[
https://issues.apache.org/jira/browse/TIKA-3940?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17641933#comment-17641933
]
Tim Allison edited comment on TIKA-3940 at 12/1/22 2:07 PM:
------------------------------------------------------------
Current idea is to set LC_ALL=C.UTF-8 in env in ProcessBuilder on Linux OS?
This will fix the problem in the spawned process, but it will not fix the
problem in the main process.
Is this a "won't fix" with guidance on setting locale properly for Java?
was (Author: [email protected]):
Current idea is to set LC_ALL=C.UTF-8 in env in ProcessBuilder on Linux OS?
> FileSystemPipesIterator failing on linux when LC_CTYPE != UTF8
> --------------------------------------------------------------
>
> Key: TIKA-3940
> URL: https://issues.apache.org/jira/browse/TIKA-3940
> Project: Tika
> Issue Type: Task
> Reporter: Tim Allison
> Priority: Minor
>
> On Linux systems with LC_CTYPE set to C or something not UTF-8, the pipes
> iterator fails on filenames that include non-ASCII characters:
>
> {noformat}
> java.nio.file.InvalidPathException: Malformed input or input contains
> unmappable characters: ??????????????????????????????.txt {noformat}
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
