From fc31aa3c4f99d6348e7c32a3e3c52c68b26ca700 Mon Sep 17 00:00:00 2001
From: David Wendt <45795991+davidwendt@users.noreply.github.com>
Date: Tue, 4 Jun 2024 10:19:30 -0400
Subject: [PATCH] Add overflow check when converting large strings to lists
 columns (#15887)

Fixes a couple places where strings columns are converted to lists column as binary -- chars are represented as INT8.
Since lists columns only support `size_type` offsets type, this change will throw an error if the size of the chars exceeds max `size_type` values.

Authors:
  - David Wendt (https://github.com/davidwendt)

Approvers:
  - Mike Wilson (https://github.com/hyperbolic2346)
  - MithunR (https://github.com/mythrocks)

URL: https://github.com/rapidsai/cudf/pull/15887
---
 cpp/src/io/utilities/column_buffer.cpp |  4 ++++
 cpp/src/reshape/byte_cast.cu           | 11 ++++++++---
 2 files changed, 12 insertions(+), 3 deletions(-)

diff --git a/cpp/src/io/utilities/column_buffer.cpp b/cpp/src/io/utilities/column_buffer.cpp
index e5d4e1a360f..27fc53fbc9e 100644
--- a/cpp/src/io/utilities/column_buffer.cpp
+++ b/cpp/src/io/utilities/column_buffer.cpp
@@ -191,6 +191,10 @@ std::unique_ptr<column> make_column(column_buffer_base<string_policy>& buffer,
         auto data      = col_content.data.release();
         auto char_size = data->size();
 
+        CUDF_EXPECTS(char_size < static_cast<std::size_t>(std::numeric_limits<size_type>::max()),
+                     "Cannot convert strings column to lists column due to size_type limit",
+                     std::overflow_error);
+
         auto uint8_col = std::make_unique<column>(
           data_type{type_id::UINT8}, char_size, std::move(*data), rmm::device_buffer{}, 0);
 
diff --git a/cpp/src/reshape/byte_cast.cu b/cpp/src/reshape/byte_cast.cu
index 1b05a9744fa..3dfa0b65814 100644
--- a/cpp/src/reshape/byte_cast.cu
+++ b/cpp/src/reshape/byte_cast.cu
@@ -135,9 +135,14 @@ struct byte_list_conversion_fn<T, std::enable_if_t<std::is_same_v<T, cudf::strin
         input.size(), output_type, stream, mr);
     }
 
-    auto col_content     = std::make_unique<column>(input, stream, mr)->release();
-    auto const num_chars = col_content.data->size();
-    auto uint8_col       = std::make_unique<column>(
+    auto const num_chars = strings_column_view(input).chars_size(stream);
+    CUDF_EXPECTS(num_chars < static_cast<int64_t>(std::numeric_limits<size_type>::max()),
+                 "Cannot convert strings column to lists column due to size_type limit",
+                 std::overflow_error);
+
+    auto col_content = std::make_unique<column>(input, stream, mr)->release();
+
+    auto uint8_col = std::make_unique<column>(
       output_type, num_chars, std::move(*(col_content.data)), rmm::device_buffer{}, 0);
 
     auto result = make_lists_column(