Skill queuing/Difference in terms of spamming skills. Our server has changed emulator around 4-5 months ago from eAthena to rAthena, I have noticed something that bothers me and that affects my game play as a whole. In rAthena there is no buffer or queuing on skill execution, meaning you would need to hit your mouse button exactly after the after cast delay, of a previous skill ending, to achieve a perfect transition from first cast to second cast, which is seemingly impossible. On our previous client however, it is readily possible to achieve that perfect transition, meaning that there has to be some sort of buffer, queuing window that allows you to pull that off.
An example would be if you cast magni or other partywide spells like overthrust/weapon perfection if you try to cast a skill during that time i'd still go through at the end of those animations it would queue it, in rAthena however it doesn't. As if everything has a big delay and no buffer, or queueing thus everyone got slow; you can't cast as fast as the game intended you to. You can't even cast fast enough to do the techniques you were intended to do with it. Other examples would be: You can't resurrect someone and lex fast; bless and assumptio fast, can't spam jt/ds as fast.
Out of curiosity, I have asked a friend of mine to help me test this on other emulator, that of which being Hercules. I found out that in Herc, this problem doesn't seem to exist as it felt how I spammed in our previous emulator (eAthena.), when I tested Jupitel thunder and Double Strafe. Is it possible to know what's causing this skill delay?
Additional Tests that were made in eAthena vs rAthena: Delay
In this screenshot we tested Jupitel Thunder old client vs our current client. With the same hw, same gears, same overweight to make sure there's no variance in sp while casting, both starting from 100% sp and spamming JT for 10 seconds. The screenshot at the top is the previous client, bottom is in rathena. There is 30% sp in the first screenshot, 60% in the second one.It's AT LEAST half as fast as the old client. at the very least first screenshot, the user used up 2162 sp, second screenshot, in the same time, same casting, same technique, same spell, the user only used up 1316. 2162 divided by 1316 is 1.64. Jupitel Thunder costs 47sp, which means in the second screenshot you could do 28 casts in 10 seconds, while in the in the first screenshot you could do 46. (Yes, there is user errors as nobody could count exactly 10 seconds but, it is a huge discrepancy. You can't have a difference that big just from one second or so off in your timings.)
Thus concluding that, the old client was roughly 165% faster. Through this test, we were able to tell that its not just queueing that has a problem but in general there is a delay which we could tell also by looking at the animations where the High wizard moves slower than it used to.
Queue
We also tried counting the frames:
Where the difference is huge. In this test we used the skill assumptio for High Priests. In this test consecutive assumptio casting were done: In the Old client 64 frames until acd goes off visually, and then we could cast again in the 65th frame. At rAthena 89 frames until acd goes off visually, and one can cast again at the 111th frame.
With this test we have concluded that you have to be off acd for another execution to be queued up, or maybe its not queued up at all anymore unlike how it used to be.
----
Notes:
Min_skill_delay limit is 200 in all the emulators used: rA, eA and Hercules.
I don''t know what's causing the issue. Does anyone have an idea what it could be?
We also replicated the Jupitel Thunder test on a Hercules server and found out that this particular delay doesn't exist in Herc. I'm quite curious if people here have any idea on what this particular delay is.
