[
https://issues.apache.org/jira/browse/GROOVY-7465?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14995551#comment-14995551
]
ASF GitHub Bot commented on GROOVY-7465:
----------------------------------------
Github user PascalSchumacher commented on a diff in the pull request:
https://github.com/apache/incubator-groovy/pull/176#discussion_r44221781
--- Diff: src/main/org/codehaus/groovy/runtime/ResourceGroovyMethods.java
---
@@ -831,9 +836,8 @@ public static File leftShift(File file, InputStream
data) throws IOException {
}
/**
- * Write the text to the File, using the specified encoding. If the
given
- * charset is "UTF-16BE" or "UTF-16LE" (or an equivalent alias), the
- * requisite byte order mark is written to the file before the text.
+ * Write the text without a BOM to the File without writing a BOM,
using
--- End diff --
This sentence seems a bit confusing to me.
> ResourceGroovyMethods/NioGroovyMethods BOM behavior is inconsistent
> -------------------------------------------------------------------
>
> Key: GROOVY-7465
> URL: https://issues.apache.org/jira/browse/GROOVY-7465
> Project: Groovy
> Issue Type: Bug
> Components: groovy-jdk
> Reporter: Keegan Witt
> Assignee: Keegan Witt
> Labels: breaking_change
> Fix For: 2.5.0-beta-1
>
>
> Most users would expect {{withPrintWriter}} to be a convenience method that
> behaves the same as if they had called {{new PrintWriter}} themselves, but
> this is not the current behavior:
> {code:java}
> File file = new File("tmp.txt")
> try {
> String text = " "
> String charset = "UTF-16LE"
> file.withPrintWriter(charset) { it << text }
> println "withPrintWriter"
> file.getBytes().each { System.out.format("%02x ", it) }
> PrintWriter w = new PrintWriter(file, charset)
> w.print(text)
> w.close()
> println "\n\nnew PrintWriter"
> file.getBytes().each { System.out.format("%02x ", it) }
> } finally {
> file.delete()
> }
> {code}
> Outputs
> {noformat}
> withPrintWriter
> ff fe 20 00
> new PrintWriter
> 20 00
> {noformat}
> Additionally most users would expect that there's no difference in behavior
> between NIO and traditional methods, but this is also not the case
> {code:java}
> import java.nio.file.Files
> import java.nio.file.FileSystems
> import java.nio.file.Path
> File file = new File("tmp1.txt")
> Path path = FileSystems.getDefault().getPath("tmp2.txt")
> try {
> String text = " "
> String charset = "UTF-16LE"
> file.withPrintWriter(charset) { it << text }
> println "withPrintWriter"
> file.getBytes().each { System.out.format("%02x ", it) }
> path.withPrintWriter(charset) { it << text }
> println "\n\nnio withPrintWriter"
> path.getBytes().each { System.out.format("%02x ", it) }
> } finally {
> file.delete()
> Files.delete(path)
> }
> {code}
> outputs
> {noformat}
> withPrintWriter
> ff fe 20 00
> nio withPrintWriter
> 20 00
> {noformat}
> This is because {{ResourceGroovyMethods}} have a {{writeUTF16BomIfRequired}}
> method that {{NioGroovyMethods}} don't.
> Most likely we'd want to change {{ResourceGroovyMethods}} to not add the BOM
> by default, or at least allow the user to opt out of that behavior by doing
> something like adding a boolean argument to the existing methods that would
> control that behavior. The other option would be to make the NIO methods
> consistent by having them also use {{writeUTF16BomIfRequired}}.
> This began as a
> [discussion|http://mail-archives.apache.org/mod_mbox/incubator-groovy-users/201506.mbox/browser]
> on the user mailing list.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
