GZDoom discussion (Version 2.3.1 released 2016/jan/7)

[Alekv]

Post something that can be tested, that goes for both the RenderStyle/Alpha issue and the crash. Screenshots cannot be debugged, but the crash looks like data corruption.

Unfortunately, I do not know how to test a crash, because it appears if you go through the game from the 1st level to 6-9 with this at each level you need to touch the auto-save object

Again, the bug does not appear every time, so it's hard for me to show it to you


But if you are ready to go through the game about two times (ie, only 9-10 levels) to see a bug
Then I can throw you this version of the game.

[D2JK]

Would it be possible to implement a particle generating function supporting custom sprites? I understand they would perform worse than simple particles, but I suppose better than (non-interactive) actors?

Also, a small observation about for - loops: when using multiple counter variables, declaring the type of additional counter variables (in the initialization field), will cause a startup error: Unexpected identifier, expecting ";".

For example, this works:

Code: Select all

int b;
for (int a=0 , b=0 ; a < 100 ; a++, b++)

This does not:

Code: Select all

for (int a=0 , int b=0 ; a < 100 ; a++, b++)

Are the additional counter variables automatically considered to be of the same type as the first one?

[Nash]

Should the zcajun Bots.cfg be packaged with future official GZDoom builds? As broken as they are, they're still an engine feature and currently the only way to get the CFG file is to download some old version of ZDoom...

[Gez]

It'd be better if instead of a file it could be loaded from a lump, say, CAJUNBOT or ZBOTCONF. Then it could be squirreled away in gzdoom.pk3 instead of being a loose file.

[Nash]

That's actually a brilliant idea, how come Randi didn't do that back in the day. :D

[Caligari87]

ZBOTCONF would be great for any future expandability. The current Cajun bots could be one section, other bots (if ever added) could use a different syntax.

[Rachael]

This really needs to go in the Feature Suggestions forum - or it will get lost and probably never be done simply because it'll be forgotten about by the time anyone has the chance to make a decision on it.

[Arch-vile90]

Never used a version of gzdoom lower than v 2.0 but today i tried the v 1.8 and i saw weapons are less bright than the latest versions of gzdoom, they tend to disappear in the darkness sectors.
I have a question about this: It's possible add a mode to switch the weapon light like the v1.8?

[Graf Zahl]

No, not really. That old code did some bad stuff with bright weapon frames which cannot be reinstated without causing broader problems. It also was not correct.

[ibm5155]

Why the shadow on the map feels slower while the shadows on objects feels faster?
example vídeo:
https://www.youtube.com/watch?v=6-YHYDZgP7g

[NightFright]

I see GZD v3.2 has just been released. Does anyone have a quick overview of the changes compared to v3.1?

[Rachael]

There's a 6-points highlights list right at the top of the release post in enlarged letters right under the downloads.

[NightFright]

Ah right, that eluded me since I was expecting that in the "changelog" section. ^^ Thanks a lot!

[Cromunism]

I have a couple questions about the changelog for the latest version of GZDoom. I would have posted this in the release thread in news, but it got locked before I got a chance, so I assume this is the right thread for it.

My main question is about the changelog entry "Add support for Unreal Engine 1 vertex mesh format." If I understand it right, this allows map and mod makers to use unreal engine 1 models in addition to the currently supported MD2 and MD3 format. I know pretty much nothing about this subject, so my questions are: why was specifically unreal engine 1 model support chosen, and what benefits does supporting this have over the already supported formats? I don't make maps or mods so this feature doesn't directly affect me, I was just curious about the purpose of the feature and how people might use it. It just seemed strange to me since unreal engine 1 is so old, but there's probably a good reason for it.

Second, the changelog mentions a rendering optimization that in highly detailed maps "may give a +20% performance improvement on Intel and AMD hardware" and benefits nvidia hardware also, but not as much. In this case, how detailed does a map need to be before this change would affect them? Would something like some of the larger doom 2 or TNT evilution maps count, or is it referring to higher detail than that?

[Gez]

I have a couple questions about the changelog for the latest version of GZDoom. I would have posted this in the release thread in news, but it got locked before I got a chance, so I assume this is the right thread for it.

The GZDoom 3.4 thread has been unlocked in the meantime.

My main question is about the changelog entry "Add support for Unreal Engine 1 vertex mesh format." If I understand it right, this allows map and mod makers to use unreal engine 1 models in addition to the currently supported MD2 and MD3 format. I know pretty much nothing about this subject, so my questions are: why was specifically unreal engine 1 model support chosen, and what benefits does supporting this have over the already supported formats? I don't make maps or mods so this feature doesn't directly affect me, I was just curious about the purpose of the feature and how people might use it. It just seemed strange to me since unreal engine 1 is so old, but there's probably a good reason for it.

It was an external contribution, made by someone who made a mod putting the Unreal Tournament '99 weapons in Doom. So here you have the motivation.

Second, the changelog mentions a rendering optimization that in highly detailed maps "may give a +20% performance improvement on Intel and AMD hardware" and benefits nvidia hardware also, but not as much. In this case, how detailed does a map need to be before this change would affect them? Would something like some of the larger doom 2 or TNT evilution maps count, or is it referring to higher detail than that?

The map in particular that was used as a benchmark was Frozen Time. You shouldn't really notice changes on the vanilla IWAD maps as presumably you were already getting max FPS on them.