httpd-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Ben Laurie <...@algroup.co.uk>
Subject Re: cvs commit: apache-2.0/src/lib/apr/time/unix access.c
Date Sat, 02 Oct 1999 14:21:43 GMT
"Ralf S. Engelschall" wrote:
> 
> In article <37F51166.9C3FB5B1@algroup.co.uk> you wrote:
> > rse@hyperreal.org wrote:
> >>
> >> rse         99/10/01 09:18:41
> >>
> >>   Modified:    src/lib/apr/file_io/unix dir.c fileacc.c
> >>                src/lib/apr/include apr_general.h
> >>                src/lib/apr/lib apr_cpystrn.c apr_md5.c apr_pools.c
> >>                         apr_slack.c
> >>                src/lib/apr/locks/unix crossproc.c
> >>                src/lib/apr/network_io/unix networkio.h poll.c sockets.c
> >>                         sockopt.c
> >>                src/lib/apr/threadproc/unix proc.c signals.c
> >>                src/lib/apr/time/unix access.c
> >>   Log:
> >>   Ok, today it bored me too much that APR shouts many pages with warnings under
> >>   --with-option=devel (which has maximum -W's enabled which are still practical
> >>   for Apache), so I investiged an hour and did a quick cleanup of the APR
> >>   sources.
> >>
> >>   Please guys, if you're using gcc, please _always_ at least use -Wall for
> >>   compiling Apache. You will be surprised how much mistakes it catches easily.
I
> >>   don't want to say you have to use the maximum -W approach I personally prefer,
> >>   but Apache's sources should always at least pass -Wall silently, I think.
> >>   Thanks.
> >
> > Arg. Fixing it with millions of evil casts is not better than the
> > warnings were! Also, why the pointless variable name changes (which seem
> > to be the bulk of this patch)?
> 
> Ah, there it is - I expected your complain, Ben.
> 
> 1. I looked carefully at the type-mismatch problem, doubt me. But the only
>    non-cast solution IMHO would be to change parts of the API and the change
>    would here leaded to a "password" to be "unsigned char *", so I decided
>    that it's better to not do this and keep the "eval casts" you like so much.

Grrk. This signed/unsigned char stuff really irritates me! We should use
C++.

> 2. The renaming of the variables is necessary because variables
>    like time conflict with the system time(3) prototype, etc.  Same for things
>    like stat which conflict with stat(2).

Whoops! I retract my criticism, in that case. Sorry.

Cheers,

Ben.

--
http://www.apache-ssl.org/ben.html

"My grandfather once told me that there are two kinds of people: those
who work and those who take the credit. He told me to try to be in the
first group; there was less competition there."
     - Indira Gandhi

Mime
View raw message