NAN option for gcc?

[jchludzinski]

Alright I give up! I thought that to only option I need for NAN,
nan(), nanf(), etc. is -l (after all you use #include <math.h>). But
NO ...

It works with g++. What -l option is required?

Thanks ahead of time,
John

PS> Yes, I've grepped the (infinite) man for gcc.

[jchludzinski]

Correction:

"... I thought that the only option I need for NAN,
nan(), nanf(), etc. is -lm (after all you use #include <math.h>)."

[James Hess]

jc**********@gmail.com wrote:

Alright I give up! I thought that to only option I need for NAN,
nan(), nanf(), etc. is -l (after all you use #include <math.h>). But
NO ...

It works with g++. What -l option is required?

I don't think nan() or nanf() is standard

maybe you'd have better luck with something like
atod("NaN")

?
-Mysid

[jchludzinski]

Actually nanf("") works; although nanf(NULL) gives a seg. fault. BUT
both compile using gcc -lm, whereas NAN does not. But NAN works with
g++ (no option). What am I getting by using g++ that I'm not getting
with gcc -lm?

---John

[Ben Pfaff]

"James Hess" <my*****@gmail.com> writes:

jc**********@gmail.com wrote:

Alright I give up! I thought that to only option I need for NAN,
nan(), nanf(), etc. is -l (after all you use #include <math.h>). But
NO ...

It works with g++. What -l option is required?

I don't think nan() or nanf() is standard

They're both in C99.
--
"Debugging is twice as hard as writing the code in the first place.
Therefore, if you write the code as cleverly as possible, you are,
by definition, not smart enough to debug it."
--Brian Kernighan

[Michael Mair]

jc**********@gmail.com wrote:

Alright I give up! I thought that to only option I need for NAN,
nan(), nanf(), etc. is -l (after all you use #include <math.h>). But
NO ...

It works with g++. What -l option is required?

Thanks ahead of time,
John

PS> Yes, I've grepped the (infinite) man for gcc.

Did you use C99 or gnu99 standard (-std=c99/gnu99, of course together
with -lm)?
AFAIR, these are no C89 standard library functions.
This may make a difference w.r.t. what is available from the standard
library (apart from the usual gcc argument that the library is not
their part).
Note that this is not entirely topical in comp.lang.c.
gcc specific questions are usually better asked in gcc.gnu.help.
Cheers
Michael
--
E-Mail: Mine is an /at/ gmx /dot/ de address.

[jchludzinski]

Yep, that worked: gcc -lm -std=c99. Thanks much!

---John

PS> Next time I'll direct gcc questions to gcc.gnu.help.

[Michael Mair]

jc**********@gmail.com wrote:

Yep, that worked: gcc -lm -std=c99. Thanks much!
You are welcome. Just as an aside: -lm belongs _after_ the compile
functions, see
<m3************@amoeba.parasoft.com>
(This is a message ID, you can use it to find the message, e.g.
with groups.google.com)

---John

PS> Next time I'll direct gcc questions to gcc.gnu.help.

Okay. One thing: Please quote a minimum of the message you are
referring to so that everyone sees the context. Otherwise, people
may see only your reply for some hours or days or forever without
having the least clue what you are responding to. See CB Falconer's
signature if you are using google for your newsgroup participation.
Cheers
Michael
--
E-Mail: Mine is an /at/ gmx /dot/ de address.