Racket and Geiser on Windows, strange path error

Question

When I try (run-racket) or (run-geiser) on my Windows 8.1 machine running GNU Emacs 24.3 and Racket v6.1 I get the following output:

Welcome to Racket v6.1.
default-load-handler: cannot open input file
  path: f:/c/Users/James/AppData/Roaming/.emacs.d/el-get/geiser/scheme/racket/geiser/startup.rkt
  system error: The system cannot find the path specified.; errno=3
>

Which is decidedly odd, but after browsing the backtrace I still don't have a good idea what's going on.

Indeed, since Racket starts up fine it seems to be some strangeness in the way paths are passed from Geiser to Racket. The only other time I see this kind of path strangeness is when MinGW bash is involved, and that /c/ makes me think that it probably is. In fact, that path would be perfectly valid if not for the leading f:.

I'm at a bit of a loss as to how to approach this problem, as I don't know how Geiser communicates with Racket (I just started trying to use it today), but I would really like to get it to work, as I'm quite invested in Emacs as my editor.

So my question is thus: What possible problem points should I investigate, given the above as clues?

Answer 1

After some digging through the Geiser source I found in geiser/elisp/geiser-load.el the line:

(setq geiser-scheme-dir "/c/Users/James/AppData/Roaming/.emacs.d/el-get/geiser/scheme")

Which was probably generated incorrectly by MinGW make.

I changed it to:

(setq geiser-scheme-dir "c:/Users/James/AppData/Roaming/.emacs.d/el-get/geiser/scheme")

After which everything works perfectly.

The problem was that Emacs' expand-file-name assumed that the /c/... path was a relative path due to the initial forward slash, and guessed that it should prepend the letter of the current drive to make it a proper Windows Emacs path.