Correct memory indexing issue. (#165)

* Correct memory reservation. * Fix the vmImage version for MacOS CI pipeline. Co-authored-by: Wenbing Li <10278425+wenbingl@users.noreply.github.com>
microsoft · Oct 4, 2021 · 94beef0 · 94beef0
1 parent 11c2d6b
commit 94beef0
Show file tree

Hide file tree

Showing 2 changed files with 4 additions and 4 deletions.
diff --git a/ci_build/azure-pipelines/mshost.yaml b/ci_build/azure-pipelines/mshost.yaml
@@ -129,7 +129,7 @@ jobs:
 
   - job: MacOSWheel
     pool:
-      vmImage: 'macOS-latest'
+      vmImage: 'macOS-10.15'
 
     strategy:
       matrix:

diff --git a/operators/tokenizer/blingfire_sentencebreaker.cc b/operators/tokenizer/blingfire_sentencebreaker.cc
@@ -7,6 +7,7 @@
 #include <locale>
 #include <codecvt>
 #include <algorithm>
+#include <memory>
 
 KernelBlingFireSentenceBreaker::KernelBlingFireSentenceBreaker(OrtApi api, const OrtKernelInfo* info) : BaseKernel(api, info), max_sentence(-1) {
   model_data_ = ort_.KernelInfoGetAttribute<std::string>(info, "model");
@@ -41,10 +42,9 @@ void KernelBlingFireSentenceBreaker::Compute(OrtKernelContext* context) {
 
   std::string& input_string = input_data[0];
   int max_length = 2 * input_string.size() + 1;
-  std::string output_str;
-  output_str.reserve(max_length);
+  std::unique_ptr<char[]> output_str = std::make_unique<char[]>(max_length);
 
-  int output_length = TextToSentencesWithOffsetsWithModel(input_string.data(), input_string.size(), output_str.data(), nullptr, nullptr, max_length, model_.get());
+  int output_length = TextToSentencesWithOffsetsWithModel(input_string.data(), input_string.size(), output_str.get(), nullptr, nullptr, max_length, model_.get());
   if (output_length < 0) {
     ORT_CXX_API_THROW(MakeString("splitting input:\"", input_string, "\"  failed"), ORT_INVALID_ARGUMENT);
   }