[
https://issues.apache.org/jira/browse/IO-171?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12625409#action_12625409
]
Matt Benson commented on IO-171:
--------------------------------
wow. Round and round. Maybe we should leave things as-is, and provide runtime
tests that require a writable location and test the case-sensitivity of a given
volume at RT. Maybe that's useful and maybe it isn't, but at least we can be
explicit about what we're doing and provide better guarantees. Note: AFAICT,
you must know (or be able to create) an existing file in order to make Java 5
on OS X tell you via canonical files whether the disk is case-sensitive. I
don't know what that actually indicates is going on in the java.io.File
implementation, obviously.
> IO assumes there are only two OSes: Windows and Unix
> ----------------------------------------------------
>
> Key: IO-171
> URL: https://issues.apache.org/jira/browse/IO-171
> Project: Commons IO
> Issue Type: Bug
> Reporter: Sebb
> Priority: Minor
>
> Just noticed that IO assumes systems are either:
> Windows-like:
> * File.separatorChar = \
> * Case-insensitive matching
> or
> Unix-like:
> * File.separatorChar = /
> * Case-sensitive matching
> It may well be true that there are only two different file sepator
> characters, but it is not true that all non-Windows systems use
> case-sensitive matching.
> For example, OpenVMS uses a file separator char of /, but originally only
> supported case-insensitive file names (always shown as uppercase).
> Current versions of OpenVMS support both upper and lower-case; the default is
> to use case-insenstive matching (and upper case names).
> Perhaps the IOCase.SYSTEM constant needs to have an "Unknown" or "Variable"
> setting for this; it would then be an error to use IOCase.SYSTEM to determine
> the case sensitivity.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
