Project

Profile

Help

Bug #136

closed

Problem with macros during multiplay

Added by Moritz Wolfart almost 6 years ago. Updated almost 6 years ago.

Status:
Closed
Priority:
Normal
Category:
Customization
Sprint/Milestone:
% Done:

100%

Company:
-
Contact person:
-
Additional contact persons:
-

Description

Hey hey!
I encountered a strange behaviour while multiplaying.
I am using Macros for movement. 8 for north, 2 for south and so on.
And if i am playing two characters in the same world, this macros behave in a strange way. I am not sure to which playing window the macros are sended. I think to the first one i logged in to the world.

Hope i could explain the problem correctly.

Best wishes!
Moritz:-)

Actions #1

Updated by Vincent Le Goff almost 6 years ago

  • Tracker changed from Feature to Bug
  • Category set to Customization
  • Assignee set to Vincent Le Goff

Ok, let me see if I have understood and can reproduce it. I will comeback to you with more questions if I'm unable to see the behavior.

Actions #2

Updated by Vincent Le Goff almost 6 years ago

  • Sprint/Milestone set to 16
Actions #3

Updated by Vincent Le Goff almost 6 years ago

Okay here's the thing. A lot of CocoMUD is still world-oriented, versus session-oriented. Worlds were here before characters or sessions. What it means is that a lot of feature are still centered on a world, and common to a world, instead of being spread on several clients as it should.

In our case, different clients are indeed created. However, only one sharp engine per world is created. Here's a very simple test:

  1. Open CocoMUD.
  2. Open a world (connect to it).
  3. Open the same world in a different tab of CocoMUD.
  4. Open the Python console in this tab (menu tools -> open Python console...).
  5. Type in the following:
    (panel.client, panel.client.factory.sharp_engine.client)
  6. Compare both numbers. Repeat these steps in the other tab.

What happens is quite simple:

  1. CocoMUD connects the first time to the world.
  2. It creates a sharp engine for this world.
  3. It then connects another client to the same world.
  4. It already has a sharp engine for this world, so both clients share the same sharp engine.

This explains the situation. Now, a quick fix isn't a simple matter: the best option would be to remove the sharp engine from the world, and place it in the session (see source:src/session.py) which is incidentally ready for the change as you can see. A client equals a session, this is the closest analogy we have, so I would think the sharp engine should be bound here.

It's not a coincidence, however, if worlds are bound to a sharp engine: most methods used by the sharp engine and its functions work on the world. Take #macro for instance which creates a macro in the world. When the world first loads, it reads the content of `config.set` file which contains the configuration specific to this world. This configuration shouldn't be read several times, just the first time the world is loaded.

Actions #4

Updated by Francisco Del Roio almost 6 years ago

Vincent Le Goff wrote:

Okay here's the thing. A lot of CocoMUD is still world-oriented, versus session-oriented. Worlds were here before characters or sessions. What it means is that a lot of feature are still centered on a world, and common to a world, instead of being spread on several clients as it should.

In our case, different clients are indeed created. However, only one sharp engine per world is created. Here's a very simple test:

  1. Open CocoMUD.
  2. Open a world (connect to it).
  3. Open the same world in a different tab of CocoMUD.
  4. Open the Python console in this tab (menu tools -> open Python console...).
  5. Type in the following:
    [...]
  6. Compare both numbers. Repeat these steps in the other tab.

What happens is quite simple:

  1. CocoMUD connects the first time to the world.
  2. It creates a sharp engine for this world.
  3. It then connects another client to the same world.
  4. It already has a sharp engine for this world, so both clients share the same sharp engine.

This explains the situation. Now, a quick fix isn't a simple matter: the best option would be to remove the sharp engine from the world, and place it in the session (see source:src/session.py) which is incidentally ready for the change as you can see. A client equals a session, this is the closest analogy we have, so I would think the sharp engine should be bound here.

It's not a coincidence, however, if worlds are bound to a sharp engine: most methods used by the sharp engine and its functions work on the world. Take #macro for instance which creates a macro in the world. When the world first loads, it reads the content of `config.set` file which contains the configuration specific to this world. This configuration shouldn't be read several times, just the first time the world is loaded.

Architectural problems...

I see. It would be tricky to solve this.

Well, we must apply this fix for this version. For the next, I will try to rework these parts to decouple all Sharp objects, including the engine.

Cheers,

Actions #5

Updated by Vincent Le Goff almost 6 years ago

Not too bad, but it might not be simple. I'll do it today or tomorrow if you want. There are only two open issues in sprint 16 so next built should be published tomorrow.

There might be some keys to check/translate into Spanish if you feel like it:

Thanks again for the contributions for this build!

Added by Vincent Le Goff almost 6 years ago

Revision ee1d7ad7 | View on GitHub | diff

Fix #136: remove the dependency of the sharp engine on world, add it to sessions

Actions #6

Updated by Vincent Le Goff almost 6 years ago

  • Status changed from Open to Closed
  • % Done changed from 0 to 100

Please register to edit this issue

Also available in: Atom PDF Tracking page