CVS on Tru64

[Steven Schweda]

> [...] it exists in 1.11.23 as well, [...]

Not what I observed, but my compilers
probably differ from your (unspecified) one.

> Anyway, in case you hadn't notices, this
> topic has been closed.

I might notice it when it happens, but with
my weak psychic powers, probably not before
then.

[BrianT_1]

I had closed this thread once before. I have no idea how it got opened again. I don't azppreciate the sarcasm.

[Steven Schweda]

> CVS 1.11.23 is what I'm describing now.

Why didn't I guess that?

> I don't azppreciate the sarcasm.

And I'm not very fond of sloppy and
incomplete problem reports, but each of us
has his own cross to bear.

Glad to see that you got this thread closed,
though.

[Dennis Handly]

>but why would the writers of cvs use an illegal construct and expect people not to notice?

You would need to check where MAP_FAILED is defined and with what value.

>Does gcc allow illegal statements detected by the C compiler installed with Digital Unix?

It may give warnings instead of errors?

[Steven Schweda]

> You would need to check [...]

And, unless you reveal which source kit
you're using ("now"), and what the actual
error message is ("now"), it'll be rather
difficult for anyone else to investigate.
And, of course, as previously suggested,
identifying the compiler in use might also be
nice. (I'll just hold my breath.)

[Dennis Handly]

>Steven: unless you reveal which source kit you're using ("now")

That might not be needed if the OS header is "broken".
On HP-UX MAP_FAILED is defined in and has an explicit cast:
#define MAP_FAILED ((void*)-1L)

POSIX requires MAP_FAILED but it isn't obvious if it has to have that cast.