Re: A proposal for Midnight Commander

[Ali Akcaagac <aliakc web de>]

On Sat, 2003-11-15 at 03:30, Alexander Varakin wrote:
> Even easier:
> #define g_malloc  malloc

Indeed, this sounds even better.

> The remaining 30% of mc's glib code usage must be the list code,
> and I think the best way to handle it is just to use glib's
> implementation.

The rest are still depending on some soon to be DEPRECATED functions
which you can verify on your own inside glib 2.3.x. I think we should
have a separate location for this stuff. Arpi already offered me some
space on mplayer.hu (also good for PR). I only like to convince a few
more people for this, so we can finally start doing the work. Let's see
how this looks like. I am not going to promise anything because MC as is
now satisfies me on my normal Workstation. It's just the idea I have
now. Maybe I can convince Arpi (shouldn't be a big prob imo) and that
Olegarch guy (who is working on 4.1.40) to merge together in one
project.

I do like Midnight Commander 4.6.0 (as I often repeated here) it's
nothing wrong only the dependency of glib which shouldn't be there imo.
I saw what Miguel replied to you and usually I have the tendency to
agree that re-using code is a nice thing. But Midnight Commander is not
big enough that it would justify it to depend on glib. For GUI material
e.g. a full Desktop Environment I would fullheartly agree anyday but not
for a console application. What benefits do you have when you are able
to shrink the size of Midnight Commander to around ~350kb but then on
the otherhand need a full blown Glib2 library with ~500kb only to use
the 5-6 functions required by Midnight Commander. I think that everybody
is going to agree here.

Let me know whether you have interest in this (this is valid for
everyone else as well), I then gonna forward this request to Arpi and
have a new ML set up for this.

greetings.