1. Take a png image and place it in Quake/qw/gfx/, let´s call it 'hej'.
2. In console, enter: hud_group1_picture "hej"
3. Start hud_editor and make sure that group1 (hej.png) is visible on your screen.
4. In console, enter: vid_restart
5. Behold a new, random image that has taken hej.png´s place.

Excellent, I can reproduce the problem here. A lot easier to find the problem and fix it now! I'll get back to you when it's fixed. Thanks

I might have found a small bug. I compiled the binary for linux on Ubuntu 15.10.

BUG: Hud elements fps and ping are not affected by item_opacity and are always non transparent.

Thanks for the great work!

Try latest nightly

Newest build:
a) only s_khz 48 works on my OpenBSD box, others are distorted (not really a problem)
b) huge performance loss:
old version: r4766~odc870b timedemo fps: 260.9
new version: r4788~4cbe6b8: timedemo fps: 153.6

Okey, we have remade the sound handling to do mixing in its own thread to be independent of the framerates. This also means that mixahead (which is just added latency) is gone. Sound latency should be much lower now than before.

What you might have to do is to set the s_desiredsamples to like 512/1024 and try again. That should hopefully lower your CPU usage too.
Just to be sure if the performance issue is sound related or not, try to start the client with -nosound and run the timedemo again and see if it makes any difference.

Feel free to PM me on IRC, same nick or you can find me in #ezquake

EDIT: Oh and yeah I should really add your OpenBSD fixes (or a version of them)

Ok, performance issues are solved thx to dimman. s_khz still only works with 48. All other values are distorted but this looks like a local problem.

I am on Debian testing.
Both ezquake 2.2 and the development build stopped working.

when I run 2.2 I get the following:
./ezquake-gl_linux-x64.glx
Segmentation fault

When I run the dev build nothing happens.
It just hangs there, no window appears. Also, under "top" I see ezquake using 100% cpu.


Any help?

Other games all work just fine.

UPDATE: I tried running the dev build under gdb. If I stop it, I get:
Thread debugging using libthread_db enabled]
Using host libthread_db library "/lib/x86_64-linux-gnu/libthread_db.so.1".
^C
Program received signal SIGINT, Interrupt.
wildcmp (wild=wild@entry=0x7fffffff2f42 "\241\005", string=string@entry=0x5a1abd3 "duel_dmm4_locktar_vs_snusetrinner[end]2965_tmp94361.mvd" at q_shared.c:353
353 while ((*string) && (*wild != '*'))


2.2 dies like this:
Program received signal SIGSEGV, Segmentation fault.
_IO_vfprintf_internal (s=s@entry=0x7fffff7ff4d0, format=format@entry=0x6b76d8 "%s/%s", ap=ap@entry=0x7fffff7ff638) at vfprintf.c:1311
1311 vfprintf.c: No such file or directory.

I have seen this start up hang with a very large demo collection sitting under /qw. Moving it somewhere else made ezquake work again. You can still browse to the new demo folder and watch demos just fine. Looks like ezquake likes to read the entire directory if it is below /qw on startup.

You are right. Moving the replays some place else fixed both 2.2 and dev. I hope this bug will be fixed in the release.

I believe the 'wildcmp' startup issue can be fixed with fs_cache 0
this issue can also happen if you have recursive symlinks.

Latest build is giving me weird chars for the first letters when typing on the console.

Windows 10

same here. First letter after pressing tilde/console button gets colored and wrong.

This crash:
http://www.quakeworld.nu/forum/topic/6525/99402/ezquake-bots-crash-on-dm1/#99402

is still reproducible in the latest dev build.

Bug report: When entering a server from qtv.quakeworld.nu via the Play Button the fakeshaft and gamma is bugged out. The gamma became alot darker and the fakeshaft became 0, despite its actual value benig 1.

Apparently, I have 3 problems I've noticed with the latest nightbuild (03/22) - Windows Build:

1) Every screenshot taken is black, or dark blue. I don't know what causes that error at all...

2) I can see every entity underwater, Or above the water whenever I set gl_turbalpha BELOW 1. I don't know if that's related to me, but that's a bit problematic. Justificative Screenshots: http://imgur.com/a/TbcoS

3) (and I think it's a problem from me, but it happens on ezQuake 2.X aswell) Whenever I record things (through Fraps for instance), the screen is extremely dark, and won't change, even if I change the gamma & contrast. It happens occasionnally when I alt-tab aswell.

Also, is it normal several HUD_* values are listed as booleans, when they act like integers ?

EDIT: Also, a suggestion would be to add a hud_itemsclock_scale CVAR since it cannot be modified (and may be really small on high con_vidheight/width).

I think I found a bug, not tested in later builds but I figured I should report it anyway. (build: r4752~590beea)

Just changed keyboard to one that has a windows key on the left side, and with sys_disablewinkeys 1 i get input lag in windows.
So for example if i write a long sentence and push backspace, it continues to erase long after I've let go of the key.

As soon as i change sys_disablewinkeys to 0, this problem disappears.

Since s_mixahead was removed I got some "jerky" sound with some sort of "click/skip" at first millisecond of sound when it's played. It's a little bit better with s_linearsample 1 but still present. Not that it's unplayable or smth like that but annoying. Probably it's my PC/soundcard as noone else is complaining about it here =)

edit:

NVM: "s_desiredsamples 1024" sorted it out

Haven't been seeing that here, can you upload a screenshot? (if it's completely strange character, can you also mention the key you pressed, and
your keyboard layout)



This was tracked down to the dm1 GPL maps, were you able to get a copy of dm1.ent?



This should be fixed in latest build.



There will be a slight delay if sys_inactivesleep 1, but setting it to 0 will cause full cpu use. I'll look into disabling the hook when ezQuake is inactive & re-enabling after, that should fix it.



Will add to the investigate list. Can you supply more info about your config (windowed/fullscreen, system config, second monitors etc)



This is deliberate, looks like /r_novis 1? gl_turbalpha is really only for maps vissed for transparent water, which is generally none (androm9 is, cmt4 isn't, for instance)



It was the default when documentation was missing to list it as a boolean, this should be fixed.



I'll have a look.

Sorry, I do not understand what you are saying.
1) The client should not crash, so it is a bug, right?
2) What should I do to fix it at least by replacing my maps? I have no idea what .ent files are and "locate" shows them only under nquakesv/qw/maps/ctf/.

Tried the latest build (released 3rd of may).
Something is wrong with the sound, it's subtle but the location for the sounds are not where they are supposed to be.
On dm4 for example, if I stand outside ssg room, next to the gl and look towards the quad wall and a player spawns at ssg, the sound plays like he is to my left, sounds a lot like a mega spawn.

Should also say I can tell the sound is different when rocketjumping (don't know what is different, it just is). It's fine on an older version (r4752~590beea).
I'll stick with that until fixed.

(Edited 2016-05-05, 08:57)

O_o

The switch to using SDL2 caused increased sound latency compared to ezquake 2.2, so the sound code was changed significantly to fix it. It no longer uses mix-ahead, and should instead output sound as your sound card driver demands it. You can play with the s_desiredsamples setting, ideally that would be as low as possible for lowest latency, but higher values may get you back to what you're used to.

I've added an extension to the /play command so you can create sounds as though a player was spawning at various locations, to let you test. If you could try again tomorrow with the latest nightly build (that doesn't exist yet):

• Stand at GL and point at quad, in the direction you are discussing above
• /play misc/r_tele4.wav @ -64 512 -296
• /play misc/r_tele4.wav @ 112 -1136 -104

That should let you compare what someone spawning at ssg & mega sounds like.

I played around with that s_desiredsamples and managed to get it to a good enough state, it's not a perfect match to how it used to sound but close enough.
Thanks.

Ruleset qcon bug / malfunction: as it disallows playing sounds, the KTX countdown sound isn't playing.

I'm lazy so I didn't read the thread, but I've noticed that the sound is not predicted correctly in ezquake 3.0 in spectator mode: if I have 100 ping and I'm speccing a guy shoting a rocket, the sound will appear 100ms after it is supposed to appear. It is very disturbing, also because the sound does not even follow the pov of the player but is generated at the place he shot.

Any way to fix that?

Is this still happening in build from 19th May? (459b00d)
Apologies, meant to post on here about it.

With the recent development in the QW scene (http://challonge.com/TrueDamage), we will need a "ezQuake 3.0 to 2.2 downgrade info" thread :|

Yeah, I think the admin is slightly misinformed... See his comment responding to Dirtbox asking about Ezquake3. Someone should tell him that Quakecon will be using an "official" ezQuake 3 build, not 2.2.

BD