244 replies to this topic

[quantumstate]

Mythos_Ruler, on 09 October 2011 - 12:06 AM, said:

Check out the angle which the bot constructs buildings. Jubot has it correct.

Ah yes, I was pi/2 out, it is corrected now for prettiness.

[quantumstate]

I have just written a rough overview of how qBot works. It is found at https://github.com/q...ucture-Overview

[Pureon]

Brilliant to see qBot expanding its territory with civic centres, and overall its economy is very strong now, quite comfortably equaling my own.

p.s. Thanks for the writing the overview

[Mythos_Ruler]

Pureon, on 15 October 2011 - 12:08 AM, said:

Brilliant to see qBot expanding its territory with civic centres, and overall its economy is very strong now, quite comfortably equaling my own.

Agreed. Great to see! qBot is becoming more and more of a challenge.

[Linux_Eki]

I am trying to make rusher from qbot. Its bit hard because I can't code a single line of javascript But I will try changing the way how qbot does things.

Some progress. Attacks in under 4 minutes on random latium map while original qbot takes 9-12 minutes for first attack. If tested on Oasis II attack times are quite the same 7-9mins for both.

Edited by Linux_Eki, 16 October 2011 - 02:11 AM.

[infyquest]

Linux_Eki, on 16 October 2011 - 01:28 AM, said:

I am trying to make rusher from qbot. Its bit hard because I can't code a single line of javascript But I will try changing the way how qbot does things.

Some progress. Attacks in under 4 minutes on random latium map while original qbot takes 9-12 minutes for first attack. If tested on Oasis II attack times are quite the same 7-9mins for both.

if you post the code somewhere, we all can try to improve qBot's code

[Linux_Eki]

You can already get code of qbot. If I manage to make something good that works like I want then I will post code for others to test.

[infyquest]

I have some working sent to quantum to be included in the qbot
If you have any other cool ideas to be implemented let us know.

[quantumstate]

Having different tactics available for qBot would be nice. In the future it should be possible to make qBot randomly pick a strategy and perhaps have them selectable on the AI options screen.

[historic_bruno]

quantumstate, on 16 October 2011 - 03:13 PM, said:

Having different tactics available for qBot would be nice. In the future it should be possible to make qBot randomly pick a strategy and perhaps have them selectable on the AI options screen.

Very nice

For the AI options, there's a ticket which includes that. I was thinking of defining the list of possible options in the AI JSON data, which becomes a series of dropdowns in the AI config box. Hmm, I'm not sure if setup options even get passed to the AIs, but if not we can add that

[plumo]

It would be nice if every civ has different tactics. f.e. guerrilla warfare for iberians etc...

[Mythos_Ruler]

qBot gives a Call Stack warning regarding WalktoCC.

[quantumstate]

Mythos_Ruler, on 21 October 2011 - 07:33 PM, said:

qBot gives a Call Stack warning regarding WalktoCC.

Could you give more details please? Also check you have the most recent qBot since there was an update yesterday affecting walkToCC.

[Pureon]

Not sure if this is the same error as reported above:

Function call failed: return value was -110100 (VFS directory not found)
Location: vfs.cpp:105 (VFS::GetDirectoryEntries)

Call stack:

(error while dumping stack: No stack frames found)
errno = 0 (No error reported here)
OS error = 0 (no error code was set)

Error produced at the end of the match loading screen using latest public svn and latest qbot.

Edit: I'm using the autobuild exe - which hasn't been updated for a few days now, so that's probably what's causing the error...

[Mythos_Ruler]

I'll wait for the next autobuild and test it again.

[historic_bruno]

Mythos_Ruler, on 22 October 2011 - 02:05 PM, said:

I'll wait for the next autobuild and test it again.

The problem is nothing to do with the engine but qBot assumes you will name the AI directory "qbot" instead of something like "quantumstate-qBot-93af3f2" (it's explicitly referenced by _init.js).

[Pureon]

historic_bruno, on 22 October 2011 - 07:07 PM, said:

The problem is nothing to do with the engine but qBot assumes you will name the AI directory "qbot" instead of something like "quantumstate-qBot-93af3f2" (it's explicitly referenced by _init.js).

Confirmed, fixed it for me. Thanks!

[Mythos_Ruler]

historic_bruno, on 22 October 2011 - 07:07 PM, said:

The problem is nothing to do with the engine but qBot assumes you will name the AI directory "qbot" instead of something like "quantumstate-qBot-93af3f2" (it's explicitly referenced by _init.js).

That worked, thanks.

[quantumstate]

historic_bruno, on 22 October 2011 - 07:07 PM, said:

The problem is nothing to do with the engine but qBot assumes you will name the AI directory "qbot" instead of something like "quantumstate-qBot-93af3f2" (it's explicitly referenced by _init.js).

Is there any way to fix this by using a relative path? In the meantime I have removed the extra directory.

Edited by quantumstate, 22 October 2011 - 09:31 PM.

[historic_bruno]

quantumstate, on 22 October 2011 - 08:50 PM, said:

Is there any way to fix this by using a relative path? In the meantime I have removed the extra directory.

I would think including source files in all subdirectories of the given module is the best solution, unless there's a reason to have AI directories that don't get used