[SOLVED] Cannot get cl_simpleitems_postfix to work

As far as I understand, cl_simpleitems_postfix allows me to load different models in place of the defaults.
For example, considering the following files:

model.md3 (default)
model_simple.iqm (included in release)
model_foo.iqm

Then cl_simpleitems_postfix "_foo" should cause model_foo.iqm to be loaded in place of model.md3.
However, Xonotic seemingly ignores the cvar and always loads the models with the "_simple" postfix.

I've attached a test case, which should replace the health models with red squares when you set

Code:

cl_simple_items 1
cl_simpleitems_postfix "_foo"

Does the test case work correctly for anyone?
Is my setup incorrect?
Is this a bug?

Any help is appreciated.

Looking at deox's simple items pack, it comes to my mind that you should name your file: model_simple_foo. The game may be looking at the second sufix.

Looking at deox's simple items pack, it comes to my mind that you should name your file: model_simple_foo. The game may be looking at the second sufix.

I gave it a try, to no avail. According to the code, it should only add the given postfix to the name, without a "_simple" in the middle.
But I cannot find the cause of the failure to load the "_foo" models.

deox's pack rewrites the existing "_simple" skin files, which is the only method I know to work.
However, that causes a competition of resources, forcing players to handle file conflicts of multiple icon sets.

Yeah, the code is broken due to a typo. Commit 900a3c0 introduced the postfix cvar, but notice that it is wrongly named:

Code:

var string autocvr_cl_simpleitems_postfix = "_simple";

This should read "autocvar". This means that the QC variable doesn't point to the right cvar, hence the constant value.

I'm reporting this to the bug tracker, that's a good excuse to try and use gitlab

EDIT: There it is, https://gitlab.com/xonotic/xonotic/issues/86

deox's pack rewrites the existing "_simple" skin files, which is the only method I know to work.
However, that causes a competition of resources, forcing players to handle file conflicts of multiple icon sets.

yeah, i tested this with Mario when investigating alternatives, and found the cvar was rendered useless. So, I just overwrote the files.

This should read "autocvar". This means that the QC variable doesn't point to the right cvar, hence the constant value.

Yes, that's it, thank you very much. With the "a" inserted, it works correctly. I should have noticed that myself, sorry for the extra work.
Since you asked for repository access, I assume you already made the fix, ready to be pushed. Otherwise, I'll give it a shot.

yeah, i tested this with Mario when investigating alternatives, and found the cvar was rendered useless. So, I just overwrote the files.

With the fix, every set can use its own models, enabling players to easily switch between them.

i'm aware, good. This seems like a good time to bring up the problems smilecythe and some others have had, that simple items dont show up with r_showsurfaces 3 (i believe that was what it was).

i'm aware, good. This seems like a good time to bring up the problems smilecythe and some others have had, that simple items dont show up with r_showsurfaces 3 (i believe that was what it was).

Well, r_showsurfaces replaces all shaders with single colors. If the simple model is just a square with a "deformVertexes autosprite" shader, you end up with a colored square.
Someone will have to make actual 3D models to make them work with r_showsurfaces.

haha no not happening that's ridiculous

Yes, that's it, thank you very much. With the "a" inserted, it works correctly. I should have noticed that myself, sorry for the extra work.
Since you asked for repository access, I assume you already made the fix, ready to be pushed. Otherwise, I'll give it a shot.

No sweat, I noticed it as soon as I looked at the code. I didn't make the fix, so if you have repo access, feel free to push it. I say the first one of us who has time for that should do it and update this thread

I'll try to remember tomorrow.

EDIT: As for r_showsurfaces, damn, that's a good point. I don't think much can be done about that! r_showsurfaces is originally meant to be a debugging feature...

I didn't make the fix, so if you have repo access, feel free to push it. I say the first one of us who has time for that should do it and update this thread

Done. At least I think so, I'm new to gitlab as well

EDIT: As for r_showsurfaces, damn, that's a good point. I don't think much can be done about that! r_showsurfaces is originally meant to be a debugging feature...

i know, but it's smilecythe's preference to set the graphics that way for competitive play for some insane reason, so i tried to provide for him

Closed issue #86, thanks sev.