> It seems Windows doesn't have lstat, though... Windows API behaves differently from POSIX. Testing is required, but unfortunately I don't have any Windows machines. Maybe on this OS delete-file and delete-file* already act consistently.
> your ticket did get through, three times. Oh, sorry, I didn't notice. Thanks! On Fri, Jan 2, 2026 at 5:30 PM Peter Bex <[email protected]> wrote: > On Thu, Jan 01, 2026 at 02:54:38PM +0100, Felix Winkelmann via > Chicken-hackers wrote: > > Hi! > > > > Thanks for the patch. I'm not sure how to proceed, though. I think > > "file-exists?" should be as portable as possible and do not perform > > operating-system specific stuff like symlinks, etc., so in the > > interest of transparency, I would say the current behaviour is > > the best one (until you need the opposite, of course). > > The patch uses lstat instead of stat, which is also in POSIX, so the > function itself is not OS-specific. It seems Windows doesn't have > lstat, though... > > > So I'm undecided, what do others think? There still is the posix file > module > > that gives one full control over how symlinks are handled (or not). > > > > In any case, "delete-file*" should work like "delete-file" for > > dangling symlinks, in my opinion. I will prepare a patch for > > that. > > From a user perspective, the two procedures should behave the same, except > that the starred version returns #f whenever the unstarred one would throw > an exception. > > PS: Regarding the 500 error on Trac, your ticket did get through, three > times. I closed the duplicates of https://bugs.call-cc.org/ticket/1863 > and removed some old sessions (a known issue with Trac due to which it > sometimes just gets "stuck") > > Cheers, > Peter >
