wxWidgets bounty - please share your thoughts!

Hello folks,

As some of you may already have seen, there is the thread "Necessary projects: (e.g. WxWidgets)" going which resulted in a suggested bounty for a port of wxWidgets. Troels and myself agree that this would really be worth a bounty as a proper port of wxWidgets would open us for a lot of fresh software instantly or with far less porting effort.

Just to name a few wxWidget users:

- AudioEvolution 5 (until v4 also on AOS, now driven by wxWidgets and so only avaibale for 'the others')
- VLC (wxWidgets would us instantly give the GUI. Of course, there still would be enough to be ported left ;) )
- Code::Blocks (an IDE)
- etc.. see Google or the wxWidgets page for more


Now we want to ask you for your thoughts about how this bounty should actually be shaped. For a successfull bounty, there have to be a good description of the requirements to fullfill the bounty, so please think about these basic questions and share your thoughts, thank you:

- What is the goal of the bounty?
I don't mean something like "a port of wxWidgets". We know that already ;) But more precisely, what shall this port be able to do to fullfill the bounty? wxWidgets consists out of many smaller specialized parts.. not only GUI stuff, but also networking, audio and whatnot. While there are many applications which just use wxBase, there are also many which uses more than one or even all featured modules of wxWidets. So we need a well defined goal. Maybe something like "the AOS port of wxWidgets must be usable to compile a working copy of <HelloWorld>" or something.

Of course, as a complete wxWidgets port is really much work, we shouldn't place a "killer-app" in place of <HelloWorld>, but something realistic.. which also leads to the next even more technical question:

- Which modules of wxWidgets have to be ported?
See above. Just a bit more refined here.

- Which timeframe should be given?
I personally think it's important to define a timeframe for a bounty and to cancel the bounty and refund the donations if the timeframe can't be hold (one or another day more shouldn't be a problem of course). Otherwise a bounty could run for years or ad infinitum which wouldn't help any of us.

- Timeframe and requirements for intermediate milestones?
I also think such a fairly big project has to be divided into milestones, which maybe can even be released as intermediate versions until the full bounty is done. This is of course only to answer after the overall goal and timeframe is set..

- Would it be a good idea to dived the bounty further in even more than one bounty?
As said, wxWidgets is huge.. so maybe it would make sense to think about dividing wxWidgets into several bounties. E.g.
wxBounty #1: Port of wxBase + wXwhatnot, app X has to be compilable with the port and actually work.
wxBounty #2: Port of wxAudio, app Y has to ...

You got the point..
I'd even say this question is highly related to the goals of this bounty.. and i wouldn't do 10 different bounties right now, but one bounty after each other for wxWidgets.. first get all the basic stuff, then open a new bounty for other parts (audio, networking, whatever).




Ok, thats all for now .. i surely forgot one or another important question to ask, so if you think about it, tell it here too

@Cyborg

That sounds good. Please note that the wxWidgets-aos project on sourceforge is targeted at all Amiga OS variants, and the intention is to keep it that way. Also, distributing the bounties to all contributors may be a bit of a challenge since this project will likely require several people to work on it.

Hans

@Hans

I guess the port should be divided in smaller parts like wxBase (IIRC and if the code did not changed too much since then this part "only" need merging with latest code version).
Then it would be a good idea to find relatively small application that only depends on few (2 or 3 max) wxWidget subsets and create a bounty to port those parts. Check would be that the considered application should compile and work.
The more time will be passing the more parts would already be ported so check applications would be able to be a bit bigger too (because even if dependencies are big, only a few of them remains to port)...

@Cyborg

My motivation for wXwidgets is for CodeBlocks and KiCad which is a PCB design program.
http://www.lis.inpg.fr/realise_au_lis/kicad/
http://kicad.sourceforge.net/wiki/index.php/Main_Page

I'm trying to find out what portions of WxWidgets it uses.

The austhor of KiCad tells me this:
..uses mainly wxCore (wxBase + obviously the graphic layer) , wxGL
(opengl layer) and some of wxNet (wxSocket client and server)
and wxSahshWindows from wxAdvanced

@abalaban

Quote:

IIRC, wxWidgets has a test-suite that could be used to confirm whether various components work. WxBase had the file access stuff mostly done. I made a few fixes foe the OS4 version, but it should be changed to use 64-bit disk I/O. However, wxWidgets supports unicode in some shape or form, and I don't think that anything was done in that area (kind of hard when the OS' own unicode support isn't up to scratch).

Hans

@Hans

Yes using the provided test suite might be a good idea . I did not looked at it, I just hope it's not based upon a scripting language we don't have a port yet (say TCL for example)...

About the 64bits DOS functions that's not a big deal to convert I can even do it if you wish. Now for the Unicode support maybe we can use the codesets.library ?

@Cyborg

Hi, I use the following parts / dll's in AE5:

wxbase28d_vc_custom.dll
wxbase28d_xml_vc_custom.dll
wxmsw28d_adv_vc_custom.dll
wxmsw28d_core_vc_custom.dll

I use GUI, gfx, xml, I/O and perhaps things I forgot!

I would skip wxAudio/wxMedia, since that's hardly used.

Good luck with the project! Unfortunately I don't have the time to help out.

This might breathe new life into the wxMozilla project as well as bring Mozilla to the Amiga. At least I'd like to see Mozilla on the Amiga.

see: http://wxmozilla.sourceforge.net/

@ferrels

Everyone would like Mozilla but I don't think that would be the main point of a wxWidgets port.
wxMozilla (which I never knew existed!) also seems pretty dead to me, so even if we got wxWidgets would a port be feasible??

@Troels

donated :)

BTW the paypal receipt still says something about Allegro (was the same for gnash) ;)