams_optimized()` with empty tensors and extreme num_bins value
Fixed torch.remainder Crash Flashfuzz
FPE in `torch.remainder`
Fixed torch.acos Inconsistent Greyfox
torch.acos produces different values using cpu and cuda for input with torch.complex128 dtype
Fixed torch.nn.ReplicationPad1d, torch.nn.ReflectionPad1d,torch.nn.ReplicationPad3d, torch.nn.ReflectionPad3d,torch.nn.ReplicationPad3d, torch.nn.ReflectionPad3d Crash Greyfox
torch.nn.ReplicationPad and torch.nn.ReflectionPad does not support float16
Fixed torch.sigmoid Nan Greyfox
torch.sigmoid producing nan for tensor of negative complex numbers on cpu
Fixed torch.asin Inconsistent Greyfox
torch.asin returns incorrect value with complex input on cpu
Fixed torch.histc Crash Greyfox
torch.histc not checking min > max on cuda for int8 tensors
Fixed torch.nn.LogSoftmax Inconsistent Centaur
`torch.nn.LogSoftmax` produces inf value on CPU but not on GPU
Fixed torch.nn.functional.celu Inconsistent Centaur
`torch.nn.functional.celu` produces inconsistent output on CPU vs CUDA
Rejected torch.floor_divide False positive Centaur
`torch.floor_divide` causes inconsistent precision
Rejected torch.fmin False positive Centaur
`torch.fmin` has inconsistent overflow behavior on CPU and GPU
Rejected torch.det False positive Centaur
`torch.det` returns inconsistent results for complex tensor on CPU vs GPU
Rejected torch.Tensor.put_ False positive Flashfuzz
Inconsistent behavior between CPU and GPU implementations of `torch.Tensor.put_` method
Rejected torch.Tensor.asin_ False positive Flashfuzz
Inconsistent Complex torch.Tensor.asin() Results Between CPU and GPU
Rejected torch.lu_unpack False positive Flashfuzz
Segmentation Fault in torch.lu_unpack() with bfloat16 Tensor
Rejected torch.cumsum False positive Flashfuzz
`torch.cumsum()` on bfloat16 tensor with dtype=torch.int8 produce inconsistent results between CPU and GPU
Rejected torch.linalg.lstsq False positive Greyfox
torch.linalg.lstsq generating different solutions on CPU and GPU
Rejected torch.digamma False positive Greyfox
torch.digamma produces inconsistent result between cpu and gpu for 0.00615 (float16)
Rejected torch.nansum False positive Greyfox
torch.nansum does not work with complex numbers on CPU
Rejected torch.nn.LocalResponseNorm False positive RBF
`torch.nn.LocalResponseNorm` raises `RuntimeError` on CPU for float16 inputs (parity issue with CUDA)
Rejected torch.nn.Linear False positive RBF
torch.nn.Linear shape mismatch with large/incompatible inpu
Rejected torch.index_select False positive RBF
torch.index_select out-of-bounds triggers inconsistent runtime errors
Rejected torch.addr False positive RBF
`torch.addr`: CPU/GPU Inconsistency with `int64` input and `float16` vector (CPU returns `inf`, GPU returns valid values)
Fixed torch.angle Crash Greyfox
torch.angle cuda version not implemented for bfloat16 and half (float16)
Confirmed torch.floor_divide Inconsistent Greyfox
torch.floor_divide not triggering ZeroDivisionError on GPU
Confirmed torch.nn.AvgPool2d Crash Greyfox
torch.nn.AvgPool2d works with long on cpu but not gpu
Confirmed torch.add, torch.matmul, torch.linalg.pinv Silent RBF
Silent Heap Buffer Overflow in standard ops (add, pinv) via implicit view resizing
Confirmed torch.clip Inconsistent RBF
`torch.clip` has checks for float16 scalar overflow on CPU but not on GPU
Confirmed torch.nn.PixelUnshuffle Inconsistent RBF
`torch.nn.PixelUnshuffle` throws RuntimeError: Stride calculation overflowed on CUDA with empty input and large downscale factor (CPU succeeds)
Rejected torch.linalg.cholesky_ex Inconsistent Not a bug RBF
`torch.linalg.cholesky_ex`: Inconsistency between CPU (Inf) and CUDA (NaN) for infinite input, with silent NaNs on CUDA
Confirmed torch.nn.functional.PoissonNLLLoss Inconsistent RBF
`torch.nn.functional.PoissonNLLLoss` returns `NaN` imaginary part on CPU (vs 0 on GPU) for overflowing complex inputs
Fixed torch.nn.functional.pdist Inconsistent RBF
`torch.nn.functional.pdist` (p=0) with `inf` inputs produce wildly different values
Confirmed torch-ldexp;torch-ldexp--long Crash Agentic-PBT
torch.ldexp raises NotImplementedError for a bool exponent (eager), while torch.compile and the decomposition accept it
Pending torch-distributions-cauchy;torch-distributions-cauchy--long Functional Agentic-PBT
torch.distributions.Cauchy.log_prob returns -inf (and NaN gradients) for large |value - loc| / scale because the square overflows
Pending torch-distributions-normal Functional Agentic-PBT
torch.distributions.Normal.log_prob returns NaN for very large or very small scale because scale**2 overflows/underflows
Confirmed torch-distributions-laplace Functional Agentic-PBT
torch.distributions.Laplace cdf/icdf return a zero gradient at the median (value == loc, p == 0.5)
Confirmed torch-distributions-laplace Functional Agentic-PBT
torch.distributions.Laplace.icdf returns -inf (and loses accuracy) for small tail probabilities
Confirmed torch-special-bessel-y0;torch-special-bessel-j1;torch-special-bessel-y1;torch-special-bessel-j0;torch-special-airy-ai Functional Agentic-PBT
torch.special.bessel_j0 / bessel_j1 / bessel_y0 lose ~9 significant digits in float64: the Cephes p1evl denominators drop their implied leading 1
Confirmed torch-special-logit Functional Agentic-PBT
[CPU] logit_backward and six activation backward kernels return a different gradient at NaN depending on tensor length
Rejected torch-maximum;torch-maximum--long Functional Agentic-PBT
torch.minimum and torch.maximum return a zero whose sign depends on the tensor length
Fixed torch-trace;torch-trace--opus-4-8 Crash Agentic-PBT
`torch.trace` raises `NotImplementedError` for `float16` input on CPU — no kernel registered for `Half`, no documented exclusion, NumPy handles it fine
Confirmed torch.unique Functional Agentic-PBT
torch.unique returns duplicate sub-tensors when the input contains NaN
Confirmed torch.arctanh Functional Agentic-PBT
torch.arctanh under torch.compile loses all precision for small inputs (returns 0.0)
Confirmed torch-clamp--long Functional Agentic-PBT
torch.clamp with a bound unrepresentable in the tensor dtype: eager raises, torch.compile ignores the bound
TensorFlow Deep Learning 61 bugs
38 Confirmed (Awaiting Fix) 12 Fixed 3 Pending 8 Rejected
Confirmed tf.linalg.svd Crash Centaur
`tf.linalg.svd` crashes on CUDA raising SIGABRT after facing OOM
Fixed tf.experimental.numpy.cumsum Inconsistent Centaur
`tf.experimental.numpy.cumsum` handles overflow inconsistently on CPU and GPU
Confirmed tf.linalg.matrix_rank Inconsistent Centaur
`tf.linalg.matrix_rank` produces inconsistent output on CPU vs GPU with `tol=6`
Confirmed tf.nn.local_response_normalization Inconsistent Centaur
`tf.nn.local_response_normalization` returns incorrect output
Confirmed tf.pow Inconsistent Centaur
`tf.pow` returns inconsistent value on CPU vs GPU
Confirmed tf.nn.depthwise_conv2d Crash Centaur
`tf.nn.depthwise_conv2d` crashes with large `strides` values when ONEDNN is enabled
Confirmed tf.raw_ops.NotEqual Inconsistent Centaur
Inconsistent behavior for `tf.raw_ops.NotEqual` between CPU and GPU with non-broadcastable shapes
Confirmed tf.linalg.eigh Crash Centaur
`tf.linalg.eigh` faces OOM on CUDA with 6D tensor and raises SIGABRT
Confirmed tf.linalg.lu_solve Inconsistent Centaur
`tf.linalg.lu_solve` returns inconsistent output from CPU vs GPU
Confirmed tf.raw_ops.MatrixDiag Crash Centaur
`tf.raw_ops.MatrixDiag` crashes with `Aborted (core dumped)` with a 6D tensor as input on GPU
Confirmed tf.linalg.pinv Crash Centaur
`tf.linalg.pinv` crashes with after raising SIGABRT with a 6D tensor due to OOM on CUDA
Confirmed tf.raw_ops.QuantizedConv2D Crash Flashfuzz
`tf.raw_ops.QuantizedConv2D` crashes with assertion failure instead of proper error handling
Confirmed tf.raw_ops.Mfcc Crash Flashfuzz
`tf.raw_ops.Mfcc` causes segmentation fault
Confirmed tf.raw_ops.Conv3DBackpropInputV2 Crash Flashfuzz
Segmentation Fault in `tf.raw_ops.Conv3DBackpropInputV2`
Confirmed tf.raw_ops.MaxPool Crash Flashfuzz
Segmentation Fault in `tf.raw_ops.MaxPool` with oneDNN and `NCHW_VECT_C`
Confirmed tf.raw_ops.QuantizeAndDequantizeV3 Crash Flashfuzz
Segmentation Fault in `tf.raw_ops.QuantizeAndDequantizeV3`
Confirmed tf.raw_ops.ResourceSparseApplyProximalAdagrad Crash Flashfuzz
Crash in `ResourceSparseApplyProximalAdagrad`
Confirmed tf.raw_ops.SparseBincount Crash Flashfuzz
Crash in `SparseBincount` due to overflow
Fixed tf.raw_ops.AvgPool3DGrad Crash Flashfuzz
Floating Point Exception in `AvgPool3DGrad`
Confirmed tf.raw_ops.MaxPool3D Crash Flashfuzz
FPE in oneDNN `MaxPool3D` with Invalid Dimensions
Confirmed tf.raw_ops.CropAndResizeGradBoxes Crash Flashfuzz
`Segmentation Fault` in CropAndResizeGradBoxes
Confirmed tf.raw_ops.HistogramFixedWidth Crash Flashfuzz
Segmentation Fault in `tf.raw_ops.HistogramFixedWidth`
Confirmed tf.raw_ops.ResourceSparseApplyAdagradDA Crash Flashfuzz
Crash in `tf.raw_ops.ResourceSparseApplyAdagradDA`
Fixed tf.raw_ops.ResourceSparseApplyMomentum Crash Flashfuzz
Segmentation Fault in `tf.raw_ops.ResourceSparseApplyMomentum`
Confirmed tf.raw_ops.TensorScatterMin Crash Flashfuzz
Fatal crash in `tf.raw_ops.TensorScatterMin` with malformed indices
Confirmed tf.raw_ops.TensorScatterMax Crash Flashfuzz
Fatal crash in `tf.raw_ops.TensorScatterMax` with malformed indices
Confirmed tf.raw_ops.SparseSegmentSumGradV2 Inconsistent Flashfuzz
Inconsistent Error Handling in `tf.raw_ops.SparseSegmentSumGradV2` Between CPU and GPU Implementations
Fixed tf.raw_ops.SparseToDense Inconsistent Flashfuzz
Inconsistent Error Handling in `tf.raw_ops.SparseToDense` Between CPU and GPU Implementations
Confirmed tf.raw_ops.BiasAdd Crash Flashfuzz
Crash in `tf.raw_ops.BiasAdd` when executing on GPU
Fixed tf.raw_ops.MaxPoolGradGradWithArgmax Crash Flashfuzz
Crash in `tf.raw_ops.MaxPoolGradGradWithArgmax` when executing on GPU
Confirmed tf.conv2d_backprop_filter_v2 Internal Exception Flashfuzz
`tf.conv2d_backprop_filter_v2` Fails with "no registered kernels" Error
Confirmed tf.linalg.triangular_solve Inconsistent Greyfox
Significant Discrepancy in `tf.linalg.triangular_solve` Results Between CPU and GPU
Confirmed tf.linalg.lstsq Inconsistent Greyfox
`tf.linalg.lstsq` producing outputs with large inconsistencies between CPU and GPU with `float32` tensors
Confirmed tf.math.floormod Inconsistent Greyfox
`tf.math.floormod` not throwing `Integer division by zero` error on GPU for tensor of int64 dtype
Confirmed tf.cast, tf.experimental.numpy.ascontiguousarray, tf.experimental.numpy.array, tf.saturate_cast, tf.image.convert_image_dtype Inconsistent RBF
Inconsistency (CPU vs. GPU) when casting NaN to int32 across multiple APIs
Confirmed tf.math.unsorted_segment_max Inconsistent RBF
`tf.math.unsorted_segment_max` fails to propagate NaNs on GPU (inconsistent with CPU)
Confirmed tf.eye Crash RBF
`tf.eye` with `TF_ENABLE_ONEDNN_OPTS=1` causes crash (Aborted/SegFault) on large `num_columns` instead of OOM
Confirmed tf.raw_ops.ParallelConcat Crash RBF
`tf.raw_ops.ParallelConcat` crashes (`SIGABRT`) due to integer overflow in shape calculation
Fixed tf.raw_ops.AudioSpectrogram Crash RBF
`tf.raw_ops.AudioSpectrogram` Segfaults with negative `window_size` and `stride`
Confirmed tf.raw_ops.ImmutableConst Crash RBF
`tf.raw_ops.ImmutableConst` causes Segfault (`SIGSEGV`) with large shape dimensions
Confirmed tf.experimental.numpy.amax Crash RBF
`tf.experimental.numpy.amax` crashes (`SIGABRT`) with specific int64 overflow in shape
Fixed tf.io.encode_png Crash RBF
`tf.io.encode_png` crashes process (`SIGABRT`) with `CHECK` failure when input has 0-dimension
Confirmed tf.linalg.det,tf.linalg.slogdet,tf.linalg.logdet,tf.linalg.cholesky,tf.linalg.inv Crash RBF
Aborted (core dumped) in `tf.linalg.det/slogdet/logdet/cholesky/inv`
Fixed tf.histogram_fixed_width Crash RBF
segmentation fault when tf.histogram_fixed_width receives large `value_range` and `nbins` on CPU mode
Fixed tf.math.argmax,tf.experimental.numpy.argmax False positive Centaur
`tf.math.argmax` throws `InvalidArgumentError` with valid `axis` of `int16` dtype
Fixed tf.raw_ops.SparseSegmentSqrtNGradV2 Inconsistent Flashfuzz
Inconsistent Error Handling in `tf.raw_ops.SparseSegmentSqrtNGradV2` Between CPU and GPU Implementations
Rejected tf.nn.conv2d_transpose Crash Centaur
`tf.nn.conv2d_transpose` crashes with "Illegal instruction (core dumped)"
Rejected tf.math.pow False positive Greyfox
Inconsistent Results Between CPU and GPU for `tf.math.pow`
Rejected tf.experimental.numpy.nansum False positive Greyfox
Discrepancy in `tf.experimental.numpy.nansum` with `dtype=float16` on CPU vs GPU
Rejected tf.compat.as_str_any False positive RBF
Misleading output when an invalid encoding Parameter Passed to tf.compat.as_str_any
Rejected tf.compat.dimension_value False positive RBF
tf.compat.dimension_value returns a Tensor when given a Tensor, despite documentation stating it returns only int or None.
Rejected tf.compat.forward_compatible False positive RBF
tf.compat.forward_compatible does not validate year/month/day ranges as documented
Rejected tf.concat False positive RBF
When a single tensor and out of range axis is passed to tf.concat it silently succeeds
Rejected tf.experimental.numpy.asarray False positive RBF
CPU/GPU inconsistency when casting bfloat16 to int8 for the api tf.experimental.numpy.asarray
Fixed tf.raw_ops.QuantizedMaxPool Internal Exception Flashfuzz
Unclear Error Message in tf.raw_ops.QuantizedMaxPool with oneDNN Enabled
Confirmed tf.gather Inconsistent Greyfox
Discrepancy in Error Handling for `tf.gather` on CPU vs GPU Leading to Undefined Behavior
Pending tf.raw_ops.SparseFillEmptyRowsGrad Inconsistent RBF
tf.raw_ops.SparseFillEmptyRowsGrad CPU-GPU inconsistency for out-of-range reverse_index_map
Fixed tf.experimental.numpy.ones, tf.experimental.numpy.zeros Crash RBF
`tf.experimental.numpy.ones` crashes (segfault/abort) when INT64_MAX is passed as shape
Pending tf.experimental.numpy.var, tf.experimental.numpy.all, tf.experimental.numpy.any, tf.experimental.numpy.amax, tf.experimental.numpy.amin, tf.experimental.numpy.count_nonzero, tf.experimental.numpy.max, tf.experimental.numpy.nanmean, tf.experimental.numpy.ptp, tf.experimental.numpy.nansum, tf.experimental.numpy.reshape, tf.experimental.numpy.promote_types, tf.experimental.numpy.var, tf.experimental.numpy.std Crash RBF
`tf.experimental.numpy.var` Causes `SIGABRT` on Empty Tensors with Large Dimensions
Confirmed tf.math.reduce_variance Crash RBF
`tf.math.reduce_variance` causes `SIGABRT` on large shapes containing zero
Pending tf.image.non_max_suppression Crash RBF
`tf.image.non_max_suppression` causes `SIGABRT` (Check Failure) with negative `max_output_size`
JAX JAX Autodiff / ML 38 bugs
17 Confirmed (Awaiting Fix) 17 Fixed 2 Pending 2 Rejected
Confirmed jax-scipy-special-betaln;jax-scipy-special-beta;jax-scipy-special-betaln--again;jax-scipy-special-beta--again;jax-scipy-special-betaln--long;jax-scipy-special-beta--long Functional Agentic-PBT
betaln(a, b) is too large by exactly min(a, b) when min(a, b)/max(a, b) is subnormal
Rejected jax-numpy-histogram Functional Agentic-PBT
jnp.histogram assigns values lying exactly on an interior bin edge to the wrong bin
Fixed jax-scipy-stats-pareto Functional Agentic-PBT
jax.scipy.stats.pareto returns -inf/inf/nan at the support boundary when scale is small relative to loc (support mask computed on unstandardized x)
Confirmed jax-scipy-stats-expon Functional Agentic-PBT
jax.scipy.stats.expon.logcdf returns -inf for small positive x (log1p(-sf) cancellation)
Confirmed jax-scipy-stats-laplace Functional Agentic-PBT
jax.scipy.stats.laplace.cdf gives NaN gradients in the tails (unselected branch overflows)
Fixed jax-jit Functional Agentic-PBT
jax.jit: negative static_argnums is not applied when the argument is passed by keyword (static_argnames not inferred)
Fixed jax-scipy-special-zeta;jax-scipy-special-zeta--long Functional Agentic-PBT
jax.scipy.special.zeta: the custom_jvp rule returns a different primal value than the function (nan vs finite for s < 1)
Confirmed jax-scipy-stats-poisson;jax-scipy-stats-poisson--long Functional Agentic-PBT
jax.scipy.stats.poisson.entropy returns NaN gradients for large mu (and small mu in float32)
Confirmed jax-scipy-special-betaln Functional Agentic-PBT
jax.scipy.special.beta returns nan / wrong sign at multiple defined points (negative integers, ±inf, -0)
Confirmed jax-scipy-special-i0;jax-scipy-special-i1 Functional Agentic-PBT
jax.scipy.special.i1 overflows to inf for large finite float64 input
Confirmed jax-scipy-stats-uniform Functional Agentic-PBT
jax.scipy.stats.uniform.cdf/pdf/logpdf silently map NaN input to finite values (pdf(nan) == 1.0)
Confirmed jax-scipy-stats-vonmises Functional Agentic-PBT
jax.scipy.stats.vonmises.pdf returns nan at kappa=0.0
Fixed jax-scipy-special-rel-entr Functional Agentic-PBT
jax.scipy.special.rel_entr(0.0, nan) returns inf instead of nan — SciPy compatibility mismatch
Confirmed jax-scipy-special-kl-div Functional Agentic-PBT
jax.scipy.special.kl_div returns nan instead of inf at x=inf, y=0.0
Confirmed jax-scipy-stats-cauchy Functional Agentic-PBT
jax.scipy.stats.cauchy.ppf/isf: no quantile-domain handling — ppf(0) returns large finite value (wrong sign in float32), q outside [0,1] returns garbage
Confirmed jax-scipy-stats-chi2 Functional Agentic-PBT
jax.scipy.stats.chi2.logpdf returns NaN at x=0 and x=inf when df=2 (inconsistent with gamma.logpdf)
Confirmed jax-scipy-stats-dirichlet Functional Agentic-PBT
jax.scipy.stats.dirichlet.logpdf returns -inf at simplex boundary where density diverges to +inf (alpha<1)
Confirmed jax-scipy-stats-multinomial Functional Agentic-PBT
jax.scipy.stats.multinomial.logpmf/pmf return -inf for all rows of a batched input (sum reduces over all axes, not axis=-1)
Confirmed jax-scipy-special-erfcx Functional Agentic-PBT
jax.scipy.special.erfcx returns spurious 0.0 in a band near x≈26.6 (should be ~0.0212; scipy/torch agree)
Fixed jax-scipy-special-xlog1py;jax-scipy-special-xlogy;jax-scipy-special-xlogy--opus-4-8 Functional Agentic-PBT
`jax.scipy.special.xlogy` and `xlog1py` return `0.0` instead of `nan` when `x=0` and `y=nan`
Fixed jax-numpy-abs;jax-numpy-abs--opus-4-8 Functional Agentic-PBT
`jnp.abs(inf + nan*j)` returns `nan` — violates IEEE 754 / C99 "infinity dominates" rule for complex magnitude
Fixed jax-scipy-special-digamma--opus-4-8 Functional Agentic-PBT
`jax.scipy.special.digamma(0.0)` returns `nan` — should be `-inf`
Confirmed jax-scipy-special-loggamma--again Functional Agentic-PBT
jax.scipy.special.loggamma returns nan instead of inf at x=0.0 and x=-0.0
Confirmed jax-scipy-special-zeta--again Functional Agentic-PBT
jax.lax.zeta gives wrong results for negative non-integer q (Euler-Maclaurin breakdown)
Fixed jax.Array.sum Functional Agentic-PBT
jax.Array.sum silently ignores the `initial` argument
Fixed jax.fwd_and_bwd Functional Agentic-PBT
jax.fwd_and_bwd silently discards keyword arguments
Fixed jax.lax.pmean Functional Agentic-PBT
`lax.pmean` with unequal `axis_index_groups` divides every group by the size of the first group
Fixed jax-scipy-special-gammaln--long Functional Agentic-PBT
jax.scipy.special.gammaln(-inf) returns +inf instead of -inf — sign-wrong at infinity edge
Confirmed jax-scipy-special-rel-entr;jax-scipy-special-kl-div;jax-scipy-special-kl-div--long Functional Agentic-PBT
jax.scipy.special.rel_entr / kl_div lose all precision for p close to q and return nan for large p (p*log(p) - p*log(q))
Pending jax-scipy-special-gammaincc;jax-scipy-special-gammaincc--long Functional Agentic-PBT
jax.scipy.special.gammaincc returns values far outside [0, 1] for large a (continued fraction stops unconverged)
Pending jax-random-poisson Functional Agentic-PBT
jax.random.poisson: samples for large lam (>~2e5) have the wrong variance (+46% at lam=1e8)
Fixed jax-scipy-stats-t Functional Agentic-PBT
jax.scipy.stats.t.logpdf is wrong for large df (lgamma(df/2) - lgamma((df+1)/2) cancels): logpdf(0, df=1e7) = 0 in float32
Fixed jax-numpy-linalg-eigh Functional Agentic-PBT
jnp.linalg.eigh silently accepts (…, 1, n) / (…, n, 1) input and returns the decomposition of a broadcast n x n matrix
Fixed jax-scipy-special-expn Functional Agentic-PBT
jax.scipy.special.expn(n, 0) returns 1/n instead of 1/(n-1) for n >= 2
Rejected jax-numpy-floor-divide;jax-numpy-floor-divide--long Functional Agentic-PBT
jnp.floor_divide / divmod return a quotient larger than x1/x2 for some float16 and bfloat16 inputs
Fixed jax-scipy-special-expi--again;jax-scipy-special-expi--long Functional Agentic-PBT
jax.scipy.special.expi(0.0)` returns nan instead of -inf
Fixed jax-scipy-stats-poisson--again Functional Agentic-PBT
jax.scipy.stats.poisson.logpmf` applies the integrality test to `k` instead of `k - loc`
Fixed jax#1;jax_1 Functional Agentic-PBT
`jnp.add.reduceat` and `jnp.multiply.reduceat` ignore the requested accumulator dtype
vLLM LLM Serving 34 bugs
20 Confirmed (Awaiting Fix) 9 Fixed 5 Pending
Confirmed vllm-tool-parsers-gigachat3-tool-parser Functional Agentic-PBT
[Bug]: gigachat3, granite and granite-20b-fc tool parsers lose tool calls when the whole message arrives in a single streaming delta
Pending vllm-reasoning-olmo3-reasoning-parser;vllm-reasoning-deepseek-r1-reasoning-parser;vllm-reasoning-ernie45-reasoning-parser;vllm-parser-abstract-parser Functional Agentic-PBT
[Bug]: Reasoning parsers mishandle markers in multi-token streaming deltas
Confirmed vllm-tokenizers-detokenizer-utils Crash Agentic-PBT
fix(logprobs): guard convert_ids_list_to_tokens against out-of-range ids
Confirmed vllm-utils-argparse-utils Functional Agentic-PBT
[Bugfix] Parse human-readable ints with Decimal to avoid off-by-one
Confirmed vllm-utils-argparse-utils Crash Agentic-PBT
[Bugfix] Avoid an IndexError for trailing dotted CLI args
Confirmed vllm-reasoning-olmo3-reasoning-parser Functional Agentic-PBT
[Bugfix][Reasoning] Olmo3: keep literal think tags in streamed content
Fixed vllm-engine-arg-utils-engineargs;vllm-engine-arg-utils-engineargs--reported;vllm-engine-arg-utils-engineargs--gpu;vllm-engine-arg-utils-engineargs--long Functional Agentic-PBT
[Bugfix] Let an optional Literal flag accept the None it advertises
Confirmed vllm-tool-parsers-internlm2-tool-parser;vllm-tool-parsers-internlm2-tool-parser--long Functional Agentic-PBT
[Bugfix][Tool Parser] Return early for unlisted tool names in InternLM2 parser
Confirmed vllm-utils-collection-utils Functional Agentic-PBT
[Bugfix] Invalidate LazyDict cache on __setitem__
Confirmed vllm-utils-collection-utils Functional Agentic-PBT
[Bugfix][Core] Preserve None values in swap_dict_values
Confirmed vllm-tool-parsers-utils Functional Agentic-PBT
[Bugfix] Fix extreme case in pythonic parser argument json safety
Fixed vllm-sampling-params-structuredoutputsparams Crash Agentic-PBT
[Bugfix][Structured Outputs] Reject empty `structural_tag` at request validation
Confirmed vllm-tool-parsers-xlam-tool-parser;vllm-tool-parsers-hunyuan-a13b-tool-parser Functional Agentic-PBT
[Bug]: xlam and hunyuan_a13b tool parsers drop the arguments of any tool call that follows an empty-argument call (streaming)
Confirmed vllm-reasoning-basic-parsers Functional Agentic-PBT
[Bugfix] Count reasoning tokens when the start token is absent
Confirmed vllm-entrypoints-openai-completion-protocol-completionreques Functional Agentic-PBT
[Bugfix] reject empty prompt lists in CompletionRequest
Confirmed vllm-entrypoints-openai-completion-protocol-completionreques Functional Agentic-PBT
[Bugfix][Frontend] Avoid mutating OpenAI request extra args
Fixed vllm-reasoning-deepseek-v3-reasoning-parser Functional Agentic-PBT
[Bugfix][Frontend] Count reasoning tokens for Harmony, DeepSeek-V3 and Step3 parsers
Confirmed vllm-reasoning-ernie45-reasoning-parser Functional Agentic-PBT
fix(reasoning): recompute ernie45 response-end index after stripping prefix
Fixed vllm-v1-structured-output Functional Agentic-PBT
[Bugfix] Escape control characters in xgrammar choice grammar
Confirmed vllm-v1-structured-output Functional Agentic-PBT
[Bugfix] Make Lark-to-EBNF quote handling escape-aware
Confirmed vllm-tool-parsers-abstract-tool-parser-toolparser Functional Agentic-PBT
[Bugfix] Avoid mutating tool parameters in _get_tool_schema_defs
Confirmed vllm-tool-parsers-internlm2-tool-parser;vllm-tool-parsers-internlm2-tool-parser--long Crash Agentic-PBT
[Bugfix] Limit split to first marker in internlm2 tool parser
Pending vllm-model-executor-models-qwen3-moe Functional Agentic-PBT
[Bug]: Qwen3MoeSparseMoeBlock rejects its documented 1D input
Pending vllm-model-executor-layers-rotary-embedding-fope Crash Agentic-PBT
[Bug]: FourierRotaryEmbedding pads sin with 1 and corrupts unrotated channels
Confirmed vllm-lora;vllm-lora--again;vllm-lora--long Functional Agentic-PBT
[Bugfix][LoRA] Skip no-LoRA tokens (index -1) in CPU torch bgmv ops
Pending vllm.multimodal.audio.AudioSpec Functional Agentic-PBT
[Bug]: AudioSpec(target_channels=1) does not produce 1D mono output for single-channel 2D audio
Fixed vllm.utils.system_utils.suppress_stdout Crash Agentic-PBT
[Bug]: vllm.utils.system_utils.suppress_stdout crashes when sys.stdout has no file descriptor
Fixed vllm.utils.system_utils.suppress_stdout Crash Agentic-PBT
[Bug]: vllm.utils.system_utils.suppress_stdout silently suppresses stderr when sys.stdout is bound to fd 2
Pending vllm.plugins.lora_resolvers.filesystem_resolver.register_filesystem_resolver Functional Agentic-PBT
[Bug]: FilesystemResolver loads adapters outside VLLM_LORA_RESOLVER_CACHE_DIR (path traversal via LoRA name)
Fixed vllm.models.deepseek_v4.common.mm_preprocess.DeepseekV4VLImageProcessor Functional Agentic-PBT
[Bug]: DeepSeek-V4 VL dummy image is not the worst case for multimodal profiling
Confirmed vllm.v1.sample.thinking_budget_state.ThinkingBudgetStateHolder Functional Agentic-PBT
[Bug]: ThinkingBudgetStateHolder maps every request to logits row 0 when spec decoding is on but a step has no drafts
Confirmed vllm.v1.sample.thinking_budget_state.ThinkingBudgetStateHolder Functional Agentic-PBT
[Bug]: ThinkingBudgetStateHolder.sync_batch leaks stale state on unidirectional moves
Fixed vllm_engine_reported/vllm-llm--engine-again;vllm_engine_reported/vllm-llm--engine-cutoff-window Crash Agentic-PBT
[Bug]: vLLM hangs indefinitely with low `num_gpu_blocks_override`
Fixed vllm_engine_reported/vllm-llm--engine-again;vllm_engine_reported/vllm-llm--engine-cutoff-window Crash Agentic-PBT
[Bug]: `block_size=8` triggers Triton CompilationError in FlexAttention kernel; other backends correctly reject
SGL SGLang LLM Serving 33 bugs
16 Confirmed (Awaiting Fix) 1 Fixed 15 Pending 1 Rejected
Confirmed sglang-srt-parser-code-completion-parser Functional Agentic-PBT
fix(completion): treat None suffix as empty in FIM prompt builder
Confirmed sglang-srt-function-call-qwen25-detector Functional Agentic-PBT
[Fix] Keep normal text preceding the tool-call opener in streaming parsers
Confirmed sglang-srt-function-call-qwen25-detector;sglang-srt-function-call-function-call-parser-functioncallpa;sglang-srt-function-call-llama32-detector;sglang-srt-function-call-gigachat3-detector;sglang-srt-function-call-hermes-detector--long;sglang-srt-function-call-llama32-detector--long;sglang-srt-function-call-mistral-detector--long Functional Agentic-PBT
[Bug] 9 tool-call parsers drop arguments when one streaming increment has more than one complete call
Pending sglang-srt-function-call-glm4-moe-detector;sglang-srt-function-call-glm4-moe-detector--long Functional Agentic-PBT
[Bug] 8 tool-call parsers give wrong or missing calls in token-by-token streaming
Confirmed sglang-srt-function-call-llama32-detector;sglang-srt-function-call-lfm2-detector--long;sglang-srt-function-call-llama32-detector--long Crash Agentic-PBT
fix(function_call): skip non-object entries in parse_base_json
Confirmed sglang-srt-function-call-mimo-detector;sglang-srt-function-call-mimo-detector--long Functional Agentic-PBT
[Bug] MiMo tool-call streaming: text after the last tool call is silently dropped; split bot-token flushes markup into content
Confirmed sglang-srt-function-call-trinity-detector Functional Agentic-PBT
fix(function_call): strip think tags split across stream chunks in Trinity
Confirmed sglang-srt-parser-reasoning-parser-reasoningparser;sglang-srt-parser-reasoning-parser-reasoningparser--long Functional Agentic-PBT
[Fix] Cohere Command4: strip echoed `<|START_THINKING|>` in streaming and flush the final text block
Confirmed sglang-srt-function-call-mimo-detector;sglang-srt-function-call-qwen3-coder-detector;sglang-srt-function-call-mimo-detector--long;sglang-srt-function-call-poolside-v1-detector--long Crash Agentic-PBT
fix(function_call): handle non-finite numbers in tool-call argument coercion
Rejected sglang-srt-layers-deep-gemm-wrapper-configurer Functional Agentic-PBT
[Bug] JIT DeepGEMM is enabled on SM121 (GB10 / DGX Spark): the SM120 exclusion checks `== 120` instead of the consumer-Blackwell family
Confirmed sglang-srt-parser-harmony-parser;sglang-srt-parser-harmony-parser--long Functional Agentic-PBT
[parser] Hold back harmony markers split across streaming chunks
Confirmed sglang-srt-constrained-xgrammar-backend--again;sglang-srt-constrained-xgrammar-backend--long Functional Agentic-PBT
fix: truncate xgrammar accepted_tokens in place on rollback
Pending sglang.srt.function_call.apertus2509_detector.Apertus2509Detector Functional Agentic-PBT
Apertus2509Detector (streaming) drops the second tool block and trailing text when they arrive in the same chunk`
Confirmed sglang-srt-entrypoints-ollama-serving--long Crash Agentic-PBT
[Ollama] Map options.seed to the sampling_seed field
Confirmed sglang-srt-function-call-deepseekv3-detector--long Functional Agentic-PBT
Tool-call parsers lose or corrupt data at streaming chunk boundaries (multiple detectors)
Confirmed sglang-srt-function-call-hunyuan-detector--long Crash Agentic-PBT
[Fix] hunyuan: Parse tool arguments with JSON Schema type arrays
Confirmed sglang-srt-function-call-llama32-detector--long;sglang#1 Functional Agentic-PBT
[Fix] function_call: Flush pending tool calls at end of stream in the Llama 3.2 detector
Confirmed sglang-srt-function-call-llama32-detector--long Functional Agentic-PBT
[Bugfix] Preserve single quotes inside JSON values in Llama32 streaming tool calls
Confirmed sglang-srt-function-call-mistral-detector--long;sglang_1 Functional Agentic-PBT
Tool-call parsers lose or corrupt data at streaming chunk boundaries (multiple detectors)
Fixed sglang_0517/sglang-srt-function-call-deepseekv32-detector--0517;sglang_0517/sglang-srt-function-call-deepseekv4-detector--0517 Crash Agentic-PBT
[Bug] deepseekv32/deepseekv4 tool-call parsers drop calls when the streamed output has two chunks
Pending sglang-srt-function-call-function-call-parser-functioncallpa Functional Agentic-PBT
[Bug] DeepSeekV31Detector.structure_info() omits `<|tool▁calls▁begin|>`, so its own detect_and_parse returns no tool call
Pending sglang-srt-function-call-glm4-moe-detector;sglang-srt-function-call-glm4-moe-detector--long Functional Agentic-PBT
[Bug] Glm4MoeDetector streaming emits invalid JSON when a non-string parameter's value is not valid JSON
Pending sglang-srt-function-call-glm4-moe-detector;sglang-srt-function-call-glm4-moe-detector--long Functional Agentic-PBT
[Bug] Glm4MoeDetector changes string-typed argument values that look like JSON ("true" → "True", "1.50" → "1.5")
Pending sglang-srt-entrypoints-openai-protocol;sglang-srt-entrypoints-openai-protocol--long Functional Agentic-PBT
[Bug] ChatCompletionRequest.set_json_schema mutates the caller's response_format schema (reused schemas change behaviour)
Pending sglang-srt-function-call-cohere-command4-detector;sglang-srt-function-call-cohere-command4-detector--long Functional Agentic-PBT
[Bug] cohere_command4 detector (streaming) drops the tool call when text and the whole action block arrive in one chunk
Pending sglang-srt-function-call-deepseekv31-detector;sglang-srt-function-call-deepseekv31-detector--long Functional Agentic-PBT
[Bug] deepseekv31 detector (streaming) drops the text that shares a delta with the start of a tool call
Pending sglang-srt-function-call-deepseekv31-detector;sglang-srt-function-call-deepseekv31-detector--long Functional Agentic-PBT
[Bug] deepseekv31 detector (streaming): greedy name regex drops a call and attaches the next call's arguments to it
Pending sglang-srt-function-call-hunyuan-detector;sglang-srt-function-call-hunyuan-detector--long Functional Agentic-PBT
[Bug] hunyuan detector (streaming): control characters in string arguments produce invalid, corrupted JSON
Pending sglang-srt-function-call-trinity-detector Functional Agentic-PBT
[Bug] trinity detector removes `<think>` / `</think>` from inside tool-call arguments
Pending sglang-srt-entrypoints-ollama-serving Functional Agentic-PBT
[Bug] Ollama streaming endpoints mangle the output when `--incremental-streaming-output` is on
Pending sglang-srt-utils-hf-transformers-common Crash Agentic-PBT
[Bug] `get_hf_text_config` raises AttributeError when `thinker_config` holds a dict `text_config`
Pending sglang-srt-parser-harmony-parser Functional Agentic-PBT
[Bug] HarmonyParser streaming: arguments of a tool call on the analysis channel are emitted as reasoning
Pending sglang-srt-constrained-xgrammar-backend--again;sglang-srt-constrained-xgrammar-backend--long Functional Agentic-PBT
[Bug] XGrammarGrammarBackend._sanitize_structural_format` skips `optional`, `star`, `plus`, `repeat`, `dispatch`, `token_dispatch` and `token_triggered_tags`, so a `null` `json_schema` inside them is rejected
ArangoDB Graph Database 24 bugs
1 Confirmed (Awaiting Fix) 10 Fixed 12 Pending 1 Rejected
Fixed aql-functions-string-substitute; co-author's Agentic-PBT run Functional Agentic-PBT
AQL `SUBSTITUTE()` returns `null` for an empty input string
Fixed aql-functions-string-substitute; co-author's Agentic-PBT run Functional Agentic-PBT
`SUBSTITUTE` returns null and an ICU warning for an empty input string
Fixed co-author's Agentic-PBT run Functional Agentic-PBT
`TO_NUMBER` converts a valid subnormal decimal string to zero
Pending co-author's Agentic-PBT run Functional Agentic-PBT
`APPEND(null, [])` returns null instead of the other input array
Fixed co-author's Agentic-PBT run Functional Agentic-PBT
REVERSE corrupts a single astral Unicode character
Pending co-author's Agentic-PBT run Functional Agentic-PBT
`FIND_FIRST` and `FIND_LAST` fail to find present control characters
Fixed co-author's Agentic-PBT run Functional Agentic-PBT
`REGEX_MATCHES` reports a match for an empty string that cannot match
Confirmed aql-operators-array-comparison-operators;aql-operators-array-comparison-operators--long Functional Agentic-PBT
AT LEAST (<n>) (<expression>) is true even when n is larger than the array's length
Fixed aql-high-level-operations-limit--again;indexes-and-search-indexing-working-with-indexes-persistent--again Functional Agentic-PBT
Persistent index with deterministic SORT and offset LIMIT incorrectly returns an empty result
Pending DATE_DIFF();DATE_ISO8601() Functional Agentic-PBT
ArangoDB DATE_ISO8601() applies the sign of a negative UTC offset only to its hours, not its minutes
Pending NGRAM_MATCH() Functional Agentic-PBT
ArangoDB `NGRAM_MATCH()` counts one attribute n-gram many times (stale DP diagonal)
Pending TO_ARRAY() Functional Agentic-PBT
ArangoDB AQL `SORTED()` replaces distinct elements that compare equal
Fixed JSON_PARSE();PI() Functional Agentic-PBT
`JSON_PARSE` exponent numbers can read beyond the input string
Fixed JSON_PARSE() Functional Agentic-PBT
`JSON_PARSE` can misround fractions and large integer tokens
Pending aql-execution-and-performance-query-optimization-optimizer-r Functional Agentic-PBT
AQL `use-indexes` uses a sparse index for a range condition that can match `null` when an `OR` branch proves non-null, dropping rows
Fixed aql-functions-string-substitute Functional Agentic-PBT
AQL `SUBSTITUTE()` with several search values skips a match that starts right where the previous replacement ended
Fixed aql-operators-arithmetic-operators Functional Agentic-PBT
AQL arithmetic on 64-bit integers from documents loses precision; the same expression on literals does not
Pending aql-operators-ternary-operator Functional Agentic-PBT
AQL `remove-redundant-calculations` treats `(a ? b : c) ? d : e` and `a ? b : (c ? d : e)` as the same calculation
Pending develop-http-api-import Functional Agentic-PBT
`POST /_api/import` stores duplicate attribute names from document and tabular input
Pending develop-http-api-import Functional Agentic-PBT
`/_api/import` with `type=documents` and `complete=true` commits a partial import when a line is not valid JSON
Rejected indexes-and-search-indexing-working-with-indexes-persistent--again Functional Agentic-PBT
AQL `COLLECT` served from a persistent index merges the array values `[]` and `[null]` into one group
Pending aql-operators-ternary-operator--again Crash Agentic-PBT
AQL ternary operator evaluates an untaken operand when it contains an array expansion with a subquery
Pending arangodb#1;arangodb_1 Crash Agentic-PBT
`DATE_ROUND` with a large positive amount terminates the server with SIGFPE
Pending arangodb#1;arangodb_1 Functional Agentic-PBT
`DATE_ROUND` assigns pre-1970 timestamps to the following bucket
cronie System Daemon 13 bugs
12 Confirmed (Awaiting Fix) 1 Fixed
Confirmed crontab-1-description;crontab-1-t;crontab-5-extensions Functional Agentic-PBT
Support input from a pipe for `crontab -T -`
Confirmed anacron-8-t Functional Agentic-PBT
`anacron -T` error stream contradicts its manual
Confirmed anacron-8-t;anacrontab-5-start-hours-range;anacrontab-5-description Functional Agentic-PBT
anacron: START_HOURS_RANGE, RANDOM_DELAY and PREFERRED_HOUR are matched by prefix, so other variables such as START_HOURS_RANGE_OLD are rejected or silently change scheduling
Confirmed anacron-8-t;anacrontab-5-description Functional Agentic-PBT
anacron: misspelled period macros such as `@monthlyXYZ` are accepted as `@monthly`, and a job with an unknown macro runs although anacron says it is skipped
Confirmed anacrontab-5-start-hours-range Functional Agentic-PBT
anacron: START_HOURS_RANGE hours are not range-checked (and overflow `int`), so invalid ranges pass `anacron -T` and silently stop all jobs
Confirmed cronnext-1-t Functional Agentic-PBT
cronnext: `-t` and `-q` parse the epoch time with `atoi()`, so times from 2038-01-19 on are truncated (Y2038)
Confirmed crontab-5-extensions Functional Agentic-PBT
@daily`/`@midnight` and `@hourly` do not set `DOM_STAR`/`DOW_STAR`, unlike their documented equivalents (cf. #52)
Confirmed crontab-5-description Functional Agentic-PBT
`cronnext` ignores a stepped day-of-week / day-of-month field
Confirmed crontab-5-description Functional Agentic-PBT
reversed ranges are accepted but never match
Confirmed crontab-5-cron-tz Functional Agentic-PBT
`cronnext` ignores `CRON_TZ`
Fixed cronnext-1 Functional Agentic-PBT
cronnext reports a time one hour late when skipping to a month with a different DST state
Confirmed cronnext-1 Functional Agentic-PBT
cronnext can skip an entire matching day across a DST spring-forward transition
Confirmed anacrontab-5-description Crash Agentic-PBT
anacron: a large `RANDOM_DELAY` value crashes anacron with SIGFPE or makes it loop forever
PostgreSQL Database Engine 10 bugs
2 Fixed 8 Pending
Pending pg_backup_start Functional Agentic-PBT
pg_backup_start accepts a newline in the label, and the backup will not start
Fixed last_value Functional Agentic-PBT
first_value/last_value/nth_value return NULL under EXCLUDE TIES when the current row is outside its frame
Fixed last_value Functional Agentic-PBT
first_value/last_value/nth_value return NULL under EXCLUDE TIES when the current row is outside its frame
Pending json_object Functional Agentic-PBT
an empty JSON_OBJECT rejects the ON NULL and WITH UNIQUE KEYS clauses that its documented syntax allows
Pending transaction_timestamp Functional Agentic-PBT
transaction_timestamp() and statement_timestamp() differ for a parameterized first statement, which the documentation says return the same value
Pending has_language_privilege Functional Agentic-PBT
has_language_privilege returns true for a nonexistent language OID when the current user is a superuser, where a named non-superuser gets null
Pending avg Functional Agentic-PBT
avg(double precision) overflows on cancelling finite inputs whose mean is representable
Pending jsonb_object_agg_unique_strict (round 003-jsonb-object-agg-unique-strict) Functional Agentic-PBT
jsonb_object_agg_unique_strict silently discards a JSONB null value
Pending json_object (round 004-json-object) Functional Agentic-PBT
JSON_OBJECT with ABSENT ON NULL accepts a NULL key and an array, composite or JSON key for the omitted pair
Pending uuidv7 (round 009-uuidv7) Functional Agentic-PBT
uuidv7(interval) rejects timestamps 1 to 999 microseconds into the last millisecond its field can hold
Clang / LLVM Compiler / Toolchain 9 bugs
6 Confirmed (Awaiting Fix) 1 Fixed 2 Pending
Confirmed alignment-builtins Crash Agentic-PBT
[Clang] Assertion `is<TypeInfoLValue>() && "not a type_info lvalue"' failed with alignment builtins
Confirmed __is_bitwise_cloneable Crash Agentic-PBT
clang `__is_bitwise_cloneable` hangs the compiler on nested structs (exponential, un-memoized recursion)
Pending model attribute Functional Agentic-PBT
clang `model` attribute silently ignored when the variable was referenced before its definition
Pending vectorcall attribute Functional Agentic-PBT
Clang misassigns homogeneous vector aggregate arguments for SysV `vectorcall`
Confirmed vectorcall attribute Crash Agentic-PBT
clang `__attribute__((vectorcall))` crashes on any `long double` parameter (x86_64 Linux)
Confirmed Alignment builtins Functional Agentic-PBT
[clang] `constexpr` `__builtin_align_down/up` on a pointer into an array reads the original element
Confirmed builtin-convertvector Crash Agentic-PBT
`__builtin_convertvector` from `__bf16` to `long double` vectors gives wrong values
Confirmed builtin-convertvector Functional Agentic-PBT
`__builtin_convertvector` to a `bool` vector folds 0.5 to `false` in a constant expression
Fixed nonnull attribute Functional Agentic-PBT
`nonnull` whose every index is rejected becomes "all pointers" and changes code
GCC GCC Compiler / Toolchain 5 bugs
3 Confirmed (Awaiting Fix) 1 Pending 1 Rejected
Confirmed builtin-is-virtual-base-of Functional Agentic-PBT
__builtin_is_virtual_base_of reports a non-virtual base reached through a virtual base as virtual
Pending builtin-stdc-rotate-left Functional Agentic-PBT
the manual's equivalent expression for __builtin_stdc_rotate_left/right is wrong for non-power-of-two _BitInt widths
Rejected hardbool-attribute Functional Agentic-PBT
implicitly zero-initialized hardbool objects hold zero bits instead of the false representation
Confirmed macros-with-a-variable-number-of-arguments Functional Agentic-PBT
the GCC manual says , ## __VA_ARGS__ drops the comma for an empty argument; the CPP manual and GCC keep it
Confirmed __builtin_bitreverseg Crash Agentic-PBT
__builtin_bitreverseg on a dependent C++ operand gives an ICE in mangle_decl
OpenSSH Security / Network 5 bugs
1 Fixed 4 Pending
Pending ssh-1-g Functional Agentic-PBT
ssh -G prints EscapeChar ^A as \^A which does not parse back
Pending ssh-1-g Functional Agentic-PBT
ssh -G shows some LocalForward entries as DynamicForward
Pending ssh-keygen-1-s Functional Agentic-PBT
ssh-keygen -s signs certificates with duplicate options, which sshd rejects
Fixed ssh-1-g--long Functional Agentic-PBT
StreamLocalBindMask ignores Host and Match blocks, and the last value wins
Pending openssh#1;openssh_1 Functional Agentic-PBT
ssh-keygen -F and -R miss a hashed known_hosts entry when the host name's case differs