bcc (An ACS Compiler)

[Positron]

Thanks for helping out here, Monsterovich.

Hi Positron, stumbled upon this project on the ZDoom forums and I must say I am very excited about BCC. Finally, there are some direly-needed features I desperately need (namely structs and enums) in an ACC compiler and the best thing of all (well, my current impression at least) is that BCC looks like it's pretty much ACC with extra features "tacked on" (as awesome as the other variants of ACC compilers available out there, I just cannot migrate because my ACS codebase would require heavy porting work to adhere to these compilers' vastly different syntax). So it seems perfect for my purposes.

(BTW, I don't mean "tacked on" as a bad term :) - maybe it's not the best word to use in this situation but I am actually happy that BCC appears to remain mostly compatible with normal ACC)

The projects I am working on contain a lot of code and I am pretty deep in the production timeline, so I just cannot afford to be set back to "port" my code to a compiler that is so syntactically different... BCC however, looks like I won't have to deal with that stuff!

I don't think I'm using any ZDoom-specific functions (ie something that won't build for Zandronum ACC)... I guess if there's something that won't build, I could run a diff between BCC and ACC and if the changes are small, possibly re-introduce said ZDoom-specific function and make my personal fork of BCC...

I haven't actually tested BCC yet, but I'm going to have to create a branch for my project to make sure my ACS codebase builds perfectly with BCC and possibly fix any tiny errors that might pop up... and then on to rigorous testing of my game to make sure every feature remains working. ;) Once that's done, I'm going to do some heavy re-write to my inventory system (second video) to make use of structs (you don't want to know how I hacked ACC to make this... using multiple global arrays and synchronizing them together to support the data struture I need for my items - UGLY!).

This went on longer than it should have - sorry you had to read all that. ;D Just wanted to say thanks for putting this out... can't wait to test BCC!

EDIT: Upon looking at BCC's source... this isn't really based off ZDoom's current version of ACC, is it? Hmm. That might be problematic for me then... I'm using ZDoom's ACC and heavily rely on it (for example, I use named scripts exclusively). I'm not sure what are the differences between Zandro's ACC and ZDoom's ACC but I imagine the gap to be big? Oh well - I'll get to test and see if I run into any problems.

phpBB [video]

The [video] tag is deprecated, please use the [media] tag

phpBB [video]

The [video] tag is deprecated, please use the [media] tag

Hello, Nash. Good stuff in those videos, man.

With bcc, I tried to keep it mostly compatible with existing ACS code. There are differences, like block-level scoping and short-circuit evaluation for the logical operators. I feel like these things were important to implement than having the older behavior. So keep these in mind if you plan to use this compiler.

EDIT: Upon looking at BCC's source... this isn't really based off ZDoom's current version of ACC, is it? Hmm. That might be problematic for me then... I'm using ZDoom's ACC and heavily rely on it (for example, I use named scripts exclusively). I'm not sure what are the differences between Zandro's ACC and ZDoom's ACC but I imagine the gap to be big? Oh well - I'll get to test and see if I run into any problems.

The compiler was started from scratch. It's a new code base. There are still things to clean up and fix, which I'm currently doing.

I keep the compiler updated to work with the features of Zandronum. It's the port I play on and am more familiar with. Zandronum is behind on ZDoom features. It doesn't have named scripts yet, so I haven't implemented them into bcc yet.

For instance, named scripts DO work. I'm not sure if you wanted to know anything else.

You can use a string instead of an integer as the script number, but it's not the same. The script number will be whatever the index of the string is. If you generate the string dynamically using StrParam() and use this string to execute your script, it won't work because the index will be different, even though the string content might be the same. If you get a string from another library and use that, it might not work because the other library might have the string indexes allocated differently. These also apply when using strings as cases inside a switch statement.

[MidnightWolf]

Thanks for this great stuff. While I understand your goal is to be compatible with Zandronum, I would be really happy if you added Zdaemon's two team score functions described here: http://downloads.zdaemon.org/zdaemon.acs

These are Zdaemon's only exclusive ACS functions and they are very unlikely to change in the future.
Zdaemon is behind Zandronum in functionality, and scripters there could benefit from BCC as well. I imagine it is not too big effort to add these functions yet it would make a big difference for modder folks over Zdaemon.

Thanks in advance

[Slim]

Thanks for this great stuff. While I understand your goal is to be compatible with Zandronum, I would be really happy if you added Zdaemon's two team score functions described here: http://downloads.zdaemon.org/zdaemon.acs

These are Zdaemon's only exclusive ACS functions and they are very unlikely to change in the future.
Zdaemon is behind Zandronum in functionality, and scripters there could benefit from BCC as well. I imagine it is not too big effort to add these functions yet it would make a big difference for modder folks over Zdaemon.

Thanks in advance

Kinda curious as to why someone'd be asking someone who made something for Zand make it also work with ZDaemon, helping a port unrelated to this one is like asking a nieghbor to leave their door open to robbers so they can break in.

[MidnightWolf]

Robber parallelism is kinda wrong to be used here (idk how that conclusion even remotely related). I never understood the hate between Zandronum and Zdaemon, we are doom players after all, just our tastes differ a bit.

This is an ACS compiler, so any port using ACS is related at least to some extent (with that said, ZDaemon). The author wants to keep up with Zandro features, which does not exclude Zdaemon because it is behind Zandronum.
These two functions are just minor differences.

Back on topic, all I asked was the addition of two functions which does not harm Zandronum or BCC in any way, but would make the life easier for folks over ZDaemon too. Let the BCC author decide the fate of this request IMO.

[Wirtualnosc]

Can't these be added through std.acs already?

[MidnightWolf]

I tried that already, Zdaemon's team score functions belong to the internal functions section in theory (their codes are negative numbers), yet the game gives unknown p-code error when I try to execute a script containing the functions. I believe BCC support for functions is hardcoded, but correct me if I am wrong.

EDIT: Nevermind, I was dumb, std should look like this:

//Extension functions
int GetTeamScore( int ) -= 19620;
void SetTeamScore( int, int ) -= 19621;

[RaveYard]

THIS IS GREAT!
The best thing about this is, that it is pretty much like ACS
but I still kinda wish there was more to this.

Like: classes, pointers, function pointers..
but the most important one is ability to use struct as return value and parameter

[Ivan]

I think there is a problem with HudMessages with BCC. I noticed that if you have a hudmessage with the flag hudmsg_log, it will cause weird things to happen. For example, I have a hudmsg_fadeinout that works properly without the hudmsg_log flag, but with it added it stays on screen permanently. Doesn't happen with other compilers.

[ZzZombo]

Can't confirm, was using for a few hours a HudMessage() w/ both these flags and no problems.

[Positron]

I uploaded new binaries of the compiler. Changes include:

• Fixed: fixed-point number literals not being properly converted from their textual representation in some instances
• Updated names of database functions
• Added ZDaemon functions provided by MidnightWolf

Links:
32-bit (Click `Raw` to download)
64-bit

I couldn't quite figure out the issue you might have, Ivan. Doing some experimenting, I noticed that sometimes a fixed-point number literal will not be converted properly to a numeric value. I fixed this bug. Maybe this had something to do with your problem. Please give the new binary a try to see if it fixes that.

[Ivan]

Now, it doesn't stay forever however somehow the hudmessages do not fadeinout properly.

Code: Select all

// Save the stuff to the console
Script 988 (int num, int type) CLIENTSIDE
{
	if(ConsolePlayerNumber() != PlayerNumber())
	  Terminate;
	SetFont("BIGFONT");
	if(!type)
	{
		HudMessage(s:UpgradeNames[num]; HUDMSG_FADEINOUT | HUDMSG_LOG, UPGRADETEXTID, -1, 0.5, 0.15, 2.0, 0.5, 0.5);
//		Log(s:UpgradeNames[num]);
	}
	else
		Log(s:RewardTextBot[num]);
}

[Positron]

Now, it doesn't stay forever however somehow the hudmessages do not fadeinout properly.

Code: Select all

// Save the stuff to the console
Script 988 (int num, int type) CLIENTSIDE
{
	if(ConsolePlayerNumber() != PlayerNumber())
	  Terminate;
	SetFont("BIGFONT");
	if(!type)
	{
		HudMessage(s:UpgradeNames[num]; HUDMSG_FADEINOUT | HUDMSG_LOG, UPGRADETEXTID, -1, 0.5, 0.15, 2.0, 0.5, 0.5);
//		Log(s:UpgradeNames[num]);
	}
	else
		Log(s:RewardTextBot[num]);
}

What do you mean exactly by they don't fadeinout properly? I compiled your code, filling in the missing code where necessary, with both acc and bcc. Both compilers generate similar output. The small differences are in optimizations. For example, bcc evaluates HUDMSG_FADEINOUT | HUDMSG_LOG at compile-time instead of at run-time like instructed by acc. I don't think these optimizations have something to do with your problem.

[Ivan]

Like, it doesn't fadein and fadeout, it instantly shows.

[Positron]

Fixed a bug involving hexadecimal literals. Hopefully this fixes your problem, Ivan.

Download links:
Windows 32-bit (Click Raw to download)
Windows 64-bit

[Ivan]

Sadly it did not. Weird thing is it seems to work as intended on my end with the normal compiler for some reason.

[Positron]

That sucks, man. May you please paste the code, enough so it compiles without errors, and also upload the object file? I want to examine the object file for any anomalies.

[ZzZombo]

Can somebody make a binary from the latest commit to date, please? I really hope on Monsterovich here, BTW ;).

[Monsterovich]

Can somebody make a binary from the latest commit to date, please? I really hope on Monsterovich here, BTW ;).

https://github.com/Monsterovich/bcc/blo ... 01_x86.zip

[ZzZombo]

Thanks!

[gato606]

Doubt anyone needs these, but here you have: Macintosh Intel (i386 and x86_64) binaries:
https://www.dropbox.com/s/r8vrxmduywuxr ... .06.15.dmg

Linux and FreeBSD binaries coming soon (of course if someone won't be faster than me).

Keep in mind that these binaries are nohow official, thus Positron takes absolutely no any responsibility for them. I do though.

No Linux version yet?? Boy... that's sad
I will have to wait then