Feasible google summer of code project?

[Joseph Garvin]

I asked in #winehq where I should ask about this, and I was sent here.
So no flames if they were horribly wrong, I tried ;)

I'm an undergraduate CS student, and I've been reading the wine
newsletter for a while now because I find the project rather cool to
watch. When I started using Linux a year ago wine didn't run my
screenwriting software at all -- it's now fully functional. That's just
flippin' sweet :)

So, naturally, I'm excited at the idea of maybe hacking on wine. I'm not
a super experienced C programmer. I know Standard C but have never
actually worked on a large project in it before. I pick up new languages
rather quickly though, and I've worked in Java, C, C++, Matlab, Lisp,
and Python (my current favorite). But because I'm not a low level C guru
(yet), I looked for things I could contribute to that weren't
frightening -- some of the suggestions on the Wiki summer of code page
look like they're meant for people already familiar with the wine codebase.

That said, I'm interested in picking up the desktop integration areas
that wine currently lacks. Working trash, my documents, autorun, xdg
menu support (adding start menu items to the K-menu for instance),
winebrowser, and icon themes. Basically a good chunk of stuff on the fun
projects page.

Two questions:

1. My understanding is the mentor organizations get to choose who they
want. Is the wine project still interested in someone with my relatively
low C/wine code base experience or should I just look elsewhere?

2. Is this something feasible for a proposal? Could it realistically be
completed over the summer, say with 10 (insert your realistic number
here) hours a week of work? You guys know the code better than I do :)

[Mike Hearn]

On Wed, 01 Jun 2005 19:26:07 +0000, Joseph Garvin wrote:
> - My understanding is the mentor organizations get to choose who they
>
> want. Is the wine project still interested in someone with my relatively
> low C/wine code base experience or should I just look elsewhere?

Well, we're not the borg but yes we're interested in anybody who seems
competent and dedicated. Based on a totally random and mostly
unjustified first impression, I'd say you are!
> - Is this something feasible for a proposal? Could it realistically be
>
> completed over the summer, say with 10 (insert your realistic number
> here) hours a week of work? You guys know the code better than I do :)

You named a lot of stuff there, but yes desktop integration is absolutely
pitched at about the right level. If it helps any, when I started I didn't
even know C, I learned it specifically in order to hack on Wine. So it's
definitely do-able, especially as documentation improved a lot since I
started (well, I would say that, I wrote quite a bit ... :)

In order of difficulty from easy to hard I'd say:

- Linking Desktop/My Documents to the real locations
- XDG menu support
- Autorun (needs the Wine explorer work really)
- Icon themes

Icon themes listed as "quite hard" because directly importing icon themes
from the environment is fraught with non-code problems like visual
consistency. I have a patch to import the GNOME artwork which IMHO makes a
*ton* of difference, it looks much nicer, but Alexandre wants yellow
folders ;) But to do that I had to Gimp the icons up a bit, they weren't
usable raw.

Autorun needs integration with the HAL to monitor for new media events,
really. And that needs a process sitting in the background, which is what
the Wine Explorer thing is supposed to be. I have a patch to add that but
AJ said at WineConf it needed more design work.

XDG menu support is totally do-able, though be prepared for frustration as
you hit random bugs in desktops and distros even if you only target XDG.

Linking Desktop/My Documents to the real locations shouldn't be too hard
... in theory ....

thanks -mike

[Travis Watkins]

On 6/2/05, Mike Hearn  wrote:
[snip]
> - Linking Desktop/My Documents to the real locations
> - XDG menu support
> - Autorun (needs the Wine explorer work really)
> - Icon themes

[snip]
> thanks -mike

I could help out with the xdg menu stuff. Not with coding, but with
working around bugs and such.

[Ivan Leo Puoti]

Mike Hearn wrote:
> If it helps any, when I started I didn't
> even know C, I learned it specifically in order to hack on Wine.

So did I.

Ivan.