tag:blogger.com,1999:blog-7094652.post6956573249815468923..comments2024-03-23T14:36:09.980+00:00Comments on Neil Mitchell's Blog (Haskell etc): Exploding Patterns in FilePathNeil Mitchellhttp://www.blogger.com/profile/13084722756124486154noreply@blogger.comBlogger5125tag:blogger.com,1999:blog-7094652.post-35023485163274378582007-04-24T10:06:00.000+01:002007-04-24T10:06:00.000+01:00To be even more confusing, HFS+ filesystems can be...To be even more confusing, HFS+ filesystems can be case sensitive: it's a per filesystem setting. You're not allowed to make the system partition case sensitive though, as it would break the OS.Anonymousnoreply@blogger.comtag:blogger.com,1999:blog-7094652.post-19185986601271389372007-04-15T02:05:00.000+01:002007-04-15T02:05:00.000+01:00And on Windows a drive can mount NFS drives as wel...And on Windows a drive can mount NFS drives as well, which are case sensitive. You are indeed right that its a drive property, but with symlinks and no real way to query drives, its impossible to get right. Some sensible guessing is the only alternative.Neil Mitchellhttps://www.blogger.com/profile/13084722756124486154noreply@blogger.comtag:blogger.com,1999:blog-7094652.post-45060296612640279162007-04-15T01:02:00.000+01:002007-04-15T01:02:00.000+01:00ah, I see. That does sound safe enough... althoug...ah, I see. That does sound safe enough... although I'm noting that on my macs HFS+ filenames are not case sensitive, either... since it's really more of a filesystem behavior than a platform behavior, I guess it's hard to choose something that's always right. <BR/><BR/>Regardless, this looks like a nice library, and I'll be making use of it soon! Thanks.Anonymousnoreply@blogger.comtag:blogger.com,1999:blog-7094652.post-86288105858790936032007-04-14T11:22:00.000+01:002007-04-14T11:22:00.000+01:00Richard: In this particular case, no. The function...Richard: In this particular case, no. The function compares equality of FilePath's, and the Windows one makes the special action of mapping toLower over the string first. The safe default option is to not do this, which is also what Posix does. So any added systems will be safe if they take the Posix route.Neil Mitchellhttps://www.blogger.com/profile/13084722756124486154noreply@blogger.comtag:blogger.com,1999:blog-7094652.post-80502173133257322582007-04-14T00:14:00.000+01:002007-04-14T00:14:00.000+01:00Isn't it more "future-proof" to add an otherwise c...Isn't it more "future-proof" to add an otherwise case at the end, which raises an error, rather than replace the isPosix one?Anonymousnoreply@blogger.com