Mail Archives: djgpp-workers/1998/03/16/07:47:46
> > > I think testing for a full disk is too expensive, and also doesn't
> > > cover all the cases (e.g., a write-protected floppy that isn't full).
> >
> > Hardly too expensive, it's only done when an error occurs.
>
> But in some cases in point (a write-protected disk, for example) errors
> happen all the time, and we don't want djtar to bail out right away, so
> you will see quite a few of the errors and statfs it every time.
Yes, but isn't that the abnormal case.
> > Isn't some better diagnostics worth the time?
>
> This diagnostics will be printed anyway if the disk fills while writing a
> file (as opposed to creating a directory). Maybe djtar should also test
> if the disk is full before it actually begins to work, or before creating
> each directory?
Yes for djtar, but I'd like better errno's from mkdir, so it's more like
Unix.
> I think the limitation is not 24 slashes, it's the length of the path
> name. Anything longer than 64 characters should make DOS barf.
Hmm. "1/2/3/4/5/6/7/8/9/10/11/12/13/14/15/16/17/18/19/20/21/22/23/24/25"
is 55 long, so that's not right.
> But the problem is that this limitation is not consistently implemented
> (some DOS functions will succeed even if you exceed the limits), and
> Windows 95 (and probably 32-bit File Access in Windows 3.11) don't have
> these limitations. So I'm not sure if we should depend on this.
Uggh! So is it possible to get more information from DOS, why the operation
failed and set errno from that?
> `_truename' will give you hell and hight water if the drive is mounted
> via a network. I think it's better to use `_fixpath', if you need that.
?
9o6,
MartinS
- Raw text -