Re: [boost] equivalent paths

From: Jeremy Maitin-Shepard (jbms_at_[hidden])
Date: 2003-12-08 18:02:32

Beman Dawes <bdawes_at_[hidden]> writes:

> At 02:30 PM 12/8/2003, Jeremy Maitin-Shepard wrote:
>> I suppose it is a matter of preference; it seems that relying on these
>> stored "volume serial numbers" is not a very good idea, because in
>> determining if paths are equal, we want to know if the devices are the
>> same.  Rather than making a change though, a warning in the docs might
>> be reasonable though.

> I've run some more tests, hard formatting memory cards with a camera. It
> assigned a volume serial number of 0.

> I also ran some more floppy tests, creating floppies with duplicate files that
> report equivalence when they shouldn't. However, it looks as if even when the
> volume serial number and indexes are the same, ftCreationTime is different. Thus
> it appears we can increase reliability by including ftCreationTime.dwLowDateTime
> and ftCreationTime.dwHighDateTime in the tests.

It would be nice if there were a way to get a real `device id' rather
than this unusual `volume id'.  Perhaps st_dev returned from _stat,
which seems to be provided, would work.

[snip]

-- 
Jeremy Maitin-Shepard