In a normal command line I would put single quotes around that and it would be fine, or backwhack the /. Unfortunately, Racket disliked both of those solutions, hence the problem.
That seems strange, if you're on a Unix like GNU/Linux or a BSD. I might be misunderstanding, or maybe Racket is doing something that I'd find surprising (and is perhaps a bug), or maybe we should clarify for the list...
When you got a forward-slash in a filename (not a pathname), was it on MS Windows or in a particular Mac OS X interface context? Or maybe it wasn't a ASCII character 47 (decimal), but a Unicode lookalike glyph? Or maybe some other problem was confusing things?
On a Unix, I assume that Racket `process*`-like procedures will pass through the argument strings to the OS system call API, with at most character encoding translation, but not interpret escaping/quoting.
On a Unix, you really-really shouldn't be able to put a forward-slash (ASCII 47) in a filename. Even if the kernel/filesystem would permit it, that would break other things throughout the system. That might even create a new class of security exploits (e.g., create a filename that uses "/", "..", and/or ".", which a later privileged process interprets differently, to mutate/delete another file, under your control). (And, if it's different forward-slash Unicode character, it's asking to confuse a human in the loop, like a lookalike-glyph impostor banking Web site domain name.) Even if you got one filesystem to support it on a particular platform, on most other filesystems/protocols that OS might also end up using (e.g., mounted removable USB drives for sneakernet with other systems, some network filesystems, simple cloud storage mapping like in AWS S3) something would likely to break.
I just read that at least one Mac OS X filesystem, at a low level, permits arbitrary characters in a filename, but that they prevent some problematic characters (including forward-slash) through most interfaces. This sounds like a good call, to me.
Given the frequent desire to leverage the huge wealth of existing Posix-assuming stuff and user understanding, and to interoperate and possibly be portable... Even when we can put ASCII 47 forward-slash (and other especially problematic characters) in a filename on a particular platform, through a particular interface... usually we want to avoid that, and design our systems from the start not to do that.
-- You received this message because you are subscribed to the Google Groups "Racket Users" group. To unsubscribe from this group and stop receiving emails from it, send an email to racket-users+unsubscr...@googlegroups.com. For more options, visit https://groups.google.com/d/optout.
