-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1
On 03/16/2007 07:27 AM, Fabio Busatto wrote:
> On Thu, Mar 15, 2007 at 05:22:19PM -0400, Russ Nelson wrote:
>> Fabio Busatto writes:
>> > Otherwise, how do you suggest to make various new features
>> > available on-demand at compile time?
>>
>> The exact method is dependent upon the exact code, but most generally,
>> you provide a library of functions, some of which have the exact same
>> API, and then you link to one or the other depending on which feature
>> you want. Think "object-oriented programming" only done efficiently
>> at compile-time (or, more properly, link time).
>
> We're not discussing on software development, but on qmail evolution
> using a patch system that lets the user to decide if a patch must be
> included in the final binary.
Fabio,
Nothing against your project, do not get me wrong, I'm
even writing as a person that already tried to build a qmail
patch system (qmail-patch-o-magic, a.k.a. qpom).
> Your answer is not concrete, and maybe I'm too limited in
> programming concepts but I cannot understand how it could
> be a solution for our problem.
There is no magic solution, if you have the time and
the desire to do such a system, go ahead. If need improvement
and you allow people to change it, you will get patches. :-)
At least suggestions and comments.
[...]
>> #ifdefs are the bane of a programmer's existance. A professional
>> won't use them.
>
> Sure? Are you really sure? You're saying that about every
> software (including djb's one and the main os kernels and
> are written by novices.
It really matters if professionals will use them or
not? Just use what you have in mind, show a prototype, people
can criticize and say "that's the worst crap that we have
ever saw in our lives", or "hmm, that's cool!", either way,
start writing it, can show people what you have in mind and
would allow people with less time but same interest to help
improve it.
In the worst case, you can start thinking about a
new version and take +10 years to prepare it. :-)
> I can accept the sentence "I don't like preprocessor
> directives", but not "A professional won't use them".
You don't need to accept it, just move on. Show
the list your progress, from what I've seen so far,
a lot of people already support your idea, if you choose
the wrong direction, it wouldn't be the first time in
history of software development (and certainly it is not
going to be the last one).
> I seldom use #ifdef when I create a new software, but
> this is another case, we're not creating a new software,
> we are creating a patching system for an existing project.
So, go ahead, do it and show everybody that it
can work, people will use it or not, but don't let the
idea die, having a patch system in the same lines of
iptables patch-o-matic would be really cool.
[...]
> Maybe I just misunderstood your approach, but I want to be
> really sure of my choices before starting.
[...]
In fact, you don't. Just code it and let people
see what's going on ("Release Often, Release Early"),
how they can help, if you already design something, go
and code it and come back. :-)
Kind regards,
- --
Felipe Augusto van de Wiel <felipe@paranacidade.org.br>
Coordenadoria de Tecnologia da Informação (CTI) - SEDU/PARANACIDADE
http://www.paranacidade.org.br/ Phone: (+55 41 3350 3300)
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.6 (GNU/Linux)
Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org
iD8DBQFF+qY5Cj65ZxU4gPQRArvqAKCEtUPoeSaezondJCjcLq12axL0mwCePZJ/
jpFEO5MifFa1LeunaWLFDSI=
=wWxm
-----END PGP SIGNATURE-----
|