send an email to -dev about perhaps making it official, they could give it to the arch testers for testingmiga wrote:Ah great!
Just started a basic GUI in JAVA (http://img374.imageshack.us/img374/7270/jebuildej1.jpg)![]()
Not doing anything yet, just the GUI.
At the moment I'm trying to add a TabView for unpack, compile, install but I think I can stop it (puhhh) and have a look at abeni!

Sounds like a job for the UserRep initiative!cokehabit wrote:it sounds like a good idea though, it needs to be official though
I agree completely.mikegpitt wrote:I took a look at your gui, and I fear that it is simplifying the ebuild creation process a lot. If you look at some of the ebuilds the build process is fairly complex. The gui would need to support some way to edit the code of the ebuild too. Basically I think what you are aiming to create is an IDE for ebuilds. The initial selection of USE flags, source URI, etc, could be done in a "wizard" phase, and then from then on you could edit the contents of the ebuild.
I'm not sure if Java is the best lang to do this in to make it official, but I may be wrong. I'm just basing this statement on what gentoo tools are official.
Sounds like a good application to build upon.dol-sen wrote:There was one called abeni created by pythonhead, but it is no longer supported or functioning. He was waiting for wx windows2 I think.


Dizzutch wrote:I think the best for this project would be a "simple" GUI that can make basic ebuilds, to make it more accessible for users to roll their own ebuilds.
Code: Select all
# Copyright 1999-2006 Gentoo Foundation
# Distributed under the terms of the GNU General Public License v2
# $Header: $
DESCRIPTION=""
HOMEPAGE=""
SRC_URI=""
LICENSE=""
SLOT="0"
KEYWORDS="~amd64"
IUSE=""
DEPEND=""
RDEPEND=""

A complete rubbish argument.welp wrote:I do not think that an ebuild GUI would be a good idea. IMHO, people who write ebuilds should have at least *some* idea as to what does what in an ebuild and how an ebuild works, otherwise there's the chance that we're gonna have people going around complaining that it's not working... i'd hate to see the forums and #gentoo filled with people screaming out "help me! help me! my ebuild for foo didn't compile" and shit like that...

That might be difficult, but I don't think it would be a unovercomeble task. Remember, we don't live in 1995 anymore and we can make such a GUI kinda intelligent in using USE flags. It could for example guestimate the USE flags running a ./config or something.omp wrote:Would be too complicated. Think of all the ways JUST dependencies could be handled.
http://devmanual.gentoo.org/general-con ... index.html
Explain to me how you'll be able to fit all of that in a simple and easy-to-use GUI wizard.

And can, quite possibly, miss or choose incorrect USE flags. And the user might be an idiot who doesn't go over it. That would then result in something along the lines of everyone complaining "My dog just made an ebuild and it doesn't work!! Oh noes!!!11111cos(0)".Q-collective wrote:It could for example guestimate the USE flags running a ./config or something.

See my answer to welp, I hate to repeat myself.omp wrote:And can, quite possibly, miss or choose incorrect USE flags. And the user might be an idiot who doesn't go over it. That would then result in something along the lines of everyone complaining "My dog just made an ebuild and it doesn't work!! Oh noes!!!11111cos(0)".Q-collective wrote:It could for example guestimate the USE flags running a ./config or something.
Other than the great number of possible flaws, we don't want everyone and their dogs making crap ebuilds. Not to mention this is a lot of development power going to waste.

Fine; see below.Q-collective wrote:See my answer to welp, I hate to repeat myself.omp wrote:And can, quite possibly, miss or choose incorrect USE flags. And the user might be an idiot who doesn't go over it. That would then result in something along the lines of everyone complaining "My dog just made an ebuild and it doesn't work!! Oh noes!!!11111cos(0)".Q-collective wrote:It could for example guestimate the USE flags running a ./config or something.
Other than the great number of possible flaws, we don't want everyone and their dogs making crap ebuilds. Not to mention this is a lot of development power going to waste.
I doubt generated ebuilds would be anywhere near those created by great devs in terms of quality. Don't forget that a lot of ebuilds need extra/complex stuff, and not basic default methods, so the generator won't be able to do those. (Don't argue that we'll be able to create this, because we won't. At least not anytime soon.)Q-collective wrote:This overal has a positive effect on increasing the portage tree and quality of the ebuilds (since they are generated, not manually written... which is another argument against your argumentation).

Generators can create very complex yet very correct stuff, see DreamWeaver for example that can create very complex yet perfectly coded websites(and don't say DreamWeaver is bad just because you don't like it, because millions of others do).omp wrote:I doubt generated ebuilds would be anywhere near those created by great devs in terms of quality. Don't forget that a lot of ebuilds need extra/complex stuff, and not basic default methods, so the generator won't be able to do those. (Don't argue that we'll be able to create this, because we won't. At least not anytime soon.)Q-collective wrote:This overal has a positive effect on increasing the portage tree and quality of the ebuilds (since they are generated, not manually written... which is another argument against your argumentation).

Speaking of apps such as DreamWeaver, don't forget that a lot is up to the user, and therefore, with valid code, a site can still be complete crap. The same can apply to an ebuild.Q-collective wrote:Generators can create very complex yet very correct stuff, see DreamWeaver for example that can create very complex yet perfectly coded websites(and don't say DreamWeaver is bad just because you don't like it, because millions of others do).

omp wrote:Speaking of apps such as DreamWeaver, don't forget that a lot is up to the user, and therefore, with valid code, a site can still be complete crap. The same can apply to an ebuild.Q-collective wrote:Generators can create very complex yet very correct stuff, see DreamWeaver for example that can create very complex yet perfectly coded websites(and don't say DreamWeaver is bad just because you don't like it, because millions of others do).
This is going circles.Q-collective wrote:There are three filtering mechanisms to prevent stuff like that: Bugzilla (on which ebuilds have to be submitted to be included with official portage), Sunrise (unofficial user community portage tree, which is not supported in any way, use at your own risk and by the time you know how to use it you know about that) and the ebuild quiz's/mentoring period (to make you an official dev, which requires real knowledge).
I'm pretty sure I read a magazine which built a website with Dreamweaver, then went in to take a look at the code and found that a whole load of tags were... wrong... out-of-date, whatever, IMO, that's not a perfectly coded website.Q-collective wrote:Generators can create very complex yet very correct stuff, see DreamWeaver for example that can create very complex yet perfectly coded websites