Build and Shoot

Sonarpulse can I please get a sentence or to from you saying how it's going along for the newsletter?
Thanks!

Well I'd like to finish doing the assembly like I described to Cajun. That should allow linux x86-32 builds, which is portable enough for most users. (x86-64 Linux & Mac will work, just not be as quite efficient as they could be, but certainly no less efficient than current AoS which is also only 32-bit.) Then I will go through Cajun's other commits which I haven't merged yet and work them in, as it's about time I do that.

However, I am pretty busy with school so more difficult things like other CPU architectures and actually building a client will probably have to wait for the summer. And if Iceball has progressed enough by then I'll be less interested in making a client.

Sorry If that's kind of dismal. I just want to be honest, as most mods and indie games I've observed over the last 5+ years often die amidst unrealistic expectations and developer optimism. I will say that by the end of next summer I am 100% committed to finishing at least porting Voxlap itself, as that was my original baseline goal and I strongly believe in finishing what I started.

Sonarpulse wrote:

I will say that by the end of next summer I am 100% committed to finishing at least porting Voxlap itself

Remember to brag to Ken about it

But only after slaping him for making voxlap so bitchy!

I'm going to wrap my head around what the kv6 load and save actually do, and learn some assembly, because, heck, I wanted to learn it at one point, and this is as good a reason as any. Currently I'm frustrated in my efforts to view GAS's manual.
>:([/siez]

Yeah the more esoteric assembly related things in both GAS and GCC are embarrassingly poorly documented. Certainly if you have any questions feel free to post them here. If it's somehow relevant to Voxlap I might have already dredged up the answer.

Good luck!

You might have expected more progress... I'm now at the point that I sort of understand that extended ASM is GCC's way to mix C and ASM, and is used via the GCC specific feature of constraints (which is a way to influence the resulting assembly in general). I understand the syntax somewhat.
Also I grasp the very basics of ASM, and the AT&T and Intel syntaxes and their differences. But of course most of the ASM is SIMD instructions, so I've got way more to learn
OMG I just noticed I forgot how to reference the data at an address in AT&T.
Now my questions:
It seems I'm not grepping any "memory" in clobber lists on conservative. That might be a fix-me? Or did you manage to separately add all elements of an array all the time? I saw it was said that leaving things in Intel syntax will mitigate the need of (big) clobber lists, but isn't that just wishful thinking? From your work I understand you can't just add a constraint list to "import" the variables and leave the original Intel syntax intact? Because, MSVC apparently does allow local symbols to be referenced in assembly? Or are we in trouble, only because some parts need C++?
Some criticism:
Since the documentation is so vague, I think we should use the %c[] syntax as little as possible (which is: never?). The original code should not be in the repository. (Maybe voxlap5.cpp.orig got there by accident?)
I'm rewriting a tutorial from Intel to AT&T now. I had a quick look at the current assembly in conservative; I better have a slow look at some occasion ;

Sorry for not working on this project, but I've been very busy with school and activism lately...

But I'm hoping to be able to look at this at some point, when I get the time, I'm also working on other open source projects, actually

VladVP wrote:

Sorry for not working on this project, but I've been very busy with school and activism lately...

But I'm hoping to be able to look at this at some point, when I get the time, I'm also working on other open source projects, actually

Yes activism, liking a facebook page.

Wasting days worth of time on facebook? Calling it "activism"? ...How many girls are involved?
As I don't have facebook I can't see what cause it is, but I'm sure it's hot -I mean noble.
TBH I've been goofing around this week as well... Maybe we could use a bit more direction?

Cajun Style wrote:

Wasting days worth of time on facebook? Calling it "activism"? ...How many girls are involved?
As I don't have facebook I can't see what cause it is, but I'm sure it's hot -I mean noble.
TBH I've been goofing around this week as well... Maybe we could use a bit more direction?

It's an image... but I just realised that i can post images... xD
Aw man I was tired at that point...

^ that's actually pretty neat.

Yeah we've all been busy, which is fine. Just, if any of you do find the time to work on this, please do the assembly conversion stuff I mentioned above, as that's the absolute next step.

Duff Paddy/Lensmen hasn't posted in a couple of days but if he has the time and desire to start reverse engineering the protocol and writing the client, I'd like to have this in as good as possible shape.

Hey guys, been following the thread. The port of the network code is going really well, got connection and map download, join chat and a few other bits working. I've been updating the protocol WIKI that rakiru set up. Should have a fully working c++ version of 0.75 / 0.76 netcode within the next week. It seems very easy to embed Python into the client, so I'm heading in that direction rather than Lua. Code will go up soon on github, and I'll follow up post when I've cleaned it up and published.

Duff Paddy wrote:

Hey guys, been following the thread. The port of the network code is going really well, got connection and map download, join chat and a few other bits working. I've been updating the protocol WIKI that rakiru set up. Should have a fully working c++ version of 0.75 / 0.76 netcode within the next week. It seems very easy to embed Python into the client, so I'm heading in that direction rather than Lua. Code will go up soon on github, and I'll follow up post when I've cleaned it up and published.

> Implies that VOXLAP works on LInux
> Implies that OpenGL exists.

Duff Paddy wrote:

Hey guys, been following the thread. The port of the network code is going really well, got connection and map download, join chat and a few other bits working. I've been updating the protocol WIKI that rakiru set up. Should have a fully working c++ version of 0.75 / 0.76 netcode within the next week. It seems very easy to embed Python into the client, so I'm heading in that direction rather than Lua. Code will go up soon on github, and I'll follow up post when I've cleaned it up and published.

What are you trying to tell us? I'm confused