[
https://issues.apache.org/jira/browse/MSHARED-1453?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17901329#comment-17901329
]
ASF GitHub Bot commented on MSHARED-1453:
-----------------------------------------
gnodet commented on PR #77:
URL: https://github.com/apache/maven-archiver/pull/77#issuecomment-2502177476
I'd go for something like:
```
private void createPropertiesFile(Properties properties, Path outputFile)
throws IOException {
Path outputDir = outputFile.getParent();
if (outputDir != null && !Files.isDirectory(outputDir)) {
Files.createDirectories(outputDir);
}
StringWriter sw = new StringWriter();
properties.store(sw, null);
String nl = System.lineSeparator();
String output = Stream.of(sw.toString().split("\\R"))
.filter(line -> !line.startsWith("#"))
.sorted()
.collect(Collectors.joining(nl, "", nl));
try (Writer pw = new CachingWriter(outputFile,
StandardCharsets.ISO_8859_1)) {
pw.write(output);
}
}
```
> Canonicalize properties files for reproducible builds
> -----------------------------------------------------
>
> Key: MSHARED-1453
> URL: https://issues.apache.org/jira/browse/MSHARED-1453
> Project: Maven Shared Components
> Issue Type: Bug
> Components: maven-archiver
> Reporter: Elliotte Rusty Harold
> Assignee: Elliotte Rusty Harold
> Priority: Minor
>
> See discussion on https://github.com/apache/maven-archiver/pull/77/files
> Briefly, properties files have non-unique representations and there's no
> guarantee two JDKs from different companies and Java versions produce the
> same byte-per-byte serialization. Our current code accounts for property
> order and comments, but not variations in escaping (hex vs. UTF-8), separator
> characters, and insignificant whitespace.
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
