Show Posts

This section allows you to view all posts made by this member. Note that you can only see posts made in areas you currently have access to.


Messages - Josh @ Dreamland

1351
General ENIGMA / Re: Game Maker and LLVM
« on: June 18, 2011, 10:34:35 PM »
That post had no trollable content; would you care to try again?

...I've been taking ENIGMA mostly any direction I please at this juncture. Adding new features and points of extensibility rather than going through and implementing old ones. Wonder if Dailly will try to add real types to GM any time soon...

1352
General ENIGMA / Re: Game Maker and LLVM
« on: June 18, 2011, 10:20:34 PM »
Point taken; forgoing the possibility that they will somehow include two copies of LLVM as they did SDL on Mac, I suppose it is possible that they could even decrease the size of the runner by replacing their mess with code from a crack team of caffeine-driven FOSS programmers.

Also, I hear all sorts of shit about Dailly's modifications breaking compatibility. While I don't pay any more mind to their endeavours than I might a piece of gum stuck to the bottom of my chair, it sounds like they're breaking scripts that take advantage of arguments being assumed zero when not passed. Really, I've heard so much shit about GM8.1, it makes me think there's not much else to ruin. Of course, I suppose it all really depends on who you ask; kind of like I might gag and choke upon visiting someone's ammonia-exuding rathole home, while its owner would be completely comfortable (if comparatively incapacitated).

As far as using LLVM goes, in theory, this is a great stride for GM and could mean the dawn of a new age in which it keeps up (or even surpasses) ENIGMA's own performance, depending on ten or so trillion conditions. In practice, I won't be surprised to find they still pad the runner to the next integer megabyte with nulls and include two copies of something large.

1353
General ENIGMA / Re: Game Maker and LLVM
« on: June 18, 2011, 09:48:29 PM »
I can't wait to see what this does to the size of the runner. :3

Not to mention they're already beginning to trash compatibility with old GM's. If they play their cards right, they could very well level the playing field between us in a number of respects with this new idea of theirs. I can't wait to see how they fuck it up.

I'll mention that personally, I'm already greasing up for the size overhead V8's going to bring us. As such, I'm adding a more formal extension system that will enable it to be swapped off for practical applications. When it's on, however, I imagine our "runner" is going to shoot up to 2.5M. My ass is going to be sore for days.

1354
General ENIGMA / Re: Game Maker and LLVM
« on: June 18, 2011, 09:25:39 PM »
*shrug*

I just wasn't sure they were trying to convey an intention to actually run GML through LLVM. This should be interesting. What will that mean for execute_string()? Is it going to JIT the string you pass it?

That should be at least as interesting as my choice of V8.

1355
General ENIGMA / Re: Game Maker and LLVM
« on: June 18, 2011, 09:08:26 PM »
Heh, that's kind of dumb. Unless that means they intend to forge some sort of object code with/for LLVM out of the GML, which I'll believe when I see. Anyway, whoever wrote that article was evidently clueless; he seems to believe that their stats showing obfuscation reducing the code size by 2/3 implied the runner would be 1/3 the size.

Do they mention this at all in a more formal blog maintained by someone literate?

1356
General ENIGMA / Re: Game Maker and LLVM
« on: June 18, 2011, 09:01:46 PM »
Are you sure? It may just be me, but "Having our own [obfuscator] means we can take advantage of local knowledge of the code we’re able to cheat a little more because we know the code," doesn't really imply to me that he's talking about HTML. In fact, my only response to that statement is that I don't think any of them have been so far even as decided to use do look more like, even as obfuscation the LLVM was for not be do because pickles.

1357
General ENIGMA / Re: Game Maker and LLVM
« on: June 18, 2011, 08:43:46 PM »
Heh. I feel trolled; I like how it removes unused "functions" but doesn't mention execute_string(), which can call any function GML offers.
Probably just more smoke blowing. Like rewriting the runner in C++ in the first place. I mean, except GM4Mac, which is receiving highly positive reviews even from GM's biggest critics.

:troll:

1358
CompilerSource/compiler/components/write_font_info.cpp writes the structures containing the font data. This is where you'd insert the baseline position;

CompilerSource/compiler/reshandlers/refont.h defines that structure; add a member to it.

Most of the glyph iteration is done in CompilerSource/compiler/components/module_write_fonts.cpp; you may need to copy one of the for loops out ahead of time to calculate that member (if so, you don't need to edit the structure in refont.h).

1359
General ENIGMA / Re: Enigma Turkish Lang Pack %100
« on: June 10, 2011, 10:18:04 AM »
> translation of ENIGMA r777

umad ismavatar?

1360
Issues Help Desk / Re: I cant compile
« on: June 07, 2011, 11:07:41 AM »
For now, select "Run" instead of "Compile" to test your game. This is a bug due to missing quotes in the makefile, which has been fixed but not marked "Stable" yet.

1361
General ENIGMA / Re: Enigma status
« on: May 27, 2011, 10:11:56 AM »
I'm happy you did this.

Anyway, I've read through the tracker artifacts and responded to each. I assigned most of them to me, but on one I need more info, and two I assigned to you. The two I gave you are so small that the fix should be included in the same revision. One of them is a copy-pasta, the other is simple but is going to keep giving, so for the sake of not having thirty tracker artifacts, I assigned it to you so you can fix it next time. Just remember to define them for variant and var alike. I recommend using const var& and const variant&, really.. Bah, I'll add that to my reply.

Anyway, thanks for the heads-up. Hopefully we can get the majority of them compiling with filler functions shortly.

1362
Issues Help Desk / Re: Error: Invalid search directories returned
« on: May 25, 2011, 11:00:16 AM »
I had already tried setting LC_ALL, and told you it didn't work. That was the very first thing I tried, using Windows' very own SetEnvironmentVariable. ENIGMA isn't its own shell; it doesn't require its own shell. I install Msys exclusively for rm and mkdir.

Making sure that the locale is correct for calls to GCC should be the job of ENIGMA's exec() function. We should not be setting the locale for the entire process, as LGM was designed to use translation files (we just haven't set them up).

The only method I can see us using to fix this, and the only one I see actually working at this point, is using Windows' own locale functions, because that's all that ever works in Windows. Until I can get someone to confirm that doing so works, as opposed to what I've been told so far (it causes compile to fail catastrophically due to ten undefined macros), which, for some reason (maybe ignorance), makes me want to not commit that particular change.

1363
Issues Help Desk / Re: Error: Invalid search directories returned
« on: May 24, 2011, 01:29:49 PM »
That's an insane hack. I'm not going to assume that bash.exe is a magical program that can make calls ENIGMA.exe cannot.

1364
Proposals / Re: XBox 360?
« on: May 23, 2011, 08:17:28 PM »
As we discussed on the IRC: Always happy to have some help. TGMG has tried this (porting to XBox, not writing an emulator) and was met with some major problems, but perhaps you can overcome them if you are versed enough to be writing an emulator for the system.

We'll be around to answer any porting-related questions, but XBox is outside mine and Ism's experience; TGMG's the only one of us who has tried developing for it.

1365
Issues Help Desk / Re: Error: Invalid search directories returned
« on: May 23, 2011, 11:36:43 AM »
I said it compiled, not that it worked. I don't have a way to test if it works, and if it keeps erroring for you, I'm afraid to commit it. The Windows version has a high enough tendency to throw compile errors as it is.

The easy fix is just to read the contents of searchdirs.txt yourself and replace the text in Compilers/Windows/gcc.ey with the translation in your default language. But until I have an idea of how to get the Windows locale calls working for everybody, I can't automate it.

Maybe we can fish through the sources to replace macros that are missing with the actual code they represent, but if I recall correctly, Fede, your implementation didn't seem to define one of the most basic macros we needed. Paste your error again next you're on the IRC, and I'll make sure it has no substitute.