CVS update: samba/source

Christopher R. Hertel crh at NTS.Umn.EDU
Thu Apr 9 20:17:08 GMT 1998


> I really, really, really, can't, sorry - it's an anal thing 

   ...your dogma is standing in the way of your karma...

> with me (if I see one warning, it's all over - the code
> is broken :-).

That's bad, Jeremy.  You really should see someone about that.  We have
counselors and psychologists here in the Midwest who could help you out. 
I'm sure that there are a few in California.  ;)

On the one hand, I agree 100% that all warnings should be dealt with.  On
the other hand, if it's a *warning*, it's meant to call your attention to
a *potential* problem, not an actual problem.  Dealing with the warning
does not necessarily mean 'making it go away'.  The real problem is the
lack of code to use the ModuleID, but that is not an immediate problem.

There is nothing actually wrong with the code that gcc is complaining
about.  Really.  It's just a piece of static space that is being set aside
and, from gcc's point of view, not used.  The compiler on the SGI seems to
lose it completely (i.e., it's not there when I run strings against smbd). 

> I will check in an #if 0 around the offending
> moduleid if that's ok - with a comment explaining
> why. Then we can remove them as the other stuff
> gets added that uses them. Is that ok ?

No.  Please don't.  I'll check in a copy with the static string removed.  
I have the "real" RCS repository at home so it won't be hard to add it 
back when I've got a version that does not offend.

> Can you send me the warnings list from the SGI
> - as I've said, I can't abide warnings (I get
> none on FreeBSD/gcc-2.8 :-).

Will do, under separate cover.

Chris -)-----

-- 
Christopher R. Hertel -)-----                   University of Minnesota
crh at nts.umn.edu              Networking and Telecommunications Services


More information about the samba-technical mailing list