@Abc, Weird that it didn't changed anything... Too bad, thanks for the feedback anyway.
How much computers where you using when it crashed, 4 or 3 if I get it? Did you get the issue with only two players?
Abc wrote:his mouse stopped working
How could he notice that? Was he able to get out of fullscreen mode?
What is this "name wheel glitch" you're mentioning?
Posted: 13 Apr 2015, 02:32
Abc
Test Report log:
1) me host: passion was not able to race because of connection problem with me (irrevelant), we were 4 racers: me, mmud, mati and sc. (5 with passion)
vsync was off for all then we alternated.
all went fine until i stress tested restarts: mati crashed in the middle. then he crashed on late join. we ended up being stuck in a restart loop with perfect sync in ghost town 2 (i think that something triggered host calculation routine and it gone nuts)
2) mmud host: 6 racers: he, me, passion, sc, dolo and mati.
vsync was on for all and we alternated. i froze on stress restart, took longer to connect and i crashed again on track change, similiar to mati (coincidence?) and we just stopped because we wanted to.
3) SC host: less relevant: same players as 2), vsync alternated. Very unclear about the crashes: mmud said he crashed in town2, mati crashed in supermarket 2 with mmud, i saw them not responding on muse1 start. (no stress restart)
in the end we were mmud, me and mati and sc has left us with no host.
Posted: 13 Apr 2015, 02:38
Abc
jigebren @ 12 Apr 2015, 05:55 PM wrote:@Abc, Weird that it didn't changed anything... Too bad, thanks for the feedback anyway.
How much computers where you using when it crashed, 4 or 3 if I get it? Did you get the issue with only two players?
Sorry to hear, i believe that it may have to do with network conditions and its not lag but packets. (as your simulation didnt yield any results) Correct?
we were two players but 4 computers (technically 4 racers) (i mentioned in the report)
jigebren @ 12 Apr 2015, 05:55 PM wrote:
Abc wrote:his mouse stopped working
How could he notice that? Was he able to get out of fullscreen mode?
yes, he noticed, he wasnt able to get out of fs (actually it wasnt in fullscreen, it was just the window being drawn by GDI++ on top because its a fullscreen window(not responding applications dont obey to window status unless they respond)). keyboard appears to work, alt tab and control+alt+del, even task manager is usable as long its always on top. (as i experienced) its a freeze not crash and he decided to restart. mmud has simply logged off on his crashes. (re-volt unable to kill because of UAC: non-elevated cant kill elevated processes like a standard user)
jigebren @ 12 Apr 2015, 05:55 PM wrote:What is this "name wheel glitch" you're mentioning?
name wheel turns too fast without vsync.
All speedy glitches due to high fps are fixed.
Posted: 13 Apr 2015, 02:57
mmudshark
A little addition to Abc's post about our tests.
When SC was hosting, both mati and I were stuck in "restart" loop on ghost town 2 when sc went to next track.
None of the other racers had this problem.
This was the only issue I had during any of today's tests.
Posted: 13 Apr 2015, 04:51
jigebren
A little addition myself. I made a test using 3 computers this time: Two XP and one Win8 laptop...
First attempt Win8 was hosting (using 15.0330 build), once I think I tried to kill one revolt instance on XP during "Host Restarting..." screen, and the Win8 host hanged just as described in this topic. I had no tools on this computer to go further so I didn't get more info, but what's weird already is I'm quite sure it was the host that hanged (while most reports here lead us to think only players could hang).
Second weird thing it that we can't really kill re-volt during the Sync screen in fact (even if it isn't hanging, it will wait until the Sync is done to respond again), so I'm not sure when exactly did I killed the revolt instance in my first attempt (as I don't remember I had to wait before revolt process gets killed).
I tried a lot of times to get the same glitch afterwards but I was not lucky (or I was too much - it's a matter of point of view ) and couldn't get it hang again...
So unless there's a few racers ready to for an online test in the coming hours (I can install RVHouse to chat, etc. but to be efficient I'll have to join by IP, not through RVHouse), the only option that remain will be to release a build with more complete revolt.log info to try to pinpoint the actual location of the bug.
Posted: 13 Apr 2015, 05:54
Abc
the loading screen is designed in a way its "frozen", never noticed the taskbar flashing when you press alt+tab on the loading screen and once it finishes, responds and does minimize itself?
also: 7pm gmt (~3pm GMT -3) is the best time when it comes to racers availability.
Posted: 13 Apr 2015, 21:26
Abc
I just realized that there's a crashing chance on one player in 3 restarts (on the 3rd restart = freeze) and it's a random player and at a given condition (it cant be replicated with simulator)
Even more odd is that the host haves less chance of hanging than the players (as seen in jig tests)
the best we can do is have huki check what they changed between 15.0211 and 15.0325 (something broke there). i think that tests are no longer needed unless what said has been verified and a new build reverting the bad change is out, and i think that one test will be enough
Posted: 13 Apr 2015, 23:34
Huki
Abc @ 13 Apr 2015, 09:26 PM wrote: the best we can do is have huki check what they changed between 15.0211 and 15.0325 (something broke there). i think that tests are no longer needed unless what said has been verified and a new build reverting the bad change is out, and i think that one test will be enough
It's very likely that the problem is rather with the original code.. if something is wrong with our changes it would have been fixed already.. but debugging existing buggy code - something that worked before but mysteriously broke now, is not very easy...
Only a few more tests with better log info can help. I'm posting a new test build which will log further info. Just upload your log file after getting a hang, like sebr did earlier. Here you go: rv1.2a15.0413_test. We'll be awaiting the results...
Posted: 14 Apr 2015, 07:49
Abc
Okay, how come until 0325 everything was working great?
Posted: 14 Apr 2015, 10:53
mmudshark
5 of us tested the 15.0413 build tonight in GameRanger (not everyone can play in RVHouse) and after a complete round of each track at least one time there were no problems.
We raced a few in mirror mode and started one custom track. Tracks were restarted in the middle of the race, one track restarted a few times as quickly as the host could do that and there were absolutely no issues. All players synchronized quickly.
Some players mentioned the steering was a little different for them.
Hopefully more tests tomorrow with the same result.
Posted: 14 Apr 2015, 15:14
Touriga
Hi all,
There is one thing I have, visual memory, I can assure you all that I notice a difference for better while playing in gameranger, game runs faster and smoother to control.
I've tested 0325 / 0330 on gameranger but unfortunately with 2/3 players, no crashes ever occurred.
Could somehow ports used by these versions have changed ? Or with more players a port that isn't open and used before is now being used in these versions ? That could explain the synching problem no ?
Just my 5 cents.
Thank you Huki and Jigebren for all effort in getting this solved.
All the very best,
Touriga
Posted: 14 Apr 2015, 18:18
jigebren
Abc @ 14 Apr 2015, 03:19 AM wrote:Okay, how come until 0325 everything was working great?
[HIDE=Offtopic]Sorry but you'd better stop with that kind of stupid question / suggestion now... it's spoiling the topic, and it just show that you don't want to get how software development works. Have you ever wondered why ANY software has a bug report section, and an alpha / beta stage before a release can be considered stable? Firefox for example was working nicely for me up to 26, then a nasty glitch appeared in 27. Why? Because some parts has to evolve and they can longer fit with older code (or older hardware) afterwards. I made a detailed bug report, I was a bit upset that it didn't get much attention from the devs even though I was clearly not the only one to get it. But I didn't get mad at the Moz devs that are providing me a nice browser for free, and I was not going like: "hey, it worked for me in 26, now it doesn't in 27, just take it back to 26 you fools or I'm going to use Chrome". BTW, now Firefox is 37 and I'm still using it...
About your suggestion to check what has changed between older bugfree version and new one... erm, what do you think we're doing actually?
And revert the change and to get back to a bug free state... Ok, so you want us to release 15.0131 a second time? We're going to make rapid progress that way.
For a possible explanation, since that was your request: Here we had to take care of eg. the high FPS issue, this is something original devs likely didn't have in mind in the late 90s. Some code may not be safe from the beginning, but some recent change can increase the probability to reach a bug. Maybe this probability was even null before. Still the old code can be the real one to blame...
OS are evolving, hardware are evolving, so v1.2 code is evolving, and new bugs are likely to occurs at each new release. It is unavoidable. Either you get it and understand that the only way to go is tracking new bugs until they're solved. Or you think that everything has to remain as it was before, in that case just forget about v1.2 and go back with v1.1 Re-Volt.[/HIDE]
@Touriga
Just wondering BTW, but are you sure all reports done here after 15.0412 have actually been done with all players using only 15.0412? I'm asking since it has not been clearly stated in the reports, and we didn't make this build incompatible with older ones (and it's quite easy to mistakenly launch another build in eg. RVHouse).
Touriga wrote:Could somehow ports used by these versions have changed ?
Hmm, I strongly doubt about it.
Posted: 14 Apr 2015, 21:35
Abc
nono, its just that its odd that the code breaks just now. and I actually meant rollback the network code to 0131's.
I also needed a comprehensive answer.
I'm sorry to spoil but I think it was necessary, at least for me, i'm trying to understand what broke.
EDIT: I'm sure and confirmed that everyone on my tests were using the current test version. (at that time)
Posted: 14 Apr 2015, 22:05
Touriga
Jigebren,
No, I'm not sure that everyone was using the same ver and furthermore I suspect that some were not.
Will try to run a proper test today ... If I can gather the right people.
Thanks.
All the best,
Touriga
Posted: 15 Apr 2015, 03:35
Abc
Huki @ 13 Apr 2015, 03:04 PM wrote: Only a few more tests with better log info can help. I'm posting a new test build which will log further info. Just upload your log file after getting a hang, like sebr did earlier. Here you go: rv1.2a15.0413_test. We'll be awaiting the results...
How long it has to hang to have a successful log?
Posted: 15 Apr 2015, 03:47
jigebren
Abc @ 14 Apr 2015, 11:05 PM wrote:How long it has to hang to have a successful log?
What?
EDIT: You don't have to wait. As soon as you can notice it's hanging it should be ok already.
Posted: 15 Apr 2015, 03:55
Abc
jigebren @ 14 Apr 2015, 07:17 PM wrote:
Abc @ 14 Apr 2015, 11:05 PM wrote:How long it has to hang to have a successful log?
What?
like a crash dump, you have to wait it to finish to correctly trace a crash.
Posted: 15 Apr 2015, 03:57
mmudshark
Touriga, mati, Passion and I just finished several tests using the 15.0413 build.
We did the same tests as the 5 racers did yesterday. Restarting the same track repeatedly, restarting in the middle of a race and again there were absolutely no issues with sychronizing players.
Testing was done in both rvhouse and gameranger.
Everyone said the gameplay was smooth and the radar even worked better than in previous versions. Spectating mode still doesn't work. When someone late joins the cars racing aren't moving but show when they get hit with a weapon.
This is looking very promising.
Posted: 16 Apr 2015, 23:43
Touriga
PLEASE DELETE THIS POST
Could not reproduce and it must have been a placebo effect of changing versions.
Posted: 17 Apr 2015, 01:51
Huki
mmudshark @ 15 Apr 2015, 03:57 AM wrote: Touriga, mati, Passion and I just finished several tests using the 15.0413 build.
We did the same tests as the 5 racers did yesterday. Restarting the same track repeatedly, restarting in the middle of a race and again there were absolutely no issues with sychronizing players.
Testing was done in both rvhouse and gameranger.
Everyone said the gameplay was smooth and the radar even worked better than in previous versions. Spectating mode still doesn't work. When someone late joins the cars racing aren't moving but show when they get hit with a weapon.
This is looking very promising.
Thanks for the feedback. Spectating support is disabled in this test build, that's why it doesn't work.
About the hang, we haven't fixed it actually, we just added various log messages to try and narrow down the source of the bug.. but it seems the bug has gone into hiding for the time being. Good news for the players, bad news for us.
If it does freeze at the sync screen again with the 0413_test build, be sure to get the revolt.log file and upload it here.
Posted: 17 Apr 2015, 12:13
Touriga
If nothing has changed in this version except logging ... that is just so strange.
Last week centralized network on RV-H was down, we had just P2P network, could this be the reason why the bug is hiding ?
Can't remember if we tested this version with just P2P but we could try, we would need to block ip acquisition by RV-H. (blocking checkip.dyndns.org).
Just my 5 cents.
All the best,
Touriga
Posted: 17 Apr 2015, 23:20
Abc
That doesn't make sense because by ip it crashes too.. (as jig has tested?)
KADC has nothing to do with directplay.
Posted: 18 Apr 2015, 18:05
Touriga
Abc I almost feel tempted to say: I AIN'T TALKING TO YOU.
But I'll keep it for next time...
Posted: 19 Apr 2015, 02:03
Abc
Touriga @ 18 Apr 2015, 09:35 AM wrote: Abc I almost feel tempted to say: I AIN'T TALKING TO YOU.
But I'll keep it for next time...
I'm just trying to help.
please discharge yourself by punching yourself with that hate.
Posted: 19 Apr 2015, 03:23
mmudshark
We did several tests today. RV-Passion hosted the first set with Dolo, mati, Sebr, and I racing. Santii late joined later.
There was no freezing at synchro in any of the races and we used 3 different hosts using v1.2 only and then All Versions.
The only issue today was when the host restarted the same track a few times. When Passion hosted and restarted twice, Sebr and I had no countdown and could start early. This was fixed by going to the next track.
I am not sure if this is relevant but we did more tests to recreate the "No countdown" problem by restarting the same track a few times.
Never a freezing issue no matter who hosted and which version they used while hosting.
Beta hosting all version > alpha client = nothing
Alpha hosting all version (series 13 14 & 15) > beta client = countdown issue
Alpha 15.0413 1.2 only hosting > Alpha 15.0413 client = countdown issue.
No revolt.log created for me so nothing to post.
Posted: 19 Apr 2015, 11:52
Abc
no hang/crash, no log
I guess the condition has went away
and it made a comeback with countdown issue instead of restarts, again, I would like to review the network code
Posted: 19 Apr 2015, 15:10
Dolo
It is an old issue that i already saw when i played with the beta in a game hosted in all versions mode with an alpha (like example the 13.0820 beurk !)
I don't want really to help you because i know what to do (reset the network code to the version 1.1 and removing some little issues like you did with the beta and may be the alphas 11.4xx)
No need optimizations with an old protocol and with an old game like revolt. The new large bandwidth today with this same optimizations changes too many aspects that obviously you are not comfortable because you do not test in real conditions. (i mean with many players in the same session)
I share anyway a little video about the issue that mmud has described. http://vimeo.com/125345669
Mmud on 13.0820 (host in all versions mode)
Mati & Passion on 15.0413
Me on Beta 11.0208
Posted: 20 Apr 2015, 00:19
Abc
12 players will be always laggy.
Again, something broke in latest builds, it may remain unknown.
Dolo: if you know what to do, do you have the tools, the source code and the skills required to do your crap? (we know you're noob so I guess the answer is no, haha)
Posted: 20 Apr 2015, 01:12
Dolo
It would be useless to have skills without source code, look at you...
Posted: 20 Apr 2015, 02:09
Abc
Dolo @ 19 Apr 2015, 04:42 PM wrote: It would be useless to have skills without source code, look at you...
the source code is there, it's leaked xbox thing.
[HIDE=Warning:][Vaid]: Dolo's posts were quite polite this time and yours, abc, were not. Looking at how you wrote to him, it looks like you are the one who is acting "childish".
I'd suggest you to avoid using these sentences. I mean, they don't really add anything useful to the context of this conversation. Plus, you're at high risk to get a warning level.
abc's forewarned quote wrote:Look at your silliness. I'm the nerd here, not you, so shut your childish mouth, fool!
[/HIDE]
Posted: 20 Apr 2015, 05:55
Phantom
[HIDE=Offtopic @Abc]
Abc @ 19 Apr 2015, 05:39 PM wrote:Look at your silliness. I'm the nerd here, not you, so shut your childish mouth, fool!
Abc your level of aggression is impossible to stand and your answers are unbearable to read. You hardly ever treat somebody with respect and you pretend that by being "the nerd" you should be respected? People won't respect you just for that. With your attitudes you only gain people's disrespect, you're being silly and childish but you prefer to tell others they are silly and childish...
You love acting like if you know a lot about things you clearly have no idea about. This is not how it works. If you have behaviour problems there are many ways to treat that, but you can't be here offending people for free. We may respect you if you at least tried not to be so offensive when you answer and, if possible, stop acting like you are the guru of something you're not.[/HIDE] [Vaid edit]: Marked as offtopic.
Posted: 23 Apr 2015, 22:43
jigebren
[HIDE=Offtopic]@Phantom
+1, nice and relevant explanation.[/HIDE]
That is the most highly inappropriate and ridiculous response I've ever seen to someone trying to make you see sense from your wrongdoings. As I said at RVL... You just don't learn, do you?[/hide]
Posted: 24 Apr 2015, 10:41
Abc
[HIDE=Offtopic]
Skarma @ 24 Apr 2015, 12:22 AM wrote:
That is the most highly inappropriate and ridiculous response I've ever seen to someone trying to make you see sense from your wrongdoings. As I said at RVL... You just don't learn, do you?
Whatever... Who cares.[/HIDE]
[HIDE=Warning][Vaid]: There, here's your warn level that you really wanted. I won't bother lecturing you anymore. It will be pointless and waste of my time. Y o u - d o n ' t - c h a n g e .
This was just a warning for now, so it won't really do anything, but I have more surprises for you in the future, just in case.
By the way, I've removed your above post, it hurts my brain cells. You can still find the quote in Skarma's post.[/HIDE]
Posted: 17 May 2015, 02:48
mmudshark
We ran a full round of stock tracks today with 6 racers (everyone using the 15.0413 version) and there were no problems.
There were no stress tests, just racing like we usually do.
This may not be relevant but it's still an issue:
15.0131 is still the most popular version for most players but restarting the same track still causes the "no countdown" problem for some racers.
The track GBA Rainbow Road crashes for some when using the 15.0131 version. Passion and I raced it today and he crashed (I forgot to ask if the game hangs or they're booted back to desktop) each time and the same thing happens to SCracer. When we switched to the 15.0413 version, the track played properly.
I think the 15.0413 version with spectating working properly would be the most popular version for most racers other than the few that insist the beta is the best.
Posted: 17 May 2015, 16:07
RV_Passion
mmudshark @ 16 May 2015, 10:18 PM wrote: Passion and I raced it today and he crashed (I forgot to ask if the game hangs or they're booted back to desktop) each time and the same thing happens to SCracer.
After the loading screen, my laptop screen was black for a moment then it come back to desktop with the message: revolt.exe no longer works, or something.
Posted: 17 May 2015, 16:55
Touriga
mmudshark @ 16 May 2015, 10:18 PM wrote: I think the 15.0413 version with spectating working properly would be the most popular version for most racers other than the few that insist the beta is the best.
I totally agree with MmudShark, when players lag 15.0131 even though popular makes radar useless and gameplay suffers.
Posted: 17 May 2015, 17:36
sebr
mmudshark @ 16 May 2015, 10:18 PM wrote: [...]The track GBA Rainbow Road crashes for some when using the 15.0131 version.[...]
I think i found something wrong with the W file of this track.
About 12ko of extra data
I send a PM to Jig about it
(wait and see)
Posted: 03 Jun 2015, 02:20
Santiii727
I do not know if it was spoken before but all versions have trouble putting tracks in reverse.
Mirror Tracks works perfectly
I think a room with all versions, I need to upload 1 video
and something strange to wait player
I have evidence here
Phan Have 0131
[youtube][/youtube]
[youtube][/youtube]
This happens in 0413/0418
Thx
Posted: 04 Jun 2015, 08:13
Phantom
Go to 1:20 on the second video to see the actual issue.
Posted: 04 Jun 2015, 12:41
VaiDuX461
I'm not exactly sure, but there's possibility that you can't force older game versions to properly change reversed track type in mid online session/room. They never supported this anyway.
Unless it is possible to remotely do a full track restart on old versions (does v1.1 even have a full restart?), but I doubt this was programmed for multiplayer.
Full restart reloads all track files, including .inf definitions, which is car start position and rotation, for example.
But that's probably not the solution. The host can't send the information to use 'different track type definitions' (in this case, use STARTPOSREV and STARTROTREV instead of normal and vice versa) to older game versions, or just can't accept it. And that requires restarting the online room/session.
Reversed tracks requires to load different files and .inf definitions, so they are more complicated, while mirrored does not.
Posted: 04 Jun 2015, 16:12
Abc
track selector doesn't support reversed and mirrored changes on the fly, since 1.0.
Adding the track search in 14.0306 has clearly shown the game's inability to change track modes on the fly (between races) it was never supported, hence that embedded selector.
) and couldn't get it hang again...
but debugging existing buggy code - something that worked before but mysteriously broke now, is not very easy...
.i. 
