httpd-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Roy T. Fielding" <field...@liege.ICS.UCI.EDU>
Subject Re: cvs commit: apache/src conf.h
Date Wed, 14 Aug 1996 23:27:37 GMT
> Hmm, we've been through all this before and I'm still pretty convinced
> that gethostname is not something we should use on Solaris 2. The
> reason there's no prototype for it (Solaris 2 does have full
> prototyping) is because it's part of the BSD compatibility library.

Not in SunOS 5.5 -- it is part of the normal library.  The prototype
is the same, though, which is why it is safe to add it for any Solaris
machine (at least until we stop using gethostname).  This is not related
to the SunOS prototype issue which fails due to prototype conflicts
in the different libraries.

> -----
>      Use of these interfaces should be restricted to only  appli-
>      cations  written  on BSD platforms.  Use of these interfaces
>      with any of the system libraries or in multi-thread applica-
>      tions is unsupported.
> -----

As discussed before, that warning does not appear in SunOS 5.5.

> You've hidden the warning now and this may well bite us when we have a
> threaded server. My OS abstraction example explicity used hostname
> resolution as an example, it should use uname were possible (more
> widely available) and gethostname if not.

My experience has been that uname is completely unreliable -- frequently
giving just the hostname (not fully-qualified) or truncating the FQDN
at some arbitrary limit.  Changing this will require extensive testing.

> On of the things I'd like to get cracking on but 1.2 is effectively in
> feature freeze indefinately.

Que?  So far, the non-HTTP related feature creep on 1.2 is about 10 times
that of the HTTP changes. That is hardly what I would call a feature freeze.
The only thing we haven't added is multithreading, and I expect to see
that patch any day now. ;-)


View raw message