In article <200607041830.OAA22468@Sparkle.Rodents.Montreal.QC.CA>,
der Mouse  <mouse@Rodents.Montreal.QC.CA> wrote:
>>>> If an application unlinks a file which is opened, the file is not
>>>> deleted until it is closed, until that it exists as unnamed.  [...]
>>> There is one way in which it can be a problem without journaling: if
>>> the filesystem is made read-only after the unlink and before the
>>> close.  [...]
>> This one could be solved by deleting the file on disk when the
>> filesystem is made read-only, I think.
>
>That will "solve" the "problem" of the file remaining.  It will not,
>however, preserve the long--standing Unix semantics of continuing to
>use open-but-unlinked files.
>
>Now, one could decree that converting a RW filesystem to RO includes
>the destruction of any such files.  But it seems better to me to leave
>them unlinked and instead make sure the filesystem is not marked clean.

The transition to RO should not be allowed as long as processes have
files open for write, linked or unlinked.

christos