Compiling libpurple in Visual Studio

[Nick Hebner]

On 5/28/07, Richard Laager <rlaager at wiktel.com> wrote:
>
> On Mon, 2007-05-28 at 02:57 -0700, Nick Hebner wrote:
> > I have mtn pull'd and up'd (I'm new to mtn, but that should do it
> > right?), but there do not seem any differences.
>
> The patches look different, so presumably you're seeing differences in
> the underlying code... I don't follow.


I copied the libpurple files from where I checked them out from mtn into a
VS project, and modified them before updating. Then when I updated mtn, and
did a diff with my old copied libpurple files, and the only differences were
the ones that I had made. Maybe I did the update wrong...

> You are right about dbus, I am now just including dbus-maybe.h, and
> > not compiling the rest. If the xmlnode struct init syntax has been
> > fixed, then there are just a hand full of others that should take care
> > of it (cleaned up patches attached).
>
> I'm curious why you're removing WINAPI and WSAAPI. I have no idea what
> they do, as I'm not a win32 coder. Can you elaborate on this a bit?


 Honestly, I'm not sure either. I was getting errors from those lines, so I
tried removing them, and that worked. I don't know if that will affect the
MinGW compile... I will look more into this. I know that these specify
calling convention: WINAPI = __stdcall and WSAAPI = FAR PASCAL (which is
defined to be __stdcall for me as well).

Also, can you elaborate on the situation with S_ISDIR a bit more?


S_ISDIR is simply not defined in any windows headers, so you must check this
bit manually.

> Do you think it would be a good idea to add a VS solution to the
> > distribution?
>
> I don't really have an opinion either way, right now. You could send a
> separate patch that adds them and then we'll be able to see exactly what
> the impact would be.
>
> Richard
>
>
Nick
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://pidgin.im/pipermail/devel/attachments/20070528/197d11d6/attachment.html>