suma

Fixed in 82982 ---------- Post added at 09:35 ---------- Previous post was at 09:16 ---------- Fixed in 82985

Yes, the sniper is able to hit. This is the difficult part, which I have tried to describe in the blog - what you see on screen always needs to be thew "authoritative" version, even if it is in fact a result of interpolation. ---------- Post added at 16:27 ---------- Previous post was at 15:48 ---------- Fixed in 82910

Note: The interpolation, while also seen on MP clients, is still a "singleplayer" technology only, it does not attempt to address issues caused by network transfer. The MP side was not touched at all yet.

I attempt following bugs to get fixed: - serious performance or stability issues (game not running) - most demanded by community (vote count on the CIT is the major factor here) - bugs which get me interested somehow, or which I personally consider to be fundamental If you think some issue deserves more attention, feel free to convince more people to vote for it. As "AI warping - especially at distance" is by far the most voted issue, we have dedicated a significant effort to solve it by means of introducing interpolation, hoping besides of fixing the issue it will bring some performance gain as well. Regarding the issue about 360 field of vision for vehicles, I was not aware of it, mostly because it was not upvoted enough - now that I am aware of it, I will look at it, because "I personally consider it to be fundamental"

Could you add repro for those two problems?

Fixed in 82928. The problem was interpolation of the animation phase was sometimes generating results outside of expected 0-1 range when animation was fast, and as a result the generated matrices used for skinned were broken. (In this case the running animation is played 18x per second.)

Great, I think both of you deserve the prize - having repro is even better than having a PIX capture, and you did a great job finding it. I will check the stockroot to see what is available and contact you later this day with a list so that you can choose what you like best - and you nicks will be perpetuated in the changelog with the fix, which I hope to have ready soon :).

Can you provide the camera script? I have seen (and fixed) a similar issue before the patch was published, but perhaps my script was different.

The status bar of the PIX still shows "waiting for the profiling process 3684 to exit". I am afraid I have no idea why it does not work for you. XP Pro should be fine, it is still fully supported by the SDK. ---------- Post added at 23:00 ---------- Previous post was at 22:58 ---------- I think I see the problem: you have "Grab screenshot" as an action for the F9 key. You should change it to Set call capture / Single frame capture Direct3D.

You still have the Arma runnning, it seems (at least I see it in your task bar). You first need to close the game, then you should be able to save or analyze the capture.

Great, this shows you have the PIX attached. I suppose what you call "has saved the pic" is actually that the capture was made (I am not aware of PIX saving any pics). Now you need to close the game, and the rest is described in my post above. If everything goes well, you will have a capture, and we can proceed to the next steps.

Save the PIXRun file. Once you have it saved, you can open it (clicking it should launch PIX) and in the Detail pane (bottom right) click on Render, in the Events pane (bottom left) click on the Frame number in Bold. You should see the rendering result, and it should contain the artifact. If you have all this done, you have your homework completed, and all you need is to send the PIXRun file to me. You can place it on some public filesharing service or FTP or whatever else suits you, or you can drop me a PM and we will agree on some other transfer method (Skype, email...). I am confused you are talking about a picture. You should have a single frame capture. Once you have made the capture, you can close the game and PIX should offer you to analyze the capture. This is a good moment to save the capture ("the run file") in case anything goes wrong.

One last word: before you send me the capture, verify it really contains the artifact by running the analysis then pressing Render tab in the Details pane.

You need this: prog path F:\Program files\Bohemia Interactive\ArmA 2\Expansion\beta\arma2oa.exe Command line -mod=Expansion\beta;Expansion\beta\Expansion -cpuCount=1 -nosound As a startup folder give F:\Program files\Bohemia Interactive\ArmA 2

What did you enter as program path? (Note: you cannot enter options as part of command path, they need to be given separately on the Advanced View / Target program tab in Command line arguments.)

OK. So here is a contest: the one who will first send me a PIX single frame capture of the artifact with any build of his choice starting with 80955 will receive a little present from me (I will give you a choice of a few games by BI, or some merchandise, like "bug hunter" developer T-shirt, depending on what is available here). To make the capture, you need to download and install DX SDK, launch PIX, create a new experiment gathering "A single frame capture of Direct3D whenever Fxx is pressed". As a command line enter what you use to launch the beta, but add -cpuCount=1 -nosound, as PIX has sometimes trouble capturing threaded applications correctly, and sound system has troubles with the long frame times caused by the PIX.

A good question. I will create a poll to learn this in this topic.

Thanks a lot for your testing. I was hoping the 81423 is clean, as this would give me some range to start checking, but I still prefer knowing unpleasant truth now over spending a lot of work in vain because of false assumptions. One option I can think of now is to continue identifying the last "artifact clean" build, as I cannot see any usefull debugging technique. I suggest starting with 79600 and and 80955, and depending on the results we shall see how to proceed (I can generate more interim builds if necessary). The other option could be to find the "most artifacting" build or configuration and try to debug it directly. I was able to repro only a few frames artifact on my system, but if it would be reproducible in a more stable manner (at least a few seconds), it might be possible to do a PIX capture of the artifact. If someone can see it for a few seconds and is willing to do the capture, contact me via PM and I will send you instructions (they will include downloading and installing DirectX SDK).

Warnings from windows that you are running out of memory most often mean the disk on which you have your swap file placed is getting full.

Well, if 81785 already shows the problem, testing the interim builds I have provided has no sense and I will have to prepare new ones in the range 81423-81785. If someone still can do some testing this weekend, I would like you to concentrate on verifying 81423 is really artifact free, so that we know testing that range is really meaningful.

OMAC, it seems for some people it is reproducible quite reliably, even if the reproduction takes quite some time. For I example I am quite confident I am able to reproduce the issue within an hour when playing Bear rising scenario on one of my two computers. If I would play the scenario twice for an hour and I did not see the artifact, I would call such build to be clean of the artifact. If you do not have a repro like this on your system or if such repro is too time consuming for you to be able to contribute, then I suggest to leave the testing to others. Also note I was careful in asking to report "the earliest build where you can see it". Not seeing it is not a 100 % proof the build is OK, but if multiple people will test it and nobody can see it in that build, I think we get recognize the first broken build with a reasonable confidence.

Here is a "weekend homework", if some of you have some spare time to help me with the issue. Here is a list of interim builds in the suspect range where the issue was introduced. Please, test it, and report earliest build in which you can see the issue. I am sorry to ask this from you, I understand it is time consuming, but once I know the exact build number the issue started, I should be able to fix it pretty soon. ARMA2_OA_Build_81869 ARMA2_OA_Build_81881 ARMA2_OA_Build_81882 ARMA2_OA_Build_81883 ARMA2_OA_Build_81890 ARMA2_OA_Build_81894 ARMA2_OA_Build_81896 ARMA2_OA_Build_81897 ARMA2_OA_Build_81909 ARMA2_OA_Build_81919 ARMA2_OA_Build_81920

While knowing what types of objects do it is also interesting (I have a screenshot which seems to show a tank to be a likely cause), the most important is to know which build is the first one to show the issue. I understand testing several betas takes long, as the repro is not quick, but if someone can do it, it will help me greatly.

I think we can confidently say 81921 is faulty. Could some of you try build 81867 to check if you see the issue there or not? Knowing if that build is "good" or not would save me a lot of work, as I would know what areas of code should I concentrate on when inspecting the problem sours.

We have two statements here: The 1-st says the first with artifact is 81921, the last without it 81867. The 2-nd says the first with artifact is 80955,the last without it 79600. Anyone else can report which is the earliest build you can reproduce the artifact with?