FlixelCommunity / flixel

Community fork of Adam “Atomic” Saltsman's popular game engine Flixel. Distilled from a variety of Flash games he worked on over the last couple years, including Gravity Hook, Fathom and Canabalt, its primary function is to provide some useful base classes that you can extend to make your own game objects.
http://flixelcommunity.org/
Other
84 stars 17 forks source link

Cannot get to compile against Genome2d SWC #227

Open greysondn opened 9 years ago

greysondn commented 9 years ago

There is just an "error:null" immediately after loading the configuration file if I try to compile. If I remove the Genome2d SWC from my library, things go back to "normal" - the endless list of errors I get, including ones from Genome2d missing.

I am going to try updating FlashDevelop but I don't think that's the issue. I'm fairly up to date.

Edit: This is against the most current dev mark.

Dovyski commented 9 years ago

I experienced that problem with FlashDevelop after it was upgraded to 4.7. I solved the problem by using the "Clean project" command before compiling anything. I don't think that's a Flixel issue, but a FlashDevelop one.

What version of Genome2d are you using?

greysondn commented 9 years ago

I believe it was the current from here:

https://github.com/pshtif/Genome2D/tree/master/bin

Is there a specific one that's meant to be built against?

Dovyski commented 9 years ago

The new GPU render has been tested with Genome2D 1.0.274. After release 1.0.279, the API has changed, so Flixel Community should build just fine against it, but it will show a black screen in GPU mode. There is a pull request with a fix to make everything work with 1.0.279.

The "error:null" problem seems to be a SDK/FlashDevelop issue though.

greysondn commented 9 years ago

Two things that may be involved that I've found, still chipping away at this as I work on everything else.

First, the FPT test suite carries a copy of Flixel 2.55 and of 2.56 in its default configuration, which may have been a piece of the puzzle. So I've removed both of those so I can work with it against the current marker.

Second, this may be an upstream error but if you say that there is a version that is known to be good then I should test against that before submitting it upstream.

http://stackoverflow.com/questions/6483692/flashdevelop-halted-with-no-explicit-compile-error

If nothing else, this error is teaching me just how dependent we become on compiler vomit to clue us to what's going on. I'm afraid with no error I'm at a bit of a loss. If there's a way to force it to tell me some gist of what's actually wrong here, it'd be useful.

The IDE and SDK are both up to date. I've tried cleaning. Obsessively as I move things around and search the source trees for possible answers.

Dovyski commented 9 years ago

I've just merged the pull request I mentioned. The dev branch should now work with Genome2D 1.0.279.

About the FPT compilation, I've tried a different approach. I cloned the FlixelSandbox project and moved the FPT folders src and lib into it. The FlixelSandbox project seems to compile and inform compilation erros correctly.

You could try that and see how it goes.

greysondn commented 9 years ago

No luck, though I think it's just me struggling with something that actually is simple here given that flixel itself is precompiled in that project.

I need to start on my own work. Therefore I have decided to work from the last commit I can get to compile. Ironically, this happens to be 7232E06F

If there is anything I am able to do that may help you diagnose this error, please let me know. Naturally, the FPT will be against that commit. The good news is that the lib-dev folder is marked at ten months ago (as I write this) for last change (in FlixelSandbox) and the indicated commit is approximately seven months - in theory newer commit than even what I was pointed at? The bad news is that this is still the early side of the GPU render changes.

I imagine you're as frustrated as I am.

Dovyski commented 9 years ago

Go ahead and work against 7232E06F. Even though it is behind the GPU render, it should be straightforward to apply all GPU-related changes later, since there is minor API changes.

Regarding your problem, could you please ZIP your current working dir (including the FlashDevelop project file) and host it anywhere I can download? From there I can replicate your dev environment and dig for the problem.

greysondn commented 9 years ago

Good (bad?) news! I can compile against that SWC now. I think my current problem is related to trying to forcedebugger but once that gets cleaned up and I can confirm that compilation runs against this fine for latest dev, this can be closed.

This is in anticipation on trying to get some minor work on flixel itself done.

Edit: Genome2D_1_0_279.swc is the one referenced here and the one build works against. I'm sure dev already knows but just in case someone comes across this and wonders.

greysondn commented 9 years ago

Followup - despite the fact that this works, no objects seem to be drawn to camera in more complex situations.

I use factories to simplify creation. I'm befuddled.

I'm open to zipping the project up and sending it to you privately to analyze. We've moved to "won't compile" to "utterly WTF", which is an improvement but still not a useful state :disappointed:

Dovyski commented 9 years ago

I've seen that pattern already, but I cannot do much without analyzing the code. If you are ok sending it to me privately, that would be great. You can use dovyski@gmail.com.

greysondn commented 9 years ago

Subject: flixel-community #227 code

Sent roughly five minutes before this comment. Don't forget to check your spam filter.

Best of wishes and don't be afraid to ask me if there's something you think I could clear up or could help you with.