[QUOTE=MinIsMin;51466473]Just curious: Why is the "TTT Paint Background" fix ([url]https://github.com/garrynewman/garrysmod/pull/1269[/url]) only in the dev branch? It's working fine.[/QUOTE]
because that's how software updates work in general, you just don't push them live
:shh:
[QUOTE=WinterPhoenix;51209827]My time to shine again!
So! As I've [URL="https://facepunch.com/showthread.php?t=1516664&p=50893496#post50893496"]pointed[/URL] [URL="https://facepunch.com/showthread.php?t=1516664&p=50564400#post50564400"]out[/URL] [URL="https://facepunch.com/showthread.php?t=1497739&p=50022428#post50022428"]several[/URL] [URL="https://facepunch.com/showthread.php?t=1497739&p=49553475#post49553475"]times[/URL] [URL="https://facepunch.com/showthread.php?t=1485127&p=48950004&viewfull=1#post48950004"]now[/URL], Awesomium is garbage and the situation is getting worse all the time for it, and Cinema and Media Player specifically continue to suffer greatly because of its close to complete lack of HTML5 Video Support.
As myself and others stated [URL="https://facepunch.com/showthread.php?t=1516664&p=50893496#post50893496"]last time[/URL], we stand in support of [URL="https://github.com/Facepunch/garrysmod-requests/issues/791"]Adding an Alternative VGUI Browser Panel that uses the ISteamHTMLSurface API[/URL].
To reiterate, what we are suggesting is [B]not[/B] that this VGUI Panel replace Awesomium, especially since Garry seems very against that idea, but [B]an alternative, implemented separately to Awesomium[/B], that GMod developers could work with. It would be completely optional.
Now, as you may know from last time, [URL="https://github.com/Facepunch/garrysmod-requests/issues/791#issuecomment-240601079"]Willox responded to the suggestion[/URL], saying [I]"We're not against the idea, but don't expect it for the next update. Asking people to post non-useful responses to any [Garry's Mod GitHub Issues] isn't going to do much. Use the forums for that."[/I]
In response to Willox's statement and as the update has now been released, we would like to once again bring this up for consideration and, as we dearly hope, an implementation of some form that will finally solve this issue once and for all. It remains an issue which, if continued to be overlooked, has the ability to impact [B]at least 1.18 Million Players[/B], which is [B]approximately 10.67% of all the players that own GMod[/B] (based solely on Cinema's Workshop Subscriber Count and [URL="https://steamspy.com/app/4000"]SteamSpy's Stats[/URL] on Owner Count).
So, once again, GMod Devs! Please consider this. I thank you.[/QUOTE]
Coming soon? Please? :(
[QUOTE=WinterPhoenix;51470292]Coming soon? Please? :([/QUOTE]
Apparently after the workshop changes???
[QUOTE=WinterPhoenix;51470292]Coming soon? Please? :([/QUOTE]
Two video services have now broken in the last month that were previously supported in Cinema and Media Player. They were Vimeo and Twitch, and both are no longer supported due to Awesomium's lack of newer JavaScript and HTML5 functionality.
Furthermore, Google Chrome is now (as of version 55, rolling out as of December 2nd) officially shipping [URL="https://9to5google.com/2016/12/01/chrome-55-stable-adobe-flash-html5/"]Flash Player disabled by default[/URL], which will only expedite the process of the web as a whole abandoning Adobe Flash Player in favor of HTML5.
To be frank, Cinema and Media Player are looking to flat out lose support for all their video services very soon if this issue is not addressed. At the time of writing, [URL="https://github.com/Facepunch/garrysmod-requests/issues/791"]This Request[/URL] is the Most Thumbed Up one on the GitHub Repo, with second place, [URL="https://github.com/Facepunch/garrysmod-requests/issues/745"]#745[/URL], having [I]40 less[/I], so why has it not been treated as a priority all this time?
[B]GMod Devs[/B], for the love of god please implement this. It is time.
Thank you.
[QUOTE=WinterPhoenix;51477768]Two video services have now broken in the last month that were previously supported in Cinema and Media Player. They were Vimeo and Twitch, and both are no longer supported due to Awesomium's lack of newer JavaScript and HTML5 functionality.[/QUOTE]
I looked for these services recently. Main error is changed mp4 standard. Firstly it was started in dailymotion: supportable with awesomium flash player aren't seeking video, pseudo streaming failed and video starts from begin. ≈month before Dailymotion don't play more. (off topic: Also Twitch.TV arent't loaded anymore in HTML panels)
I found a way to fix this. We just take direct URL to video and play via own swf player. I've tried to make it with vimeo,twitch,daily motion and other. So I use web agent that support html5 video, get HTML and match URLs to video/use api/grab from player if possible(flashvars). This bet only in theory... Sometimes it will not get video tag, sometimes video aren't played because awesomium are ****(nice :D). I found a small solution: use HTML5 video tag in awesomium(support video tag as can). Same issue: work on 25% videos and not support streams...
So ≈31%(1 standard, 3 custom) of services(6 standard, 7 custom) are work now on my server. Aaand YouTube get same rarely issue with seeking now(auto seeking to start, middle.) In future maybe YT too no longer works...
Just make a full HTML5 evergreen browser in flash. No need to wait around for the devs! :v:
(this situation is ridiculous and the idea of implementing Steam's browser is a good one and should be a top priority)
Just a thought, why using C# or C++ comments in lua?
[url]https://github.com/garrynewman/garrysmod/blob/784cd57576d85712fa13a7cea3a9523b4df966b0/garrysmod/lua/vgui/spawnicon.lua#L275[/url] (I haven't seen them anywhere else in Gmod repo)
[QUOTE=MinIsMin;51479811]Just a thought, why using C# or C++ comments in lua?
[url]https://github.com/garrynewman/garrysmod/blob/784cd57576d85712fa13a7cea3a9523b4df966b0/garrysmod/lua/vgui/spawnicon.lua#L275[/url] (I haven't seen them anywhere else in Gmod repo)[/QUOTE]
Looks like it has been there since Garry pushed it -- he wasn't exactly consistent with his styling.
To be fair I like C multi line comments better, since I can easily type them using the numberpad keys that are next to each other, unlike Lua multi line comments.
[editline]5th December 2016[/editline]
And they look neater in my opinion.
[QUOTE=Robotboy655;51480157]To be fair I like C multi line comments better, since I can easily type them using the numberpad keys that are next to each other, unlike Lua multi line comments.[/QUOTE]
Nearly all code editors show them as code and not as comment because they load the lua sytnax highlighting and not the c++ one.
[B]My opinion[/B]: That makes understanding new code difficult.
[QUOTE=MinIsMin;51480181]Nearly all code editors show them as code and not as comment because they load the lua sytnax highlighting and not the c++ one.
[B]My opinion[/B]: That makes understanding new code difficult.[/QUOTE]
I never said we should be using C comments in GMod's Lua base.
I like having two kinds. I use Lua comments for removed code and c comments for notes.
Okay, I am pretty sure that [i]something in materials is fucked[/i] and I have absolutely no idea what. No matter what I do, the results make NO SENSE, and I can't explain them. Worse still, the results are completely different depending on whether I have anti-aliasing turned on or off.
The worst part is that it's so wacky and random that I don't even know where to point the finger. The only thing I can give you is basically a code dump, which I know just turns people away ("where do I even look? fuck this I'm not bothering" is what goes through my head when I come across it). That said, since I have no other options, here it is anyway:
\lua\entities\sentsandbox.lua
[url]http://pastebin.com/NKTjTk4s[/url]
See line 181 render.DrawScreenQuad() - comment it out / back in to see the difference (I recommend having gmod run windowed/noborder to see it immediately).
Not drawing the material I want to draw: everything is fine.
Drawing with anti-aliasing on:
It somehow remembers the state of the texture from the previous frame? The translate effect grows each frame instead of remaining constant, and if I disable the translate operation I can 'draw' with the pixels.
Drawing with anti-aliasing off:
It draws correctly (without retaining old data) [b]but for some reason creates a stencil for the viewmodel despite me obviously clearing the stencil multiple times[/b].
Video (let me know if your results are different):
[video]https://youtu.be/ys162hG1fMo[/video]
Even if I did something wrong (if I did please point it out), how come the results are so different based on the AA setting?
Edit: side-note, is there a way to blur while preserving alpha?
There's a chance to get _reconnect command to restart game and reconnect to last server
[QUOTE=NeatNit;51480851]... rendering ...[/QUOTE]
Looks like some sort of Z buffer issue to me. Also I have completely no idea why you are juggling around with so many render targets. What exactly are you trying to achieve?
[QUOTE=NeatNit;51480851]stuff[/QUOTE]
You could just push/pop [url="http://wiki.garrysmod.com/page/Enums/TEXFILTER"]manual antialiasing[/url] if you can't figure it out as a bandaid solution.
Watch dogs reference maybe? How many ferraris paid ubisoft for that effect guys
[QUOTE=MDave;51482299]Looks like some sort of Z buffer issue to me. Also I have completely no idea why you are juggling around with so many render targets. What exactly are you trying to achieve?[/QUOTE]
I'm trying to create a more subtle version of the morphing effect here: [url]https://facepunch.com/showthread.php?t=1543612&p=51468266&viewfull=1#post51468266[/url]
I need 2 render targets to construct the morph texture the way I want it.
Thanks for the Z buffer idea, I'll check it out. I definitely AM clearing the depth buffer but maybe it's sneaking back in from somewhere.
[QUOTE=Moat;51482874]Not sure if this has been reported or is a known issue, but net_graph 2, 3, and 4 causes some weird stuff to happen with polys. :v:
I have multicore rendering disabled.
-snip-
Actually, just tested it again while making this post with multicore rendering enabled and it seemed to fix the constant spazzing issue, but still spazzes for a couple frames when they're drawn for the first time.[/QUOTE]
Thanks for the info. Up to now I was certain this was related to queued rendering (and now I've even reproduced it.)
I gave it a quick search, but didn't find anything. Was there a console command added at some point that deleted any .gma's that you weren't subscribed to anymore?
[editline]6th December 2016[/editline]
Literally just found it, menu_cleanupgmas
[QUOTE=smithy285;51486386]I gave it a quick search, but didn't find anything. Was there a console command added at some point that deleted any .gma's that you weren't subscribed to anymore?
[editline]6th December 2016[/editline]
Literally just found it, menu_cleanupgmas[/QUOTE]
Yes, the console even tells you about it on launch when it finds unused gmas ;)
How about the CMaterial System message every second on multicore?
Can't we just ya know... remove it or put it on only on dev mode.
[QUOTE=mcNuggets1;51486646]How about the CMaterial System message every second on multicore?
Can't we just ya know... remove it or put it on only on dev mode.[/QUOTE]
idk what message you mean
[QUOTE=Willox;51486650]idk what message you mean[/QUOTE]
*CMaterialSystem::Lock: 6.03ms
[QUOTE=Willox;51486676]you've got mat_queue_report enabled[/QUOTE]
Ah well, fuck.
Is that default set to true?
It seems like the behaviour of ENTITY:SetKeyValue or ENTITY:GetKeyValues has changed.
Apparently ENTITY:GetKeyValues() also returns other types than string while ENTITY:SetKeyValue only accepts strings, so running the following code will now usually create an error.
[CODE]
for k,v in pairs(ent:GetKeyValues()) do
ent:SetKeyValue(k, v)
end
[/CODE]
I think that's highly irritating and clearly shouldn't be a thing.
Example output of the types of GetKeyValues.
[QUOTE]
>for k,v in pairs(Entity(3073):GetKeyValues()) do print(type(v)) end...
number
number
number
number
string
...
[/QUOTE]
[QUOTE=syl0r;51491424]It seems like the behaviour of ENTITY:SetKeyValue or ENTITY:GetKeyValues has changed.
Apparently ENTITY:GetKeyValues() also returns other types than string while ENTITY:SetKeyValue only accepts strings, so running the following code will now usually create an error.
[CODE]
for k,v in pairs(ent:GetKeyValues()) do
ent:SetKeyValue(k, v)
end
[/CODE]
I think that's highly irritating and clearly shouldn't be a thing.
Example output of the types of GetKeyValues.[/QUOTE]
That was always the case. The only thing that changed was that it shows more keyvalues now, and it shows their real values, instead of showing 0 for all int/float values.
Sorry, you need to Log In to post a reply to this thread.
