Re: ANN: PikoPixel pixel-art editor

[Fred Kiefer]

Hi Josh,

thank you for your detailed reply. It helped me understand the issue a
lot better.

Am 20.09.2015 um 00:06 schrieb Josh Freeman:
> On Sep 19, 2015, at 7:20 AM, Fred Kiefer wrote:
>> thank you very much for porting this great code over to GNUstep.
>>
>> I am at the moment trying to get it running on gcc with the gcc objc
>> runtime. I know this isn't the supported constellation, but I would like
>> to at least understand what features are missing in this setup.
> 
>    PikoPixel's implementation of method-swizzling doesn't work with the
> gcc runtime because the runtime's function, class_replaceMethod(),
> behaves differently than on the objc2 or Apple runtimes. That's the only
> issue I know of (though I didn't test much further on gcc after finding
> it), so in theory, a reimplementation of method-swizzling (in
> NSObject_PPUtilities_MethodSwizzling.m) that removed/reworked the
> class_replaceMethod() calls could allow PP to use the gcc runtime.

I already had figured out from the messages on the terminal that method
swizzling was causing the problems, but didn't understand why. With your
hint and a bit of googling (http://nshipster.com/method-swizzling/) I
ended up with a small change to your swizzling method. All I had to do
was to replace the two calls to class_replaceMethod() with the following:

   BOOL didAddMethod = class_addMethod(self, selector1, implementation2,
                                        typeEncoding2);

    if (didAddMethod)
      {
        class_replaceMethod(self, selector2, implementation1,
                            typeEncoding1);
      }
    else
      {
        method_exchangeImplementations(method1, method2);
      }

Strange enough this only works for the two instance method swizzling
methods, for the class methods I have to use just

method_exchangeImplementations(method1, method2);

This seems to work with the gcc runtime. I really don't understand what
I am doing here and why it should be any different.
Now PikoPixel starts up and works fine for me. Of course I also removed
all the swizzling for methods I already changed in GNUstep and I had to
fiddle with the PPBuildEnvironmentMacros.h file to enable swizzling.

>> I parallel I am digging through your GNUstep specific changes. I already
>> found the NSMatrix additions and the cairo non-anti-alias stuff. I am
>> going to add code into GNustep that supports this to make specific
>> additions for GNUstep unnecessary. In theory it should be possible to
>> just reuse your Cocoa code on GNUstep. Where this isn't possible, please
>> file a bug report against GNUstep, preferable with an attached patch.
> 
>     I kept an internal list with brief, 1-to-2-line descriptions of
> issues (now at about 17 items) that needs to be expanded into full bug
> reports on savannah when there's time - I wrote one up so far:
> http://savannah.gnu.org/bugs/?45442 . I'd be happy to send the raw list
> to you - or post it to the developer mailing list - if that would be
> useful.

Just send me that internal list, together with your code changes, this
should
help me figure out what went wrong in GNustep.

As for your bug report, I am sorry to never have replied to it. Window
ordering is one of the weak spots in GNUstep. Previously I tried to
figure out suitable values for window properties for KDE and Gnome for
each new release. But I gave up on that a long time ago. Now I just take
care that GNUstep stays usable on my own KDE environment. If you are
interested in the problem please take a look at the method
setwindowlevel:: in back/Source/x11/XGServerWindow.m.

>> Once again thank you and I will keep you updated on my progress.
> 
>    Thanks, please do!

Done :-)

Fred