C is too old? opinions?

[Jean-Marc Bourguet]

In that case the '/' will cause problems. I meant filenames without
paths. Paths are normally controlled through unmentioned compiler
options anyhow. And I have little sympathy for people who give different
files the same name.

I don't have the luxery of working only with what I've written and I've
spend to much time integrating together things which where not designed to
be integrated (they started in different companies), not to have sympathy
for people struggling with name conflict in general.

Using prefixes help somewhat but the more robust way I'm aware is putting
some generated unique id in the guard (for example combining the time, a
numeric id of the host -- MAC address? -- and a numeric id of the user).

Yours,

 

[Philip Potter]

In that case the '/' will cause problems. I meant filenames
without paths. Paths are normally controlled through unmentioned
compiler options anyhow.

Paths are often partially specified in #include directives. But
regardless of how the path is deduced, headers with the same name will
be a problem.

And I have little sympathy for people who
give different files the same name.

What alternative do you propose to avoid collisions? Should filenames be
allocated like MAC addresses? Chris Torek has already presented some
reasonable situations where it might occur.

 

[CBFalconer]

CBFalconer wrote:
.... snip ...


What alternative do you propose to avoid collisions? Should
filenames be allocated like MAC addresses? Chris Torek has
already presented some reasonable situations where it might occur.

I don't think the problem is all that big. Normally the user has
the choice of filename for at least one of the problem makers.

 

[jacob navia]

What alternative do you propose to avoid collisions? Should filenames be
allocated like MAC addresses? Chris Torek has already presented some
reasonable situations where it might occur.

Even MAC addresses will not work since a MAC address can have a LOT of
files!

 

[Philip Potter]

I don't think the problem is all that big. Normally the user has
the choice of filename for at least one of the problem makers.

Only if he wrote them himself.

 

[Joachim Schmitz]

Even MAC addresses will not work since a MAC address can have a LOT of
files!

Huh? Allocated _like_ a MAC address, MAC addresses are (supposed to be)
unique, worldwide.

Bye, Jojo

 

[Dik T. Winter]

> Huh? Allocated _like_ a MAC address, MAC addresses are (supposed to be)
> unique, worldwide.

It is good you wrote "supposed to be", because they are not. Some time
ago we got here a new laptop on the network, but it did not work because
there was already another machine on the network with the same MAC address.
That one was bought in the US and the other in Europe probably had
something to do with it.

 

[Charlie Gordon]

It is good you wrote "supposed to be", because they are not. Some time
ago we got here a new laptop on the network, but it did not work because
there was already another machine on the network with the same MAC
address.
That one was bought in the US and the other in Europe probably had
something to do with it.

The odds of such a collision be the result of chance are far below those of
winning the Euromillions Grand Prize (130 million EUR and counting)

Did you be any chance install a software on both then reprograms MAC
addresses ?

 

[Joachim Schmitz]

It is good you wrote "supposed to be", because they are not. Some time
ago we got here a new laptop on the network, but it did not work because
there was already another machine on the network with the same MAC
address.
That one was bought in the US and the other in Europe probably had
something to do with it.

That's exactly why said it this way . In my former company there was code
that relied on them to be a) unique and b) unchangable... (using the MAC as
kind of a dongle)

Bye, Jojo

 

[Richard]

That's exactly why said it this way . In my former company there was code
that relied on them to be a) unique and b) unchangable... (using the MAC as
kind of a dongle)

Bye, Jojo

Foolish people still use MAC address as the only line of security on
their wireless LANs. More fool them, MAC address imitation is very easy
and very common.

 

[Joachim Schmitz]

Foolish people still use MAC address as the only line of security on
their wireless LANs. More fool them, MAC address imitation is very easy
and very common.

Well, in defense of my old company I must say that this was 14 years ago,
ethernet wasn't that widely available and WLAN not at all...

 

[Al Balmer]

It is good you wrote "supposed to be", because they are not. Some time
ago we got here a new laptop on the network, but it did not work because
there was already another machine on the network with the same MAC address.
That one was bought in the US and the other in Europe probably had
something to do with it.

It sounds like either someone is cheating, or the MAC address of one
machine was altered. I've never heard of a case where two vendors were
assigned conflicting address blocks.

OTOH, if the programming software doesn't properly check the limits,
it would be quite easy for a vendor to overrun his allotment without
realizing it.

 

[Richard Tobin]

Foolish people still use MAC address as the only line of security on
their wireless LANs. More fool them, MAC address imitation is very easy
and very common.

Some people don't lock their doors. Whether this is a really bad idea
depends where you live, and how valuable your possessions are.

-- Richard

 

[Richard]

Some people don't lock their doors. Whether this is a really bad idea
depends where you live, and how valuable your possessions are.

-- Richard

Big difference though. People using your Wireless could land you in all
sorts of hot water.

 

[Chris Hills]

Big difference though. People using your Wireless could land you in all
sorts of hot water.

I turn off the DHCP, use MAC and TCP/IP address, and the router only
accepts those pairings. The wireless is only on when it is needed.

Most routers let you turn it the wireless off. There is no point having
the office wireless on 18:00-08:00 and the home one on 08:00 -18:00 or
thereabouts.

Naturally we use WPA encryption. Not sure what else I can do. However
your security only has to be stronger than the value of the thing it is
protecting. If ti is random snoopers there are other easier targets and
I don't think were are likely to be the target of industrial espionage.
..

 

[Dik T. Winter]

> "Dik T. Winter" <[email protected]> a écrit dans le message de
> (e-mail address removed)... ....
>
> The odds of such a collision be the result of chance are far below those of
> winning the Euromillions Grand Prize (130 million EUR and counting)
Perhaps.

> Did you be any chance install a software on both then reprograms MAC
> addresses ?

Yes, after we found the collission. Both machines were installed as out
of the box.

 

[Dik T. Winter]

> On Mon, 1 Oct 2007 12:58:58 GMT, "Dik T. Winter" <[email protected]>
> wrote: ....
>
> It sounds like either someone is cheating, or the MAC address of one
> machine was altered. I've never heard of a case where two vendors were
> assigned conflicting address blocks.

I think some of the vendors are cheating. They were cards from the same
vendor. Apparently the vendor thought that a card bought in Europe would
not enter a network in the US where the MAC address was crucial, and the
other way around.

 

[Charlie Gordon]

Yes, after we found the collission. Both machines were installed as out
of the box.

Do you remeber the makes by any chance ?

 

[robertwessel2]

The odds of such a collision be the result of chance are far below those of
winning the Euromillions Grand Prize (130 million EUR and counting)

Did you be any chance install a software on both then reprograms MAC
addresses ?

Actually if the vendors are doing their jobs, the odds of a collision
should be zero with UAAs (universally administered addresses). That
being said, more than a few (low-end) vendors have just picked a range
(not infrequently an old ID assigned to a now defunct vendor) and used
it (saves them the $4K to actually get a properly allocated block), or
have just messed up (I know of one vendor who managed an entire
production run of ~20K Ethernet cards programmed with the *same* MAC
address).

Obviously that doesn't apply to LAAs

There was originally consideration given to the idea of just having
each card pick a random 48-bit MAC, and relying on the low odds of two
NICs on a single LAN picking the same address, but that basically
didn't happen, and the managed approach we have now was selected.

 

[Dik T. Winter]

>
> Do you remeber the makes by any chance ?

No, it is already quite a long time ago.