Marius Gedminas wrote:

On Sat, Apr 03, 1999 at 12:29:12PM +0200, Peter Gerwinski wrote:

...

Don't we need an additional variable `CasePreserving' then which is false on FAT but true on VFAT filesystems?

What would it be used for?

I don't think we need it.

Actually, one can't rely on file names being stored unmodified at all. E.g., Dos (without LFN) will strip a trailing `.', except in `.' and `..', and anything beyond the 8+3 limitation, anyway. So, one must not rely on finding the file name as one stores it. All one can expect (I hope, at least) is that one can access the file by the same file name one used for storing (i.e., it will be modified the same way...)

If one wants to canonicalize a file name, one can try `FExpand'...

...

Even worse: This needn't be a constant for one OS but depends on the path you are accessing. A file on a FAT filesystem mounted under Linux has LFN=CaseSensitive=CasePreserving=false.

I really think there is no need to go into that level of details. Other Linux programs do not care for this, why should programs compiled with GPC be an exception?

I agree. I don't think a Linux user mounting a FAT FS with a file `foo' would expect it to show up when expanding `F*' (better mount it umsdos, anyway ;-).

Frank

-- Frank Heckenbach, frank@fjf.gnu.de, http://fjf.gnu.de/ GPC To-Do list, latest features, fixed bugs: http://agnes.dida.physik.uni-essen.de/~gnu-pascal/todo.html