Jump to content

twostars

Administrators
  • Content Count

    1745
  • Joined

  • Last visited

Posts posted by twostars


  1. Changelog

    • Fixed a crash that occurred when a lot of data was received by the client (e.g. zoning into Moradon).
    • Fixed a couple of cases where large data was no longer compressed, causing much more network traffic than usual.
    • Inaccurate Blinks (which appears to be caused when playing with a low frame rate) are now better handled and should no longer report "Failed Blink" in the client (as opposed to the regular error "<skill> failed").

  2. Issue with it crashing should be fixed as of this restart, in addition to other necessary server-side optimisations.

    I'll re-enable batching which may still cause unintended side-effects, but we'll see how it is. I can always turn it back off at any point. :)


  3. Good to hear. Looks like the client's decryption couldn't deal with such large data (lots of players in Moradon), so I've temporarily disabled batching until I can figure out why that is.

    With it disabled, what's the status on the Blink issues? Is it failing less than before?


  4. Changelog

    • Drains can now only be used once per target per 60 seconds, meaning that only 1 drain can be applied to a single monster (or NPC) in that 60 seconds regardless of who is trying to use it.
      • This intends to discourage players from forming parties specifically to abuse the effectiveness of drains (e.g. on Ultima), while still giving them value.
      • Note that this does not apply to player targets.
    • Fixed the remaining cause of debuffs being cured before they're applied in the client.
      • This caused debuffs (e.g. Divide Armor) to 'stick' and not expire visually without a relog/zone change.
    • Fixed a bug with erroneous data being sent to the client unnecessarily.
    • Reduce batched packet delay to better line it up with movement to try to minimise movement delays.
    • Implemented measures to prevent visual 'bounce-backs' after being teleported.
    • Removed nation restriction for monster quest credit.
      • This intends to discourage scenarios like players swapping to one nation to kill Ultima to ensure credit.
    • Fixed a bug causing loot to be inaccessible when the monster's location is considered unusually higher than the player.

  5. Changelog

    • Debuffs can no longer be cured before they're applied in the client.
      • This caused debuffs (e.g. Divide Armor) to 'stick' and not expire visually without a relog/zone change.
      • Special thanks to Sick for helping to track down and consistently reproduce this.
    • Cubed targets can no longer be affected by any AoE whatsoever.

  6. 4 hours ago, Nath said:

    I'm not following you either. I go all the time Jr > Eslant > Moradon > CZ. Unless you're counting Eslant as EMC, I never go there.

    Anyway, it happened again last night. I was in CZ with both attack/defense enchant scrolls and got summoned into Juraid (weapon enchant scroll was gone as soon as I landed into Juraid). I haven't played since last night so you can check out that time frame.

    Again, this might be just visual (client side?) I will pay attention now checking my AP to see if it remains the same even after the scroll is gone.

    Sorry, that instance I was referring to was actually BDW not Juraid, that's why it took you back to EMC instead of Eslant.

    Your current case is very interesting. It looks like it was removed in CZ with the others as if you clicked them off at 2:35:31~2:35:33, and all except the Weapon Enchant scroll reinserted fresh immediately after.

    The next hit I see is you manually rebuffing with new Weapon Enchant & Armor Enchant scrolls at 2:53:35 (in Juraid).

    Juraid started & you were teleported in at 2:50:00, which carried across your other buffs fine.

    So it actually happened 15min prior to entering Juraid - for some reason, at ~2:35:31, it clicked of all of your scrolls (500033 [HP Scroll 1500], 500030 [Scroll of Armor 300], 500034 [Scroll of Attack], 500035 [Speed potion], 500049 [Weapon Enchant scroll]) and I assume you manually reinserted the rest, as the next hits were applying new scrolls (500035 [Speed potion], 500030 [Scroll of Armor 300], 500033 [HP Scroll 1500], 500034 [Scroll of Attack]) - so no new Weapon Enchant scroll.

    Question is what exactly is happening here and why. It's telling me you're clicking them off and then rebuffing them all (and forgetting the Weapon Enchant scroll). And again, this is 15min prior to Juraid - when you zone into Juraid you obviously still don't have the scroll here because it was removed then.


  7. Unfortunately I'm unable to reproduce it using the steps provided. I had a look over your logs around that time and it appeared your character had the debuff up until after you left Juraid - rather, you exited Juraid, the (same) scroll was restored (in EMC), you teleported to CZ and 20 seconds after it restored, it removed it - registering it as if you clicked it off.

    I'm not entirely sure if this was the same session you're referring to, but it appearing as if it's clicked off (as well as some of your other buffs at the same time) suggest it might be.

    I don't want to jump to any conclusions though, so if you're able to reproduce this, can you give us a shout when exactly you do so I can chase it down in logs?

    Thanks.


  8. Changelog

    • Fixed a bug causing various data to not be sent to all players in an instance consistently.
      • This was directly responsible for the death notices not being shown to everyone.
    • Added diminishing returns to stuns, slows and skill cancellations.
      • We'll continue to monitor consecutive hit behaviour and tweak this as needed, but it should help stop stuns/slows/interrupts hitting multiple times in short windows, which we believe to be the main source of frustration experienced with this. Note that resistances are very important, so if you're being stunned/slowed a lot, you should try to get more gear with lightning resistance (for stuns) or ice resistance (for slows).

  9. Staff ranges do seem larger than they should be, but it is respecting the client's range, so I'm assuming what you're seeing is a result of that + the visual inconsistency.

    We may reduce their base range from the official range; that should keep things more in line with other melee-type skill ranges.


  10. 21 minutes ago, MERET said:

    My pc is really bad. I can't record video and also I can't take what you want us to give information when I focus on PK. GM maybe helps us about this issue. I am not the only telling that, you know.

    Just a general timeframe and your character name (I'm assuming MERET) is all I'm asking for. "Sometime within the last couple of hours" is plenty.

    Regardless, I'm sifting through your stun logs (in particular) and not really seeing anything crazy. For reference, here's what I'm looking at (a report of all stun attempts to MERET in the last 7 days, excluding those from Chaos Dungeon skills -- since they're hard rates and don't matter for anything, though the Kurian Rush passive is in much the same boat, but oh well), since I want to be completely transparent here:
    https://pastebin.com/mbDVwaP6

    What you see is that with your resistances, the rates are very low. And as a consequence, there's a ton of misses. I don't see many close proximity hits at all; they typically are still spread apart.
    One exception is:
     

    1. October 12th 2018, 08:55:47.000 190771  247 275 46  2.394726    HIT
    2. October 12th 2018, 08:55:45.000 190772  247 275 46  2.394726    HIT
    Which is extremely low (2.4%) but happened within 2 seconds. It's going to happen; it's really down to RNG for that. But occurrences of this, or generally any occurrences of hits within reasonably short timeframes do not happen often. They happen occasionally, but they're certainly not recurrent enough to consider it a rate issue.
     
    Which brings me back to the idea of using diminishing returns to try to reduce these particular cases, which I believe is where your frustration actually lies.
     
    Of course, a general timeframe so I can know which session you're talking about would help me identify where those frustrating cases lie for you, but whatever.

    Edit:
    Left in a bunch of mob skills too (those start with 3*). Those can be ignored as well.

  11. 22 minutes ago, Rougean said:

    2. I think this is just a visual bug? Like when warriors hit you from miles away. I am not sure tho, but should be easy to check.

    Yes, I'll be looking into it but I assume right now it's just a visual inconsistency. We've had to balance the responsiveness of movement a little with performance, as the client is very inefficient with how it handles movement updates, so movement/positioning is slightly (0.2s) less responsive than it was before, however it ultimately performs much, much better in larger scales (e.g. with UTC), so I'm fairly happy with it as it is. It's still 300% as responsive than it is officially, so for now - until I can spend the time fully optimising movement update logic - it's a much more performant middle-ground which doesn't cause the extra stuttering lag people were experiencing.

    I'll check its range though, just to make sure it's not doing anything wacky like Blow Arrow was at one point.


  12. 27 minutes ago, MERET said:

    By the way, you always focus on one sentence within all sentences and according to that you answer all the time. 

    I read all of it, but the issue I'm presently focused on is rates. As I'm making that change soon, I was hoping for any and all actual evidence behind your claims so I can see what it is you're actually complaining about, but if you're not willing to do that, all I can do is ignore your claim entirely as it doesn't help us in any way.


  13. 34 minutes ago, MERET said:

    A success rate of ice and stun are too much than normal. I get plenty of resistance but it doesn't work.

     

    Give me a timeframe for when you were stunned / slowed too much and the name of your character. I can tell you what your rate was and how much they missed at that time.

    I suspect, like the other cases we've looked at, it's not a case of the rates being too high (because they're not). It's the frustration that occurs when, due to RNG, they happen twice in a short span of time. Adjusting rates doesn't really fix that, and causes more problems.

    On the other hand, it's looking like I'll be implementing diminishing returns specifically to address these frustrating cases.


  14. 9 hours ago, red123 said:

    Could you tel me what's included in such reports? I've noticed "Desktop screenshot generator failed" What's that? Does it mean I've been sending screenshots from my desktop? 😱

    The crash reporting service has many features available for us to use, but all we include in the data we send is the crash dump, that's it. The crash dump is basically a memory dump of the KO process, and which part it was running at the time it detected a problem. Otherwise we'd end up with more data than we know how to deal with. Screenshots really don't help us in these cases anyway.

    I believe it should show you when you submit what's being sent to us, as well.

    As for your crash, it was fixed in the patch on the 7th. Thanks for your report. :)

×