Extracting zipped file from ResourceStream throws error “Invalid stored block lengths”

问题

I am trying to extract a ZIP file from my current JAR using:

InputStream resource = getClass().getClassLoader().getResourceAsStream(name);

This get the correct InputStream, but it gives an error when I try to unzip it using the following code (I'm storing each file into a Hashmap<file, filename>):

public static HashMap<String, String> readZip(InputStream inputStream) throws IOException {
    byte[] buffer = new byte[1024];
    HashMap<String, String> list = new HashMap<>();
    ZipInputStream zipInputStream = new ZipInputStream(inputStream);
    ZipEntry entry = zipInputStream.getNextEntry();
    while (entry != null) {
        if (!entry.isDirectory()) {
            StringBuilder stringBuilder = new StringBuilder();
            while (IOUtils.read(zipInputStream, buffer) > 0) {
                stringBuilder.append(new String(buffer, "UTF-8"));
            }
            list.put(stringBuilder.toString(), entry.getName());
        }
        zipInputStream.closeEntry();
        entry = zipInputStream.getNextEntry();
    }
    zipInputStream.closeEntry();
    zipInputStream.close();
    return list;
}

However when I try to do this, I get this exception (on IOUtils.read)

java.util.zip.ZipException: invalid stored block lengths
   at java.util.zip.InflaterInputStream.read(Unknown Source)
   at java.util.zip.ZipInputStream.read(Unknown Source)

Am I doing this wrong? I've done plenty of googling of the error, and I didn't see anything related to my issue.

回答1:

Thanks to @PaulBGD's answer above. It saved me hours to figure out what happened to my system I add the following new snippets into my pom.xml file (which I didn't realize is the root cause before reading Paul's answer):

<resources>
  <resource>
    <directory>src/main/resources</directory>
    <filtering>true</filtering>
    <includes>
      <include>**/*.version</include>
      <include>**/*.properties</include>
      <include>**/*.xml</include>
      <include>**/*.csv</include>
      <include>**/*.txt</include>
      <include>**/*.gif</include>
      <include>**/*.json</include>
      <include>**/*.xlsx</include>
      <include>rythm/**</include>
    </includes>
  </resource>
</resources>

However I didn't take Paul's answer directly, instead I don't think those xlsx files should go through the resource plugin's filtering pipeline, thus I added another resource into pom:

  <resource>
    <directory>src/main/resources</directory>
    <filtering>false</filtering>
    <includes>
      <include>**/*.xlsx</include>
    </includes>
  </resource>

And it fixed my problem without changing the source encoding setting from UTF-8 to ISO-8859-1

回答2:

After a few more hours of searching, I decompiled the maven-resources-plugin and noticed it used UTF-8 encoding by default. I quickly looked up the encoding I would need (ISO-8859-1) and put it in my pom.

<properties>
    <project.build.sourceEncoding>ISO-8859-1</project.build.sourceEncoding>
</properties>

Now the zip file copies into the jar perfectly, no corruption at all.

回答3:

Change spring-boot-starter-parent to 2.0.4.RELEASE. It worked for me.

  <parent>
        <groupId>org.springframework.boot</groupId>
        <artifactId>spring-boot-starter-parent</artifactId>
        <version>2.0.4.RELEASE</version>
        <relativePath />
    </parent>

来源：https://stackoverflow.com/questions/25235229/extracting-zipped-file-from-resourcestream-throws-error-invalid-stored-block-le