Our ReVolt Pub

The oil issue was fixed and also earlier versions cannot join but there was a major problem when 6 of us raced using the 15.0330 version this morning.

1. Fatal crash (revolt crashed) for some racers during synchro at start of a race.
2. Fatal crash during synchro when trying to rejoin lobby session.
3. 6 racers and each of them had at least one fatal crash during the races. Some racers had more.

Touriga hosted again later with SCracer and I.
With the same host and only 3 of us racing there was only one problem unlike the earlier races with 6 racers. My revolt froze once when the host restarted the same track. This was done several times and the issue only occurred the one time.

mmud

I crashed a few times but never during hosting, as mmudshark said during synch.

But my crashes were not only revolt but also windows, had to reboot pc.

But one thing has to be said, while the 3 of us were racing the gameplay was very smooth.

Thanks for all the great work.

All the very best,

Touriga

Ouch, that doesn't sound good.

We'll try to double check the code but I'm not sure that'll be enough to find out the reason of that issue. Don't hesitate to give any more info if you have (like eg. confirming that the host itself never crashes, or if you find any way to make it reproducible).

Since there was also the code change to support high FPS, maybe it could be worth to report whether you're using VSync or not (I doubt it has any effect but just in case...).

Sounds bad indeed, thanks for the report. Does the crash only happen at the synchro screen? The black screen that says "Synchronizing..." or "Host restarting..."? Never during the race?

Maybe mmud has a better memory than me, crashes were never during race but now you got me with the synchronizing or host restarting, can't be sure.

My V-sync is off.

Myself as host never crashed.

I would say that with 6 players with 0330 shouldn't be very difficult to reproduce, yesterday each race we lost 1 or 2 player due to revolt or pc crash.

I'm on Windows 7.

I'll try to find 6 for a test this afternoon, wish me luck.

As i have observed, since i only run Re-Volt online in multiplayer mode, this is an issue with this newer release, and can happen with any number of players in an online race...and only happens when the host restarts a race, or selects a new track to start.

it happens as the count down is waiting for the players to sync and the names of the players waiting are displayed, and blinking cars are reserved and displayed on the starting grid for the player. Once the waiting period expires, the players are gone and so are the blinking cars from the starting grid and the count down begins.

the starting grid is affected as the missing cars do not affect where the other players start from and so the front row may be missing at the start of the race if they were the players that got dropped.

if i had to guess, i'd start by looking at the blinking car mod that was made recently. Maybe if we try removing that function as a test. Functionally, it's very rare to see cars blinking in the game anyway, so i don't think having this function on or off changes the experience in online races.

Good luck, please let me know if there's anything i can do to help test.

mmudshark @ 29 Mar 2015, 10:35 PM wrote: The oil issue was fixed and also earlier versions cannot join but there was a major problem when 6 of us raced using the 15.0330 version this morning.

1. Fatal crash (revolt crashed) for some racers during synchro at start of a race.
2. Fatal crash during synchro when trying to rejoin lobby session.
3.  6 racers and each of them had at least one fatal crash during the races.  Some racers had more.

Touriga hosted again later with SCracer and I.
With the same host and only 3 of us racing there was only one problem unlike the earlier races with 6 racers.  My revolt froze once when the host restarted the same track.  This was done several times and the issue only occurred the one time.

mmud

I know that balint and me had the same issue with 15.0325 (i was never host)

With 3 players everything is ok

with 5 players the crash only happen at the black screen that says "Synchronizing..." or "Host restarting..." and Never during the race, But if I try to late join I can spectate the race without crash, and the next "Host restarting..." = crash

I didn't tried 330 yet

note : win8 64bit with "emuatefullscreen" so revolt always work as v-sync off for me

Win7 64 bit. I've always raced with VSync off as the races are smoother for me.
My crashes happened at "Host restarting..." and I don't recall a crash during a race. Others crashed at synchronizing before the race started. At least that's what my screen showed.

We ran several races this morning (Mar 30) with two different hosts and didn't have anyone crash. Most of the races were with 3 players but we did a few with 4 racers and there were no problems. We will test more when we can get 5 or 6 racers.

Thanks for the clarifications...

I can just add that I tried myself in a 2 players LAN session, and even with an artificially very degraded connexion (bandwidth: 20Kbit/s delay: 500ms, 1 packet dropped out of 4), I get the blinking/ghost for remote car, but I could restart same or another race and join late the session without any crash (in one case the remote host car was not moving when spectating after joining late, but even in that case the host restart worked).

My only conclusion so far is that either it needs at least 3 players to occurs, or the 2 computers I used are too slow to make it happen and it has something to do with high FPS.


We could try to remove the blinking car indeed, but I think this is just an indicator of something already going wrong rather than the real source of the crash.

Out of curiosity it could be nice to try with VSync turned ON, honestly this is almost a shot in the dark, but at least it would remove this doubt (I do understand though it's not easy to ensure that all the players in the session used it).

I would like to advise that a fps limit of 60 should be hardcoded, the game is designed to run at 60 and lower speeds, beyond causes all kind of glitches, signals appearing in slow motion, flashing, they appear in wrong places, camera turning gets sluggish or too fast, spinning wheel is "flash gordon", oil introduced with 0325, crashes introduced in 0325. (my guess is that attempting to optimize the high fps things broke some other things / caused more bugs than fixes)
Also this is a very common issue: for example, having an insanely high fps on a shooter like half-life causes very low jitter but insane lag (unable to move). re-volt engine is designed to run at 60 fps, only UI and "scripting" are affected by >60 fps, not the rendering.

Advice: Hardcode a max of 60 fps (that ensures no glitches and possibly no tearing (most of the time the tear shows anyway) and reverting the novsync optimizations.

btw, the "ghost" car in multiplayer is good and a bit awkward because that one is used in time trial !

@Abc: As for me I would like to advise you that you don't pretend to know what you're talking about when you obviously don't... First the game is not "designed to run at 60 FPS". For your information, the internal physics engine run at 125 FPS. Second, our "attempt to optimize the high fps" is precisely hardcoding that fps limit, so... Now you see how ironically inconsistent it is to have both your advice (which comes a bit late) and your guess that it "brokes some other things / caused more bugs than fixes" in the same post.

And all the glitches you mention are precisely what should already be fixed by this last build (but have you only tried, or even just read the changelog?).

As soon as we get 5 or 6 racers for our next test we will be sure that everyone has VSync ON. Eliminating any "possible" causes will eventually lead to a solution.

p.s. I don't see why fps would suddenly now be an issue when it hasn't been for the last 14 years or so.

jigebren @ 31 Mar 2015, 12:35 AM wrote: @Abc: As for me I would like to advise you that you don't pretend to know what you're talking about when you obviously don't... First the game is not "designed to run at 60 FPS". For your information, the internal physics engine run at 125 FPS. Second, our "attempt to optimize the high fps" is precisely hardcoding that fps limit, so... Now you see how ironically inconsistent it is to have both your advice (which comes a bit late) and your guess that it "brokes some other things / caused more bugs than fixes" in the same post. 

And all the glitches you mention are precisely what should already be fixed by this last build (but have you only tried, or even just read the changelog?).

Ok, you're right, I'm sorry and surprised. i didn't intend to pretend to know (i know my knowledge limits) and sorry for not do further testing on the new builds. Call my post speculation (and partially truth) well, its interesting that the advice comes late and has been already used, still it's pretty weird to have these bugs when they didn't ever happen before.
FYI, i think i have the highest fps of the whole community (Anybody can reach <800 fps?)
I'll just test then.

Sorry jig but some days it's difficult to get people together at the same time for the tests.
Today was only the 2nd time we had more than one racer crash during the game. Passion hosted with Shur, Touriga, and I playing. The plan was to start with vsync OFF as usual, do some races to see if someone crashed, and then try again with vsync ON. After the first race I crashed at "Host restarting...". When I rejoined and host restarted, everyone but Passion (the host) crashed.
Shur did not return from the crash and Touriga had to leave so we did not have a chance to retest with vsync ON.
My car handles very differently with vsync on and I'm not the only one who has said this so I'm hoping this isn't the cause of the issue.
Passion, Touriga, and I all have Win7 64 bit. I didn't get a chance to ask Shur which OS he uses.

Thanks for the tests so far.. we'll wait for the tests with v-sync on, then we'll have some clue.

mmudshark @ 3 Apr 2015, 03:58 AM wrote:My car handles very differently with vsync on and I'm not the only one who has said this so I'm hoping this isn't the cause of the issue.

Was this the case even in previous versions or only the latest release?

Racing with vsync ON has always been more difficult Huki in all versioins. It is the same for everyone else in today's races as well.

6 racers vsync ON
Passion hosting again. Kosztya, matiwrc, Santii, Cosmo, and me.
mati and Kos crashed at first race during syncro.
Cosmo crashed the next race when "host restarting..."
Mati, Santi, and I raced the next track, but Santi crashed when we tried to restart the same track.
Only Passion (the host) and I didn't crash this session.

We did another test a little later with 7 racers.
The host was using the beta version and the rest of us used the 15.0330.
No crashes the entire session. We tested with vsync OFF for some races, then ON.
Restarting the same track and no crashes.

mmudshark @ 4 Apr 2015, 01:13 AM wrote: We did another test a little later with 7 racers.
The host was using the beta version and the rest of us used the 15.0330.
No crashes the entire session. We tested with vsync OFF for some races, then ON.
Restarting the same track and no crashes.

Interesting.. what if everyone including the host has the latest version, but the host uses all-versions mode? Is there any crash then?

Well, think I have found a bug in the online code that might have something to do with the current crash as:
- it should only affects the players but not the host.
- it was introduced in the last release.

This bug should occurs after a player - that is not a spectator - is leaving, then all other players (excepts the host) can try to access an array beyond its limits. Don't know if this situation (a player is leaving) actually makes sense to those who faced this bug already... I'm afraid I don't feel like it explains everything so far though, so maybe it fixes only half the issue, it's hard to tell for now.

Expect another release soon anyway.

Racing with vsync ON has always been more difficult Huki in all versions. It is the same for everyone else in today's races as well.

That's definitively weird, as the last build should had smooth over a lot of VSync ON /OFF differences.

I'm quickly posting a test build that is supposed to fix the crash. Download it here. It is recommended that everyone in the session use this build. And of course, use v1.2-only mode as usual. Let us know if this one definitely fixes the crash or not (there is a small chance it won't..)

I was in two tests of the 15.0404 so far today with all racers using this version.

The first was only with Touriga (hosting), RV_Passion and myself and I didn't notice any problems.

The 2nd was with Dolo (host), Cosmo, Passion, Balint, and myself.
Cosmo crashed (revolt and computer) at sync on the first track. Balint crashed at start of the next track during sync. He rejoined as we were in the middle of the next race so Dolo restarted and Balint crashed again at sync.

Just to try to get a better idea, can you give a rough estimation of how long it takes to crash after the black sync screen appears? Right at this screen, or rather something like after 15-20 seconds?

jigebren @ 5 Apr 2015, 03:52 AM wrote: Just to try to get a better idea, can you give a rough estimation of how long it takes to crash after the black sync screen appears? Right at this screen, or rather something like after 15-20 seconds?

Any answer about this could help us: does it hang at the sync screen for a really long time (more than 10 seconds) before crashing, or does it crash immediately, within the usual 5 seconds of sync time?

It's hard to give an answer
we get the black screen with "syncronising" or "host restarting" and nothing more
after a long time we need to kill revolt with "CTRL + ALT + SUPR" because "ALT + TAB" don't work

I can't tell you if it's appen within the usual 5s or not because i never get windows message "revolt not responding"

Wait, you mean there's no crash actually?

It's quite different then (even though we still don't have a clue about it in that case either). A crash gives the idea that the application is suddenly killed with (or without) and error message form the OS. If the application just stops responding, even if it hangs the computer, calling it a crash was definitively misleading (in fact the application may still be running in an infinite loop for example).

We'll try to look at it again with that new information in mind...

jigebren wrote:Just to try to get a better idea, can you give a rough estimation of how long it takes to crash after the black sync screen appears? Right at this screen, or rather something like after 15-20 seconds?

Sorry I didn't answer this earlier but I didn't notice it until now as I wasn't around for any tests today.
My answer is the same as sebr's. The application does hang the computer (rather than crash so I have been using the wrong terminology) and other than hard booting or using ctrl/alt/del to "log off" is the only way to "unfreeze" it.
I thought a "crash" was when a program became unresponsive. I checked Event Viewer logs there are no event reports for Revolt.

Weird, that reminds me of an issue I faced with my firewall...

When I launch Re-Volt online mode, if I had not allowed Re-Volt in my firewall already (and I have to do this at each new release since the binary file change is detected), then a pop up is displayed for the authorisation request. But when RV is in fullscreen, the pop up can't be displayed (RV keeps the focus), but at the same time that pop up is also blocking RV (which is waiting to connect to the network). As a result, my computer become totally unresponsive.

[ctrl+alt+del], [Alt+Tab], [Alt+F4], etc. don't work. The only trick I found to regain control in that case was to press the "Sleep" key on my keyboard to force my computer to go in Sleep mode, then I wake it up. Strangely, in some case I think I was also able to simply answer the firewall pop up blindly (using the keyboard arrows and space key) and it worked.

As a solution, during my online tests I kept in mind to only use windowed mode, as in that case I would not definitively lose focus in case of pop up message.

Afterwards Huki made a modification to prevent that issue.

Alpha 14.0208 wrote:    [Multi]
    - Do the multiplayer session start / join in the background to avoid blocking firewall messages.

All this just to say that it could be an idea to run in windowed mode during your test for now. At least if it hangs you may not lose control on the computer.
(and if you're an advanced user you could even try to get some info about the running (and hanging) RV process )

So it sounds like it's not a crash actually, but the game stays at the sync screen for ever, for no specific reason. Very strange, but here's a new test build (download).

This build also logs some basic info about what happens at the sync screen, you can find the revolt.log file in your temp folder (just type "%temp%&#092;revolt.log" in your folder's address bar). Here's a sample: In case this build still hangs, it would be nice if everyone (both host and clients) upload their log file after killing re-volt.

Hey Huki !
Come in RV House, make a room and test the build with the others players !
You will see the log by yourself like that... tssss

Dolo @ 6 Apr 2015, 05:15 PM wrote: Hey Huki !
Come in RV House, make a room and test the build with the others players !
You will see the log by yourself like that... tssss

DEVS Don't race, they maybe don't have time or are busy with life.
don't say they don't care about development, they provide the community the tools required to debug stuff.
now the problem with the community is that they're not geek-like

Test his own work is part of development ...
It's not about racing for fun ...

Hey Dolo ! For once your comment has been of some use. You're stupidity is so strong that you've been able to make Abc sound like a pertinent and mature guy, and that was definitively unexpected. Thanks for him.


Now back to the topic. Once we get the first results from the build posted by Huki yesterday, either it fixes the issue and it's the end of the story (still it was the opportunity to find and fix another online bug), or at least it will give us a better clue of what's really going on.

Yay noobs

Dolo @ 7 Apr 2015, 02:30 PM wrote:Test his own work is part of development ...
It's not about racing for fun ...

I do race for fun when i feel like. please don't tell something that's not really true. i think you got me wrong for being too freak/geek/nerdy.

So does it work with the latest build or does it still hang? We're still waiting for some results.. the log file is not the most important, but we really need to know whether the problem still exists with the last build or not.

lots of players (like me) only do multiplayer races on week-end

you may need to wait a little

Last_Cuban hosting.

First race with only 3 racers. LC, mati and myself.
mati froze at synchronization and had to kill revolt but LC and I had a normal race.

Next race with LC, mati, monty, and myself.
mati froze again at synchronization but LC, monty, and I had a normal race.

mati checked the revolt.log but there was nothing created for today. The only entry was
Compilation date 20:18:25, Mar 19 2015

I hosted the next test with mati and monty joining.
mati changed his monitor to 60hz and did not freeze this time. After the race he set it back to 75hz and froze again.
Touriga joined the test and raced with no problems the first race.
When mati late joined the race with the monitor back to 60hz it froze on synchro again. Touriga froze at synchro this time too.
Neither Touriga nor mati had any new entries in their revolt.log. Both showed Compilation date 20:18:25, Mar 19 2015.

strainge mmud
I only race ofline yet with last test build and get this as revolt.log

Ok, thanks for the feedback. As far as I'm concerned I'm still clueless though...

@mmudshark, do you have an idea, between LC, mati, yourself, and also Touriga, of who has the fastest and the slowest computer (I mean, roughly, as far as track loading time is concerned for example)? Does mati tend to be the slowest?

And out of curiosity, what kind of screen can toggle from 75 to 60 Hz? Is refresh rate an option for some gaming-oriented screens? I presume it's not still a CRT screen (those values are a bit low for a CRT anyway).

@Sebr, I've edited your post as it was still breaking the invisionfree code... I presume the log file would be really interesting only if it shows the "RemoteSyncTimer" entries as in Huki example anyhow. Still I don't know why Touriga and mati both have their log file almost empty.

Too bad there is no log.. maybe it doesn't get recorded when re-volt is launched from lobby. In a short while I'll be sending a build that logs the info inside the profiles folder, that is guaranteed to work even when launched from lobby, so stay tuned for that..

new revolt.log

I start a race with Turbo has host
then Passion, SCracer late join
everything was good

and Dolo join and I crash at "host restart..."

sebr @ 11 Apr 2015, 10:56 PM wrote: new revolt.log

I start a race with Turbo has host
then Passion, SCracer late join
everything was good

and Dolo join and I crash at "host restart..."

Thanks sebr, from this log it appears you crashed at the end of loading screen, not in the sync screen... is that right?
Can you tell me what command lines you used for re-volt? Specifically did you use -sload?
Another thing: was it really a crash, or it hangs forever and you had to kill it?

it was at the sync screen, didn't see loading screen

I use " -emulatefullscreen -nomip" I don't use -sload

was not realy a crash I had to kill revolt

Just a clarification:
When you stalled you were restarting Market1, right? Just to be sure can you confirm it was going on exactly either way 1 or way 2: After playing a few tracks,
1. you played Martket1 just once and it worked,
2. you played Martket1 two times and it worked,
then you restarted Martket1 and it stalled at "host restart..." screen (you didn't even see the track loading screen).

French version:
Ou si tu préfères: aviez-vous déjà fais une ou alors deux parties avec Market1 juste avant de recharger ce circuit et que ça plante?
Et tu ne vois pas l'écran de chargement du circuit après l'écran noir de Synchro (alors que normalement tu le vois bien, on est d'accord?)

sebr @ 11 Apr 2015, 11:32 PM wrote: it was at the sync screen, didn't see loading screen

I use " -emulatefullscreen -nomip" I don't use -sload

was not realy a crash I had to kill revolt

Ok, just to get the sequence clear:
- The host started Market1 --> no hang.
- Then the host restarted same Market1 again --> still no hang
- Then at the next restart (when Dolo joined), it hanged?

1) I forgot that I have a special command line set on Rv-House : -window -emulatefullscreen -nomip
this way I can use all revolt versions

2) we start Market2 (Passion, Turbo and me) and before the end of 1st lap SCracer late join
So Turbo restart Market2 and we finish the race
Turbo restart the race (dolo late join) and it hanged

But i'm not shur that Turbo didn't restart market2 a 3rd time, but start ghost2

@ jig
I wasn't around early enough to do any testing today.

LC, mati, Touriga, and myself all load fast in the lobby and also in races.

Screenshot from mati on pc using XP Pro.

Huki @ 11 Apr 2015, 03:13 PM wrote: Ok, just to get the sequence clear:
- The host started Market1 --> no hang.
- Then the host restarted same Market1 again --> still no hang
- Then at the next restart (when Dolo joined), it hanged?

Dolo is Re-Volt Virus!

Abc @ 11 Apr 2015, 11:06 PM wrote:

Huki @ 11 Apr 2015, 03:13 PM wrote: Ok, just to get the sequence clear:
- The host started Market1 --> no hang.
- Then the host restarted same Market1 again --> still no hang
- Then at the next restart (when Dolo joined), it hanged?

Dolo is Re-Volt Virus!

The alpha developpement is the virus... And i'm the antidote !

All I can say is, for an antidote you don't have much curing effect...

Well, thanks to last feedback we at least have a better idea of where the issue can or cannot to take place. Just found some glitches in the code, that was worth a quick test build.

Download: Re-Volt_15.0412.zip

(This one is just a try to see if it brings any change, the revolt.log file has not extra information so there's no need to post it this time.)

Just for info during my own quick tests I stalled a few times at "Waiting for host", that is to say after the black screen and the track loading, for a reason I still don't know (but I think I already seen that anyway when simulating a very bad network, just as I was doing now as well).

Dolo @ 11 Apr 2015, 07:13 PM wrote: The alpha developpement is the virus... And i'm the antidote !

Who cares....

santi and I tested with our netbooks and desktops. (we have a netbook and a desktop each)
netbooks don't see each other but desktops see everyone, we used 0412 on gameranger.
first we tried toy1 without vsync (first time took a bit to load on the netbook for some reason (im using the same folder on both computers)), all fine until we restarted same track, his desktop crashed, he had to reboot manually because his mouse stopped working. no crashes on toy2 with vsync on.
we stopped and started again the game with vsync on on toy2 without his desktop and his netbook crashed when he latejoined, two more latejoins didn't crashed on the same map.
then vsync off on market2, all ok and we stopped.
can the rvh peeps retest again with this version?done[HIDE=Unrelated glitches ]
we also noticed a glitch with the ghost car in multiplayer: some parts are transparent when they shouldn't, my theory is that it has to do with black being alpha. (bertha wheels and some panga areas viewing from front/behind)
Screenshots
@Jig: separating makes it inconsistent, did you have any issue? (should mention)
@Abc: Don't you have a </br> inserting in the middle of the link? (not in preview, but after posting)
@Jig: Odd, Bug with links system? There's also a space between "com/" and "5UMAFKd", shouldn't happen

also I noticed a slight fps variance which I found weird/odd: on race I had 300-200 fps on nhood1 and then 1600 when finished and 1000 when moving the camera with the race finished.[/HIDE]name wheel glitch is confirmed fixed.
[HIDE=Specs]
My Specs

Santi's netbook specs:
Bangho Suma E11IS2
Intel Atom N455 1.66ghz
ram kingston 1gb ddr3
Intel GMA 3150
winxp sp3 pro 32-bits

Santi's desk specs:
Intel Core 2 Duo E7200 2.53GHz
Kingston 2gbx2 ddr2
gpu amd radeon hd 5670 512mb gddr5
win7 pro x64 sp1[/HIDE]