On Sat, Nov 17, 2001 at 12:15:00AM +0100, Neal H Walfield wrote:
> > Those void functions you changed to error_t can never fail, so I don't see
> > the point. The actual diskfs_make_node changes seem ok.
>
> They can never fail today, however, I thought it made the
--- Roland McGrath <[EMAIL PROTECTED]> wrote:
> I don't see anything wrong with the current diskfs_make_node signature.
> It can only return failure for a malloc failure.
>
The advantages I see of using a signature that returns an error are:
1. Any users of make_node,
I don't see anything wrong with the current diskfs_make_node signature.
It can only return failure for a malloc failure.
___
Bug-hurd mailing list
[EMAIL PROTECTED]
http://mail.gnu.org/mailman/listinfo/bug-hurd
I changed the signature of diskfs_make_node to return an error_t like
diskfs_make_peropen or diskfs_make_protid.
I'm holding back on the libnetfs stuff until I know how make_protid should
act wrt it iouser struct.
2002-04-02 James A. Morrison <[EMAIL PROTECTED]>
ext2fs:
> Those void functions you changed to error_t can never fail, so I don't see
> the point. The actual diskfs_make_node changes seem ok.
They can never fail today, however, I thought it made the interface a
bit more consistent. I guess, however, you are right, there is no
need.
Here
Those void functions you changed to error_t can never fail, so I don't see
the point. The actual diskfs_make_node changes seem ok.
___
Bug-hurd mailing list
[EMAIL PROTECTED]
http://mail.gnu.org/mailman/listinfo/bug-hurd
The way we use diskfs_make_node is currently unsafe: we almost never
check the return value, however, we really should. The following
patches fix this. However, rather than simply correcting the
problems, I opted to fix the error at its root: the interface. Now,
we return an error code and
