ET Jaymod Bug Release

[the cake is a lie]

Thanks @nUllSkillZ,

 

Issue opened for landmine spotting/defusing conflict. I was able to recreate the issue without much trouble. The bug happens with all three landmine types in jaymod. This issue does not occur in default ET (etmain).

 

You can even trigger the conflict as the defuser. Just step onto the mine as you would to defuse it, spot the mine with binoculars (it won't change to a flag but it will say "Landmine Spotted" and show up on the command map), and then try to defuse it with pliers.

 

Thanks to everyone who has submitted issues so far! Please continue to submit jaymod bugs in this thread, to this github repo, or contact me on jay1 as nullSkillZ did yesterday.

[nUllSkillZ]

Can't delete the post.

Edited May 19, 2022 by nUllSkillZ

[Haruhi Suzumiya]

I have been running into an issue a lot picking up weapons.  Standing, crouching, prone, none of it works.  I am using client side prediction for pickups, could always try server side but I hear client side is more accurate.

[the cake is a lie]

Errors like this are usually some kind of network issue or sometimes a clipping issue. There isn't much that can be done in the mod to resolve these kind issues directly but I'm planning to add some more debugging tools to help detect and understand the causes and effects of the various kinds of lags and desync issues in ET/jaymod.

 

7 hours ago, Haruhi Suzumiya said:

I hear client side is more accurate

I haven't looked into it but I would guess client side is faster (because it is predicted) but less accurate (the server may not agree that you picked it up because of your true position).

 

-----

 

Also, opening a bug for the unofficial use of commands 'silently' in team and fireteam chat. This will preserve the intent of restricting \! commands to admins of the appropriate level.

 

Keep the bugs coming everybody!

[Haruhi Suzumiya]

I'll test both client and server side and see what comes  of it.

 

[the cake is a lie]

Here is a fun one:

 

The crosshair placement in the Windows version of the Jaymod game client is slightly different from the Linux version. Crosshair placement in the Windows version is identical to stock etmain and other mods. The Linux version has every crosshair offset by 1 pixel upward. This cannot be fixed by using cg_crosshairY because of how the client scales cg_crosshairY before applying the transform (this will be fixed though). This difference exists in both stock Jaymod 2.2.0 and the custom version that F|A distributes on it's servers.

 

Usually the clients are built at the same time or at least using the same code base so that you get the same game no matter what OS you are playing on.

For whatever reason, this was not the case with the Windows and Linux clients of Jaymod 2.2.0.

 

The windows client was built from code version 1967:

 

The linux client was built from code version 1965:

 

The only difference I am aware of so far between the builds is the crosshair being off-center but there are probably other differences. Without the sources for revs 1965 and 1967 there is no way to be sure. I don't understand how this particular mistake happened but it is probably here in line 2289 of rev 1965:

 

trap_R_DrawStretchPic( x + 0.5 * (cg.refdef_current->width - w), y + 0.5 * (cg.refdef_current->height - h) + 1, w, h, 0, 0, 1, 1, hShader );

[nAbish]

48 minutes ago, the cake is a lie said:

Here is a fun one:

 

The crosshair placement in the Windows version of the Jaymod game client is slightly different from the Linux version. Crosshair placement in the Windows version is identical to stock etmain and other mods. The Linux version has every crosshair offset by 1 pixel upward. This cannot be fixed by using cg_crosshairY because of how the client scales cg_crosshairY before applying the transform (this will be fixed though). This difference exists in both stock Jaymod 2.2.0 and the custom version that F|A distributes on it's servers.

 

Usually the clients are built at the same time or at least using the same code base so that you get the same game no matter what OS you are playing on.

For whatever reason, this was not the case with the Windows and Linux clients of Jaymod 2.2.0.

 

The windows client was built from code version 1967:

 

The linux client was built from code version 1965:

 

The only difference I am aware of so far between the builds is the crosshair being off-center but there are probably other differences. Without the sources for revs 1965 and 1967 there is no way to be sure. I don't understand how this particular mistake happened but it is probably here in line 2289 of rev 1965:

 

trap_R_DrawStretchPic( x + 0.5 * (cg.refdef_current->width - w), y + 0.5 * (cg.refdef_current->height - h) + 1, w, h, 0, 0, 1, 1, hShader );

I'm glad you're around, Cake. Love seeing stuff like this. Hope you can figure it out. Would this increase acc. any? xDDDD

[DIM]

On 5/10/2022 at 10:04 AM, the cake is a lie said:

- remove FPS dependent recoil

- remove FPS dependency from turn spread

 

did i understand correctly that these fixes will fix the crosshair twitching at different fps below 125? for example, i play with 62/76/85 fps on some maps for technical reasons, and the spread from bullets increases than with a stable 125 fps, visually and physically.

if i didnt understand it correctly, then please explain it to me. thanks.

[the cake is a lie]

Generally yes. Everyone will get the same recoil (pistols, MG42, sniper) and the effect of mouse movement on bullet spread will be standardized.

 

There will still be a tiny bit of variation because of how floating point math works but it shouldn't be noticeable.

 

This is how spread currently grows when making a steady 90 degree turn at various frame rates and how the patch (dotted line) standardizes spread to be most similar to com_maxFPS 166:

 

 

The slight twitching of the crosshairs and how it varies with FPS can't be entirely standardized because of how it interacts with the display refresh rate but it will hopefully be better than it is now.