Re: acl for cvs, access from internal machines

[Mark D. Baushke]

Edward Peschko <address@hidden> writes:

> Well, it looks like my patch for acl grafted onto CVS has 
> pretty much sunk without a trace... wrong forum? philosophical 
> issues again? no interest?

It is the right forum, (I actually have your message in a queue of
things to look at right behind a patch by Dieter Mauer that deals with
keyword substitution, but which turned up an odd problem that I have
been trying to track down in that his keyword2-20k test does something
odd).

I do have a few issues with your patch in that you are not presently
following the coding standards used by the rest of cvs.

For example, the functions in acl.c do not follow the coding standards
for how functions are defined. 

    int
    acl_quick_check_split (dir, file, op)
        char *dir;
        char *file;
        char *op;
    {
    ...
    }

We are NOT YET at the point where we have moved to using full ANSI C
specifications as we are still conformant to being used by K&R C
compilers. It has been discussed, but it is not yet adopted.

Additionally, the fact that the _printit function is an illegal name
(functions with leading underscores are reserved). The intent is to have
source code that can get through all of the compilers for the platforms
that presently 'support' cvs on them... the entire world does not yet
use an ANSI C compiler.

Other formatting problems include calls to functions like xmalloc have
the space before the opening parenthesis and not after it. For example:

      /* some comment about the magic number 4 being used on the next line */
      fullpath = xmalloc (strlen(dir) + strlen(file) + 4);

(your comment is missing in at least one place).

Many function calls are missing the space before the parenthesis entirely.

Some lines are a lot longer than eighty columns without a real need for
them to be.

You also have not provided ChangeLog entires that describe what has
changed.

> If it is the wrong forum for patches, then what is the *right* forum? 

It is a fine forum to float ideas. Once the ideas have received support
we typically work with you to refine the code to be closer what is in
the rest of cvs for coding style.

Note that some other patches are floated on the address@hidden mailing
list. I have no problems with the patches appearing in either place.

> If its philosophical issues again, then what are they and how can 
> we overcome them? As for interest, well I'm using the acl patch right 
> now, and am finding it damn useful, if I do say so myself..
> :->

I don't know about the rest of the team, but I have lots of things that
are happening, so I just have not had time to really look at the patch.
To be honest, I'll be on holiday for a few weeks in July, so I probably
will not be getting to even looking closely at your patch until August
(I just do this cvs work as a hobby).

> Anyways, I figured a way around the internal machine dilemma, I think.
> If I subscribe from the internal machine, it seems to take precedence
> over the spam filter, ie: I can post from inside without worrying
> about the domain check.

That is good news indeed.

        Enjoy!
        -- Mark