httpd-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Dean Gaudet <dgaudet-list-new-ht...@arctic.org>
Subject Re: returning statuses in apr
Date Sun, 12 Mar 2000 17:12:30 GMT
i think the problem, in general, is that we know of N platforms at 2.0
release and none of them require a return status, and then someone comes
along with N+1 which requires a return status and we have to change the
API.

not that i care too much, 'cause i only care about unix ;)

(the c++ weenies are probably thinking "exceptions are much nicer than
result codes" ;)

Dean

On Mon, 6 Mar 2000, David Reid wrote:

> To be honest it makes sense to me as well.  However, we also need to include
> the proviso that if so much as a single supported platform implements the
> function in such a way that calling it may return an error code then it
> should be returning an ap_status_t on all platforms.
> 
> d.
> ----- Original Message -----
> From: "Greg Stein" <gstein@lyra.org>
> To: <new-httpd@apache.org>
> Sent: Monday, March 06, 2000 11:11 PM
> Subject: Re: returning statuses in apr
> 
> 
> > Let's modify the Design Statement to be:
> >
> >   "An APR function may return ap_status_t or void. Nothing else. Return
> >    values should be handled through [out] parameters."
> >
> >
> > Cheers,
> > -g
> >
> > On Sun, 5 Mar 2000 rbb@apache.org wrote:
> > >
> > > Every discussion we had about status codes said EVERY function returns a
> > > status code.  We could remove this requirement, but I would rather not.
> > > It makes it very easy for people using APR functions to ALWAYS know the
> > > basic setup for an apr function.
> > >
> > > Ryan
> > >
> > > On Sun, 5 Mar 2000, Life is hard, and then you die. wrote:
> > >
> > > >
> > > > Looking at lib/apr/lib/apr_md5.c I have a question:
> > > >
> > > > Why are the functions returning ap_status_t when nothing can go
> > > > wrong? (i.e. all they ever return is APR_SUCCESS). Any reason
> > > > not to revert back to returning void? I vaguely remember various
> > > > discussions around return values, but I don't recall (and couldn't
> > > > find) anything about having *every* function return a status.
> > > >
> > > > The cvs log entry where this happened reads
> > > >
> > > >   Begin to update apr/lib files to follow the standard APR function
> format.
> > > >   This is the incredibly low hanging fruit, where all I am doing is
> > > >   standardizing the return codes for functions that already returned
> int's.
> > > >
> > > > but the functions returned void before... (with the exception of
> > > > ap_validate_password(), but that's not what I'm talking about).
> > > >
> > > >
> > > >   Cheers,
> > > >
> > > >   Ronald
> > > >
> > > >
> > >
> > >
> > > Come to the first official Apache Software Foundation
> > > Conference!!!   <http://ApacheCon.Com/>
> > >
> > >
> ____________________________________________________________________________
> ___
> > > Ryan Bloom                        rbb@apache.org
> > > 406 29th St.
> > > San Francisco, CA 94131
> >
> > --------------------------------------------------------------------------
> -----
> > >
> >
> > --
> > Greg Stein, http://www.lyra.org/
> >
> >
> 
> 


Mime
View raw message