v1.4 Public Python Beta Testing

[tobi-the-fraggel]

Does that mean that all of this changes are to be seen on v1.4 when it comes out?

Well, I am not a dev but this is a beta-test. So naturally the purpose of it is to test new features which are about to be implemented. Depending on the outcome of this tests these feature will be implemented or not.

[Mats391]

We found the issue with tickets and resolved it. Fix is sent to server admins and will be applied shortly

[Captain Fubar]

Just give me that 1.4 I will masturbate to its changleog

[Scubbo]

neat features, thanks for the development gents! look forward to a stable release!

[qs-racer]

Updated outpost overrunning system. They now get overrun when:

8 soldiers are within 150m
4 soldiers are within 100m
2 soldiers are within 50m

Does this apply equally to 4km and 1km map ?
I am afraid on a 1km map that all the fob will be overrun. Especially because fob are 300m away from the mainbase, on each side. Meaning it will be 200m far from the battlefront (consider middle of the map)... Very close from the 150m overrunning system.

[=-=kittykiller]

well done you guys! the civi medic was the dogs nutsack now its the cats ****

[Mats391]

Does this apply equally to 4km and 1km map ?
I am afraid on a 1km map that all the fob will be overrun. Especially because fob are 300m away from the mainbase, on each side. Meaning it will be 200m far from the battlefront (consider middle of the map)... Very close from the 150m overrunning system.

It is same on all maps. This is why we have this beta, to see how that plays out. I played a round on Fallujah so far as only 1km map and there the Insurgents had quite some hideouts that stayed up for long. So either ranges are fine for 1km or it was bug

Meanwhile we have another fix out. Fixing minor errors mostly with realityadmin.

Spoiler for Changes:

New GameStatusHandler
Improved errorLogging once more, logs current gameState
Added logging of events that try to trigger during Loading/EndGame status and get stopped. These get logged in python_errors.log
Fixed switching in vehicles. Marks to switch on next death
Fixed sayTeam through PRISM
Fixed players with 3 character long names being kicked
Fixed issues with kill logging
Fixed ServerTeamMessage being logged
Fixed index out of bounds in realityvehicles

[CG-Delta]

Does this reset the ridiculous 2 min civi timer if I have shovel in hand and switch it to melee again?

The manual says 1 minute.
"If a civilian does any of the following listed below, he will be considered a combatant for 1 minute afterwards and may be shot without penalty:..."
Is it 2 minutes as of now?

[Mats391]

It is 2 minutes for quite some time now, the manual just is not as up-to-date as we wish it to be.

[Yosik]

Does the BF2 game engine load the python standard library from the rar at the main folder every time? In that case, isn't it possible to upgrade the interpreter to a newer version which is probably more optimized? Or give a shot to optimized alternatives like pypy?
I couldn't find information about this anywhere, I would give it a try if you'll give me some directions. I guess that the major issue here would be replacing dice_py.dll but it seems like it mainly contains Python.h functions that can be upgraded

I was also wondering why realityprism uses rcon.invokes to communicate with prism instead of events

[Mats391]

Does the BF2 game engine load the python standard library from the rar at the main folder every time? In that case, isn't it possible to upgrade the interpreter to a newer version which is probably more optimized? Or give a shot to optimized alternatives like pypy?
I couldn't find information about this anywhere, I would give it a try if you'll give me some directions. I guess that the major issue here would be replacing dice_py.dll but it seems like it mainly contains Python.h functions that can be upgraded

I was also wondering why realityprism uses rcon.invokes to communicate with prism instead of events

Updating the python library is possible, but we are yet to see to what extend. No idea if updating to a version with threading would be possible etc. and just using a slightly newer version wont be worth the risk of losing potential optimizations DICE made.
realityprism uses rcon.invokes as those seem to be executed in a different thread than normal events. Not sure if really the case tho.

[Yosik]

Updating the python library is possible, but we are yet to see to what extend. No idea if updating to a version with threading would be possible etc. and just using a slightly newer version wont be worth the risk of losing potential optimizations DICE made.
realityprism uses rcon.invokes as those seem to be executed in a different thread than normal events. Not sure if really the case tho.

Does that mean that the handling of the event 'RemoteCommand' happens in rcon_invoke's thread as well?
It's just odd to me from a design standpoint. While writing a server plugin that took info from realityrally I was forced to register 'RemoteCommand' and filter prism messages. I was under the impression that the prism devs just wanted to avoid creating a new event for rallyrally since other modules don't need to watch for rallydelete/rallycreate at this moment.
If you're going to sacrifice design for performance, wouldn't it be more beneficial to bypass the main event thread with events that aren't CPU bound(See: GIL)? rallyDelete/rallyCreate are very insignificant.

After further inspection, it seems like the dice_py.dll is taken directly from the python2.3.4 source with small modifications/reductions.
Based on my checks, threading support isn't blocked in any of the dlls/exe, it simply wasn't precompiled into the .dll from thread.c(it's available in dice_py_d.dll but I don't have the time to get bf2_r.exe to work with it today) so it should be possible to have threading after upgrading.
Taking over dice_py.dll would allow

• Exporting your own python.h that you can use to embed C code into heavier python functions.
• Threading/Multiprocessing
• Easier and faster development with python's newer features
• Less bugs, more optimizations
• Getting rid of CPython for pypy - noticeably faster, stackless, greenlets instead of threads(awesome).

[Mats391]

I remember Tema working with exchanging python and running into issues. Not sure what he tried tho.

[RAWSwampFox]

Good Morning,

While it has been quite awhile since I've played. I think I've played 3 hours in the past 55'ish days and 2 of those were last night on Al Basrah. I've read through this thread and combined that with what I heard in-game last night and had a question or two.

1. In the past, just holding an epi did not reset your timer, using it did. Based on what I'm reading here, if you are just holding it it resets your timer?
2. Along the same thought, if you were shot while holding the epi in the past, you did not martyr but it didn't reset your timer.
3. I have been shot while deploying an asset and not martyr but it was like a split second shot. So using the phone while deploying now resets the timer?
4. Does the "using the phone to keep in touch with your mama" affect squad members who put the phone to their ear or is it only in the "deploying asset" action?
5. Could a timer be added to people who get arrested 5 times? Something like what is in place for blufor players who martyr 5 civilians? My reasoning for this is that I strongly feel that due to the fact that we actually have a spawnable CIVI kit, people are picking this more and sabotaging the game either purposefully or due to the convenience factor.
6. Will Al Basrah v2 be included in the 1.4 release and set as the standard map layout? < answered in another thread.
7. Will any of the INS v2 beta things be put in? Example, the commander being able to place / move the unknown? Love that feature.

[Mats391]

1. In the past, just holding an epi did not reset your timer, using it did. Based on what I'm reading here, if you are just holding it it resets your timer?
2. Along the same thought, if you were shot while holding the epi in the past, you did not martyr but it didn't reset your timer.
3. I have been shot while deploying an asset and not martyr but it was like a split second shot. So using the phone while deploying now resets the timer?
4. Does the "using the phone to keep in touch with your mama" affect squad members who put the phone to their ear or is it only in the "deploying asset" action?
5. Could a timer be added to people who get arrested 5 times? Something like what is in place for blufor players who martyr 5 civilians? My reasoning for this is that I strongly feel that due to the fact that we actually have a spawnable CIVI kit, people are picking this more and sabotaging the game either purposefully or due to the convenience factor.
6. Will Al Basrah v2 be included in the 1.4 release and set as the standard map layout? < answered in another thread.
7. Will any of the INS v2 beta things be put in? Example, the commander being able to place / move the unknown? Love that feature.

1. It always did. We always just checked for selection. We cannot even check for when someone actually uses a weapon aka fires it.
2. It did.
3. Deploying assets resets ROE timer.
4. You can do anything you want with the phone, just not deploy hideouts, roadblocks or SPG
5. Getting arrested already gives you spawn penalty.
6. Yes
7. Not in 1.4

Today is last day of the beta, so last chance to hop on and test it yourself!

[Luiscuella]

The command of !mvote will still running?
We use the !scramble no Divsul Server too... please we want to keep it.

Good Work!

[xambone]

Mats391;2140527']
[*]Made attackers unable to "hold" a flag. If you do not have the required 2-1 ratio the flag will go back towards defenders. If neither team controls flag, it goes back towards neutral.

[*]Removed spawn point from unknown caches

This is game changing. Like Whoa

[RAWSwampFox]

Good Evening,

Originally Posted by [R-DEV]Mats391 View Post
[*]Made attackers unable to "hold" a flag. If you do not have the required 2-1 ratio the flag will go back towards defenders. If neither team controls flag, it goes back towards neutral.
[*]Removed spawn point from unknown caches

While I understand the reason for the above, it also speaks to the lower "tactical aware" player abilities now on PR. I disagree with the first one but, hey, I'm not a DEV. I understand the last one and somewhat disagree with it. If people built hideouts and were more "tactical aware", this wouldn't be an issue. I mean, hey, INS is broken anyways (from what I've been told over and over again), that is why we play it a lot, right?

[ParaPlays]

Will the battle recorder system get an update, it currently crashes all the time about 10 seconds in. Many people having the same issue, it's a shame as creating videos using it helps promote the game.

[camo]

Will the battle recorder system get an update, it currently crashes all the time about 10 seconds in. Many people having the same issue, it's a shame as creating videos using it helps promote the game.

It's fixed for 1.4.

This is the work around for 1.3.9.
https://www.realitymod.com/forum/f27-pr ... ost2118757