linux编译运行jar包崩溃

[nuclear06]

• OS info

❯ lsb_release -a
No LSB modules are available.
Distributor ID: Ubuntu
Description: Ubuntu 22.04.3 LTS
Release: 22.04
Codename: jammy

• JDK

❯ java -version
java version "1.8.0_351"
Java(TM) SE Runtime Environment (build 1.8.0_351-b10)
Java HotSpot(TM) 64-Bit Server VM (build 25.351-b10, mixed mode)

• 流程

  1. 使用README提供的命令进行构建

  mkdir target\jfx\app
  cp -r models target\jfx\app
  mvn jfx:native -DskipTests -f pom.xml

  2. 编译完成后目录如下(省略部分路径)

  .
  ├── app.properties
  ├── docs
  ├── LICENSE
  ├── models
  ├── package.json
  ├── pom.xml
  ├── src
  ├── target
  │   ├── jfx
  │   │   ├── app
  │   │   └── native
  │   └── tools-ocr-2.2.9.jar
  └── targetjfxapp
  

  3. 进入路径target/jfx/native/treehole/app/

  cd target/jfx/native/treehole/app/
  
  # 使用提供的命令下载模型
  wget https://github.com/litongjava/tools-ocr/releases/download/model-ppocr-v4/ch_PP-OCRv4_det_infer-onnx.zip
  wget https://github.com/litongjava/tools-ocr/releases/download/model-ppocr-v4/ch_PP-OCRv4_rec_infer-onnx.zip
  mkdir -p models/ch_PP-OCRv4_det_infer
  mkdir -p models/ch_PP-OCRv4_rec_infer
  unzip ch_PP-OCRv4_det_infer-onnx.zip -d models/ch_PP-OCRv4_det_infer
  unzip ch_PP-OCRv4_rec_infer-onnx.zip -d models/ch_PP-OCRv4_rec_infer

  4. 当前的目录结构

  ❯ tree -L 2
  .
  ├── lib
  │   ├── api-0.25.0.jar
  │   ├── basicdataset-0.25.0.jar
  │   ├── commons-compress-1.23.0.jar
  │   ├── commons-csv-1.10.0.jar
  │   ├── commons-logging-1.2.jar
  │   ├── fontbox-2.0.24.jar
  │   ├── gson-2.10.1.jar
  │   ├── hutool-all-5.8.11.jar
  │   ├── imgscalr-lib-4.2.jar
  │   ├── jna-5.13.0.jar
  │   ├── jnativehook-2.1.0.jar
  │   ├── logback-classic-1.2.3.jar
  │   ├── logback-core-1.2.3.jar
  │   ├── model-zoo-0.25.0.jar
  │   ├── onnxruntime-1.16.0.jar
  │   ├── onnxruntime-engine-0.25.0.jar
  │   ├── opencv-0.25.0.jar
  │   ├── opencv-4.7.0-0.jar
  │   ├── pdfbox-2.0.24.jar
  │   ├── pytorch-engine-0.25.0.jar
  │   └── slf4j-api-1.7.25.jar
  ├── models
  │   ├── ch_PP-OCRv4_det_infer
  │   └── ch_PP-OCRv4_rec_infer
  ├── tools-ocr-2.2.9-jfx.jar
  └── treehole.cfg
  

5. 运行tools-ocr-2.2.9-jfx.jar

Java运行输出

❯ java -jar tools-ocr-2.2.9-jfx.jar

2024-01-03 00:24:03.259 [JavaFX-Launcher] WARN SimpleRepository.getMetadata:212 - Simple repository pointing
to a non-archive file.
Loading: 100% |████████████████████████████████████████|
2024-01-03 00:24:03.480 [JavaFX-Launcher] WARN LibUtils.downloadPyTorch:444 - No matching cuda flavor for lin
ux-x86_64 found: cu117.
2024-01-03 00:24:03.817 [JavaFX-Launcher] INFO PtEngine.newInstance:67 - PyTorch graph executor optimizer is
enabled, this may impact your inference latency and throughput. See: https://docs.djl.ai/docs/development/infe
rence_performance_optimization.html#graph-executor-optimization
2024-01-03 00:24:03.820 [JavaFX-Launcher] INFO PtEngine.newInstance:72 - Number of inter-op threads is 8
2024-01-03 00:24:03.820 [JavaFX-Launcher] INFO PtEngine.newInstance:73 - Number of intra-op threads is 8
2024-01-03 00:24:03.938 [JavaFX-Launcher] WARN SimpleRepository.getMetadata:212 - Simple repository pointing
to a non-archive file.
Loading: 100% |████████████████████████████████████████|
2024-01-03 00:24:04.076 [JavaFX Application Thread] INFO MainForm.init:58 - primaryStage:javafx.stage.Stage@1
c648f55
#
# A fatal error has been detected by the Java Runtime Environment:
#
# SIGSEGV (0xb) at pc=0x00007fa4b4897ef4, pid=474539, tid=0x00007fa47fb20640
#
# JRE version: Java(TM) SE Runtime Environment (8.0_351-b10) (build 1.8.0_351-b10)
# Java VM: Java HotSpot(TM) 64-Bit Server VM (25.351-b10 mixed mode linux-amd64 compressed oops)
# Problematic frame:
# C [libc.so.6+0x97ef4] pthread_mutex_lock+0x4
#
# Failed to write core dump. Core dumps have been disabled. To enable core dumping, try "ulimit -c unlimited"
before starting Java again
#
# An error report file with more information is saved as:
# /home/saniter/my_project/tools-ocr/target/jfx/native/treehole/app/hs_err_pid474539.log
#
# If you would like to submit a bug report, please visit:
# http://bugreport.java.com/bugreport/crash.jsp
# The crash happened outside the Java Virtual Machine in native code.
# See problematic frame for where to report the bug.
#
[1] 474539 IOT instruction (core dumped) java -jar tools-ocr-2.2.9-jfx.jar

6. 日志hs_err_pid474539.log 内容
   logs

我的操作有哪一步存在问题吗？任何指导都非常感谢！

[litongjava]

1.请使用最版本测试一下
2.使用这个文件中的docker镜像 https://github.com/AnyListen/tools-ocr/blob/master/.github/workflows/build.yml

[nuclear06]

TL;DR:
已成功解决，为上游代码问题，希望作者更新代码。此外，在使用中发现一个截图偏移的问题。

存在的问题

1. 新的代码中使用了模型ONNX_PPOCR_V4_SERVER，但是上游项目RapidOcr-Java回滚了该更新
   请问为什么ONNX_PPOCR_V4_SERVER被注释了 MyMonsterCat/RapidOcr-Java#46
   基于该issue重新编译rapidocr-onnx-models-1.2.2.jar，rapidocr-common-0.0.7.jar
ps: 手动编译后不需要按照build.yml中下载模型文件到指定路径

2. 在手动编译jar包后依旧出现core dump
   怀疑为上游jnativehook的问题
   SIGSEGV when registering hook in JavaFX application kwhat/jnativehook#442
   怀疑为版本问题，尝试使用不同版本
   在更换版本为2.0.2时可以正常运行，并且，在全部共4个版本中仅有2.0.2版本编译出的可以正常运行

新发现的问题

在使用截图ocr时截图的框和实际内容不匹配，存在垂直下移的问题，猜测为全局缩放率导致的，