Update on 4.35 U2
-
_"Ambitious, considered all the people here’s skills…
But I like this behavior in a pilot." (more or less, Cmdr. Viper addressing to the scholars attending at the TopGun first day :D)_
With best regards to all.
-
One more report from the front…
This weekend we will go heavy on another candidate, our 30th.
There are no more known critical issues, which means all we knew that was wrong, we fixed: from very old bugs in Falcon core, to very complex bugs in our new rendering engine. Of course, new issues can show up as well, even the fixes might contain bugs.
There are still logic bugs, but the focus of this release is stability. This means we still have weird AI behaviour and some glitches. But we expect a lot less CTDs than before.
Fingers crossed here.
-
If this is truly supposed to be a simulation of real life, then remember that some weird AI behavior is realistic, because some people are weird.
-
One more report from the front…
This weekend we will go heavy on another candidate, our 30th.
There are no more known critical issues, which means all we knew that was wrong, we fixed: from very old bugs in Falcon core, to very complex bugs in our new rendering engine. Of course, new issues can show up as well, even the fixes might contain bugs.
There are still logic bugs, but the focus of this release is stability. This means we still have weird AI behaviour and some glitches. But we expect a lot less CTDs than before.
Fingers crossed here.
Don’t get this message wrong, 4.35 has not introduced weird behavior compare to previous versions
Ai « weird » behavior is mainly related to AG behavior that will not be fixed in 4.35
this behavior in AG is the same from 4.34, 4.33, 4.32 and all other falcon versions
To change this a full rewrite was needed but that’s not for 4.35
In AA I still to find a way to teach AI to not follow enemy in hostile territory for too long , but in the same time alloW them to engage in deep hostile territory when needed
This is not easy problem to solve , for now I don’t have the solution to get a reliable behavior. Problem beeing that AI are of course not « intelligent » they are just state machines
-
Is there some logic you could apply around closure rates vs duration of pursuit that might help a bit?
Sent from my iPhone using Tapatalk
-
Is there some logic you could apply around closure rates vs duration of pursuit that might help a bit?
Sent from my iPhone using Tapatalk
partially
-
Still very excited for U2 stability.
-
Tests are finished.
-
-
-
well, I’m thinking 3-4 weeks.
I mean, tests finished, don’t know if they passed or failed, were they code tests or acceptance tests?
-
-
I am thinking what you are thinking :woohoo:
I am checking the forum again and again since that post like a million times…. I cannot sleep… LOL…
-
I am checking the forum again and again since that post like a million times…. I cannot sleep… LOL…
Yeah don’t overhype
We just finished the testing part -
-
-
Sorry, didn’t want to make this a cryptic message… We have spent a long and tiresome period of testing. We finished it because we think we have a product that is as good as we think it should be.
Like I said before, not perfect, but good enough. Rest assured our team worked really hard on this one (as usual, but this time focus was on stability).
Now we are in the process of packing and doing some sanity checks, to deliver the final version. Codewise, we will not change anything except if something really serious shows up.
It will still take some days though, no need to break the reload button.
-
I am thinking what you are thinking :woohoo:
I think so Brain but me and Pippy Longstocking; what would the children look like?
-
-
i hope u2 to be the last one update