cpp版本识别重建表格时崩溃

[cty-ai]

• 系统环境/System Environment：ubuntu 1804
• 版本号/Version：Paddle： PaddleOCR： 问题相关组件/Related components：
• 运行指令/Command Code：
• 完整报错/Complete Error Message：
  Thread 3 "ppocr" received signal SIGSEGV, Segmentation fault.
  [Switching to Thread 0x7fffeb9af700 (LWP 24749)]
  0x00007fffef4334a0 in std::__cxx11::basic_string<char, std::char_traits, std::allocator >::find(char const*, unsigned long, unsigned long) const ()
  from /usr/lib/x86_64-linux-gnu/libstdc++.so.6
  (gdb) where
  #0 0x00007fffef4334a0 in std::__cxx11::basic_string<char, std::char_traits, std::allocator >::find(char const*, unsigned long, unsigned long) const ()
  from /usr/lib/x86_64-linux-gnu/libstdc++.so.6
  Upload PaddleOCR code  #1 0x000055555564d27a in PaddleOCR::PaddleStructure::rebuild_table (this=0x7fffeb99f1a0, structure_html_tags=std::vector of length 460, capacity 460 = {...},
  structure_boxes=std::vector of length 348, capacity 348 = {...}, ocr_result=std::vector of length 320, capacity 512 = {...})
  at /root/work/PaddleOCR/deploy/cpp_infer/src/paddlestructure.cpp:358
  del tmp #2 0x000055555564c79e in PaddleOCR::PaddleStructure::table (this=0x7fffeb99f1a0, img=..., structure_result=...)
  at /root/work/PaddleOCR/deploy/cpp_infer/src/paddlestructure.cpp:277
  optimize PaddleOCR  #3 0x000055555564b2a1 in PaddleOCR::PaddleStructure::structure (this=0x7fffeb99f1a0, srcimg=..., layout=true, table=true, ocr=true)
  at /root/work/PaddleOCR/deploy/cpp_infer/src/paddlestructure.cpp:106
  polish infer_rec and add ic15_dict #4 0x000055555560df32 in ocr_thread (param=0x7fffffffe260) at /root/work/PaddleOCR/deploy/cpp_infer/src/ekkocr.cpp:371
  Fixocr #5 0x00007ffff03bd6db in start_thread (arg=0x7fffeb9af700) at pthread_create.c:463
  Fix inference #6 0x00007fffeee2561f in clone () at ../sysdeps/unix/sysv/linux/x86_64/clone.S:95

源码:PaddleOCR//deploy/cpp_infer/src/paddlestructure.cpp :
170行: if (matched[td_tag_idx].size() > 0) {
172行: if (matched[td_tag_idx][0].find("") != std::string::npos &&
可能访问越界导致崩溃
170行修改为 if (td_tag_idx < matched.size() && matched[td_tag_idx].size() > 0) { 可解决

[github-actions]

This issue is stale because it has been open for 90 days with no activity.

[github-actions]

This issue was closed because it has been inactive for 14 days since being marked as stale.