Is anyone using the ABIC layer in Spring?

[hughperkins]

Is anyone using the ABIC layer in Spring?

If no-one's using it, we can probably remove it. I kindof feel guilty about clogging up Spring with something that no-one is using, and that just complicates the buildtree.

Edit: the ABIC layer is the stuff in the subdirectories rts\ExternalAI\GlobalAIInterfaces and rts\ExternalAI\GlobalAICInterfaces.

Basically, if your AI dll advertises a C function "IsCInterface", then its using the ABIC layer. If it doesnt, then it's not.

Edit2: what I'm going to do is wait a week before deleting it, and in the meantime target .Net interface development at an ABIC-free Spring.

[lale]

Is the new version of spring, going to require .NET?

- This might be a problem, since Vista includes a company configurable setting that, if set, only allows windows to run .NET programs, if the compiled code is signed by VeriSign or Microsoft.

E.g. on my company issued Vista, a can NOT run unsigned .NET programs.

[Dragon45]

It's gotta be possible to fake a signature. I mean, why the hell not? Somone's made a crack for it probably, Vista's been out for a while now.

[AF]

Not spring itself, I'd assume he means a .Net loader AI at least untill mono bindings are done.

[jcnossen]

Just out of interest.. how are you going to make a spring .NET interface for mingw without ABIC?

[hughperkins]

> Not spring itself, I'd assume he means a .Net loader AI at least untill mono bindings are done.

The mono bindings are done. You can find them in SVN in CSAI\monoloader.

> Just out of interest.. how are you going to make a spring .NET interface for mingw without ABIC?

The ABIC exists as a separate dll from Spring. It's still used by the .Net interface, but not integrated into Spring.

Basically, the way things are right now, having to maintain two interfaces within Spring complicates AI interface maintenance. We should really choose one or the other. The de facto standard is the C++ interface, and it's a good choice.

Really, the ABIC-integrated-into-spring thing came about because of the switch from msvc to mingw for Spring. Neither AF nor I were really happy about it. It was an emotional time. The ABIC integration was my attempt to exert some control over the migration to mingw, but really in the cold light of day, there's no real advantage in integrating it into Spring.

[hughperkins]

It's gotta be possible to fake a signature. I mean, why the hell not? Somone's made a crack for it probably, Vista's been out for a while now.

Nah, you cant fake a signature, at least it's not like decompiling code and replacing a conditional jmp with a nop. You'd have to obtain the private encryption key, using social engineering or an insanely powerful computer. Social engineering is possible, but is very rare for encryption keys. Look at Sony's PSPs: lots of hacks, but no-one has ever obtained Sony's private key AFAIK.

The other approach would be to hack windows itself to not require a signature. In this specific case, by default Windows doesnt require a signature; it's a company decision to activate it in lale's company.

[AF]

There are advantages to ABIC, however the work needed to actually use it is far to great for anyone to be bothered.

Whatsmore it lacks certain essential items, such as the cheat interface. Thats one of the main reasons NTai uses the C++ API.

[hughperkins]

Seems like one week became one year :-D

Those GlobalAICInterface directories are still hanging around. No-one's using them right? They're just clogging up the build? I'll delete them from svn now?

(Edit: we can always put them back later if we really feel like it)

[hughperkins]

Removed, commit 5927