Official eMule-Board: Client Analyzer Skull Matches - Official eMule-Board

Jump to content


  • (8 Pages)
  • +
  • 1
  • 2
  • 3
  • Last »

Client Analyzer Skull Matches Which mod is the worst?

#1 User is offline   James R. Bath 

  • Golden eMule
  • PipPipPipPipPipPipPip
  • Group: Members
  • Posts: 790
  • Joined: 02-August 04

Posted 22 June 2010 - 08:18 PM

Skulls are the way ClientAnalzyer (CA) reports whether a user is up to bad things. With the interest/objective of providing a scientifically reproducible way to both evaluate CA's bad mod detection and which mods need improvement/retirement, I'm going to record skulls. Preferably, they'll improve.

This was unofficially begun under another topic to determine whether a particular mod was bad. Feel free to add your own CA comparisons/matches.

My skull match for today is Xtreme vs. Morph using the CA in SharkX 1.7.
  • Xtreme: 5 skulls out of 54 total (9.26%) :flowers:
  • Morph: 19 skulls out of 43 total (44.19%) :-1:

Fakes/thieves: 9 Xtreme, 5 Morph. These are excluded from the skull and total counts because they are probably different, unofficial mods.
Queue/Known Clients at match time: 3107/4067
Number of active Downloads: 2
Note: this test did not exclude mods marked with skulls because of bad UL/DL ratios. I'll be excluding those from the counts in future matches. Based on memory, it wouldn't have had a significant impact on this match. Both mods would have had a couple less skulls.

Match Rules:
Bad UL/DL excluded, fakes/thieves excluded, only users in queue counted. In cases that there are a ridiculously large number that would be difficult to count quickly and accurately, samples are taken using File Priority or Filename as the primary sort so that a random chunk of the users are sampled as opposed to those more likely to have skulls or not.

Complete Match List
(direct links to avoid the tantrums, whining and offtopic rants of the losers :argue: :cry: )

This post has been edited by James R. Bath: 14 July 2010 - 11:49 PM

Currently recommending and using: eMule beba 2.63
For slot control only, currently recommending: Tombstone Xtended 1.0 (or higher) if you absolutely must have slot control


Posted Image

Quote

Where there is a mule there is fuel. Where there is a stool sits a fool. - Winston Churchill

-8

#2 User is offline   Stulle 

  • [Enter Mod] Dev
  • PipPipPipPipPipPipPip
  • Group: Members
  • Posts: 5804
  • Joined: 07-April 04

Posted 22 June 2010 - 08:48 PM

God damn it, do you have no life? No friends? Not enough self-esteem to already stop this nonsense? We got it, you don't like me so you keep bashing Morph!

Even using the word "scientifically" in this thread is driving the whole idea of working scientifically ad absurdum! As a matter of fact I was asked if it was possible to achieve 1:10 or even 1:12 with common MorphXT and I agreed to that given some conditions. Turns out all conditions were met. Another most interesting fact was it was in fact a MorphCA so a MorphXT with Client Analyzer that brought back these results! Any bells ringing?

It's always the user behind the client who can cause negative behavior of the client. Everybody with just a little knowledge will agree to that. As a matter of fact does MorphXT not violate any of the network rules. However, there are some people out there - possibly just as smart as you are, James - changing MorphXT in bad ways so it will not obbey the limits we included in the code or allow even worse. This is beyond our control and yet these clients identify themselves as MorphXT to the network. What makes you think you will not get any results from them? No legit MorphXT is intended to trigger any of the CA's checks and if it does I would be glad to be pointed to that in order to find possible flaws of the code. It is, however, possible - if not likely - that MorphXT can be considered bad due to a negative Up/Down ratio. This can be caused due to a great number of possible cases. Most notable of course, we got a request only scenario for most of the remote MorphXT clients. Also, MorphXT provides the user with a powerful PowerShare feature which can eventually block traffic to most/all other (non-PS) files. This is of course not desired and this is why MorphXT - just like ScarAngel - uses a mechanism to deny powersharing well spread files. Obviously, another point where MorphXT stands for fairness!

So what can the intelligent reader get from this thread? Nothing... just, yet, another example of how much James tries to discredit my work because he personally dislikes me. He goes on to post invalid arguments and publish invalid, biased test results.

Honestly, if anybody also thinks it is getting annoying - and I am looking in the direction of Wiz, Andu and possibly pP here - please voice your opinion. I really don't mind criticism but I darn well hate some random idiot noob who contributed nothing whatsoever discredit the work of my former colleagues and my work without having any real argument.

EDIT:
Just another thought, do you even realize that you are just heating the old CA vs. non-CA discussion, again?! If there was anything questionable in MorphXT it would have been brought up long ago and since it has not been you are just fueling old conflicts. What are you trying to achieve? What do you expect me to do? Do you expect me to sit quietly when you go on spreading invalid assumptions about MorphXT? At least have the balls and have a proper discussion where you do prove me wrong or just insult me personally. But keep MorphXT and my other mods out of this. I have proven I am contributing to the community. The only thing you contributed lately is reason for many rants on my side and annoyance in some other users. So please, tell me, what is your aim? Getting back at me where I hate it the most, my work?

This post has been edited by Stulle: 22 June 2010 - 09:24 PM

I am an emule-web.de member and fan!

[Imagine there was a sarcasm meter right here!]

No, there will not be a new version of my mods. No, I do not want your PM. No, I am certain, use the board and quit sending PMs. No, I am not kidding, there will not be a new version of my mods just because of YOU asking for it!
3

#3 User is offline   coluche 

  • hm ?
  • PipPipPipPipPipPipPip
  • Group: Members
  • Posts: 2274
  • Joined: 02-May 05

Posted 22 June 2010 - 09:32 PM

Quote

What are you trying to achieve?


provoking an answer out of you is my guess.

It was already in the topic with the VDSL client where I thought "Why do you honor him with answers?"

My guess is this topic would have died all alone with no answers, buried soon enough in all the old topics.

For some, being bashed seems better than no attention at all.
It's Screamin' Jay Hawkins and he's a Wild Man, so bug off!
1

#4 User is offline   Stulle 

  • [Enter Mod] Dev
  • PipPipPipPipPipPipPip
  • Group: Members
  • Posts: 5804
  • Joined: 07-April 04

Posted 22 June 2010 - 09:42 PM

Well, seems like omeringen is on the right track, he agrees with you. Nevertheless am I really having a problem with my work being discredited... it's just one of my buttons and I just can't help myself a lot if it is pushed.
I am an emule-web.de member and fan!

[Imagine there was a sarcasm meter right here!]

No, there will not be a new version of my mods. No, I do not want your PM. No, I am certain, use the board and quit sending PMs. No, I am not kidding, there will not be a new version of my mods just because of YOU asking for it!
0

#5 User is offline   ElChele 

  • Miembro con emule 0.50a
  • PipPipPipPipPipPipPip
  • Group: Members
  • Posts: 7295
  • Joined: 02-September 04

Posted 22 June 2010 - 09:51 PM

hi
One thing that I have very clear.
We are many more who appreciate your work, and I'm pretty sure that you also do.
I think that should be enough for everyone.
bye :flowers:
Posted Image Make your own ipfilter file, and manage fake files. Take in count, You are the best filter for emule.
2

#6 User is offline   Stulle 

  • [Enter Mod] Dev
  • PipPipPipPipPipPipPip
  • Group: Members
  • Posts: 5804
  • Joined: 07-April 04

Posted 22 June 2010 - 09:53 PM

Nice to hear! But I guess you know how it is, it's the unpleasant things that you can remember the best. Anyway, thanks for the compliment.
I am an emule-web.de member and fan!

[Imagine there was a sarcasm meter right here!]

No, there will not be a new version of my mods. No, I do not want your PM. No, I am certain, use the board and quit sending PMs. No, I am not kidding, there will not be a new version of my mods just because of YOU asking for it!
0

#7 User is offline   James R. Bath 

  • Golden eMule
  • PipPipPipPipPipPipPip
  • Group: Members
  • Posts: 790
  • Joined: 02-August 04

Posted 23 June 2010 - 08:57 AM

After doing a question mark search I found maybe 2 actual questions in all the ranting and personal smears that would deserve answers if I hadn't already answered them. Remedial reading 101 is in session:

View PostStulle, on 22 June 2010 - 09:48 PM, said:

However, there are some people out there - possibly just as smart as you are, James - changing MorphXT in bad ways so it will not obbey the limits we included in the code or allow even worse. This is beyond our control and yet these clients identify themselves as MorphXT to the network. What makes you think you will not get any results from them?

This reduces to 1 of 2 things. One, your making the "fake" argument again. As mentioned above, the fakes have been removed from the counts. Two, this is back to you blaming others for making your mod do bad things. If there wasn't some problem in your mod, that wouldn't be an issue. For all anyone knows, you put it there so it could be exploited free of any blame. In this case you made it possible for the baddies to become that way via config settings, so really the problem is Morph and it deserves every skull it gets. CA reports bad behavior and if your mod allows bad behavior, you need to change it so it doesn't.

View PostStulle, on 22 June 2010 - 09:48 PM, said:

Just another thought, do you even realize that you are just heating the old CA vs. non-CA discussion, again?!

If you had actually read my opening post, you'd have read the part at the top "... to both evaluate CA's bad mod detection and which mods need improvement/retirement, I'm going to record skulls." IOW, great. :flowers: Let those skulled and those skulling be open for review.
Currently recommending and using: eMule beba 2.63
For slot control only, currently recommending: Tombstone Xtended 1.0 (or higher) if you absolutely must have slot control


Posted Image

Quote

Where there is a mule there is fuel. Where there is a stool sits a fool. - Winston Churchill

-2

#8 User is offline   Stulle 

  • [Enter Mod] Dev
  • PipPipPipPipPipPipPip
  • Group: Members
  • Posts: 5804
  • Joined: 07-April 04

Posted 23 June 2010 - 09:09 AM

So avoiding the more important questions, are you?

And no, you are only excluding obvious fake. There is no identifying a client with the exact same code, putting aside the removal of the ZZ Ratio and probably adding kick/ban from upload. Additionally, if they were all actually legit MorphXT clients, MorphXT is still a lot more strict than other mods and the official! Read the code if you can.

You have no idea what is even the reason for the skulls, I take it. Well, if you got nothing sensible to say please say nothing at all because you are not helping. You are just provoking.

And now just to give you some more of the important questions... or rather asking them again:
  • What are you trying to achieve?
  • What do you expect me to do?
  • Do you expect me to sit quietly when you go on spreading invalid assumptions about MorphXT?
  • So please, tell me, what is your aim?
  • Getting back at me where I hate it the most, my work?


Just one last remark, it is easy to criticize but it is hard to understand. I know that because I know the code and you don't. Just speaking of the upload, you are comparing apples to beans if you compare MorphXT's upload to, say, SharkX' upload. To give you a car metaphor, MorphXT upload is like a very high tech tuned and modified race car engine. It works under extreme stress but you will have to be careful with it. Driving in the desert is not an option because too much dirt will just break it. Official and most other systems are more like a standard vehicle you can buy from any given car manufacturer with some quality standard. It will drive up in the North in coldest winter and down in South right in the desert. You won't make it first to the finish line but it's okay because you will do it regardless of the circumstances. MorphXT's upload is a very powerful tool and this makes it very complex. Wiz once said he attempted to write a similar system (Upload Splitting Classes) and eventually gave up, although he might have written a working one by now. Having complex code introduces many possibilities for bugs. Still, thanks to SiRoB and leuk_he, MorphXT's upload has come a long way and is as of now the most powerful upload management system available to the eMule modding community.

This post has been edited by Stulle: 23 June 2010 - 09:23 AM

I am an emule-web.de member and fan!

[Imagine there was a sarcasm meter right here!]

No, there will not be a new version of my mods. No, I do not want your PM. No, I am certain, use the board and quit sending PMs. No, I am not kidding, there will not be a new version of my mods just because of YOU asking for it!
0

#9 User is offline   James R. Bath 

  • Golden eMule
  • PipPipPipPipPipPipPip
  • Group: Members
  • Posts: 790
  • Joined: 02-August 04

Posted 23 June 2010 - 04:39 PM

View PostStulle, on 23 June 2010 - 10:09 AM, said:

So avoiding the more important questions, are you?

You mean like why there are more Xtreme mods and fewer Xtreme skulls? Or do you mean why Xtreme's code is "a lot more strict" than Morph?

View PostStulle, on 23 June 2010 - 10:09 AM, said:

And no, you are only excluding obvious fake. There is no identifying a client with the exact same code, putting aside the removal of the ZZ Ratio and probably adding kick/ban from upload. Additionally, if they were all actually legit MorphXT clients, MorphXT is still a lot more strict than other mods and the official! Read the code if you can.

Or why your mod is less strict than the easyMule and aMule mods? Seems you're pulling random excuses out. The "important question" is why Morph is worse relative to other mods? If CA is only excluding obvious fakes, it's doing it to all the clients, not just Morph, and Morph still ends up with more skulls, a lot more.

This post has been edited by James R. Bath: 24 June 2010 - 01:23 AM

Currently recommending and using: eMule beba 2.63
For slot control only, currently recommending: Tombstone Xtended 1.0 (or higher) if you absolutely must have slot control


Posted Image

Quote

Where there is a mule there is fuel. Where there is a stool sits a fool. - Winston Churchill

0

#10 User is offline   tHeWiZaRdOfDoS 

  • Man, what a bunch of jokers...
  • PipPipPipPipPipPipPip
  • Group: Members
  • Posts: 5630
  • Joined: 28-December 02

Posted 23 June 2010 - 05:48 PM

View PostJames R. Bath, on 22 June 2010 - 10:18 PM, said:

My skull match for today is Xtreme vs. Morph using the CA in SharkX 1.7.
  • Xtreme: 5 skulls out of 54 total (9.26%) :flowers:
  • Morph: 19 skulls out of 43 total (44.19%) :-1:

While it is interesting to get some data, it's of no use if you don't supply us with additional details. The main point in this case would have been the reason why those clients got marked as bad guys. If you find the Xtreme clients as sources for your downloads only (don't request something of you) but the Morph clients enter some "tit-for-tat" with you, then it's kinda clear, why they will soon get marked as "bad" guys.
I'd refrain from any accusation that you don't have sufficient evidence for - personally, I dislike Xtreme for its chunk-kick-behaviour (and of course DLP) but it's LEGAL to use.
2

#11 User is offline   Stulle 

  • [Enter Mod] Dev
  • PipPipPipPipPipPipPip
  • Group: Members
  • Posts: 5804
  • Joined: 07-April 04

Posted 23 June 2010 - 07:33 PM

Thanks Wiz, exactly the point.

James, I repeated the questions I consider important. Please reply to them and don't get distracted by further ongoing discussion. I am still waiting for the 5 questions that are numbered in my last post. And no, I don't care how many people use MorphXT or Xtreme. If I did I would not help zz_fly with the hosting via the original Xtreme SourceForge project. If I did I would never have proposed to Xman that I would pick up where he left (although I eventually withdrew that offer due to Xman's behaviour). If I did I would quit StulleMule because only a small number of users do actually use it. If I did I would not review zz_fly's code when he asks me to, offer him my own Xtreme merge to 0.50a Beta 1 (or was it 2?) for reference purpose and tell him when he should check MorphXT sources for fixes and improvements on the SafeHash code. All of which I did. You may think I am self centered but then, I probably just don't have anything to offer to you because you don't do anything useful for this community aside of making the people happy who like to see me rant about!

Also, in what way is easyMule or aMule more strict than MorphXT? Do they employ a 1:3 ratio at all time? That would indeed be more strict but AFAIK they just don't. What I was speaking about is the strict denial of PS and other releaser features in non-releaser scenarios. There is a strict ratio for unlimited friendslots and/or PS on partfile. The earlier not being in all mods that allow Slotfocus or loosely restricted slotlimiter that allow just 2 slots for instance, which is - to my mind - very unfair. Most other mods - and that means also many of those you like to promote while trying to bash Morph - do not even have a session ratio limit under any circumstances, while MorphXT enforces this when the upload is below 10 because, for instance, the upload limit is set to 10 kB/s. More strict in this sense. This is all in the code.

Your findings are not worth anything unless you provide real information like Wiz suggested. Please also follow his advise and stop these accusations unless you get sufficient and evident proof that there is code in MorphXT that qualifies it as bad. Until then you are just spreading invalid arguments to further your smear campaign, for whatever reason you have. The latter remark obviously also means I still wonder why you are doing all this. But please review and answer the numbered questions in my last post.

PS: @ Wiz: I take it you meant the Xtreme Full Chunk upload code, right? It's of course not exactly perfect if we stop a upload transfer but trying to complete chunks in as many remote clients as possible is also a desirable aim. To my mind it's a matter of taste what should be preferred.

This post has been edited by Stulle: 23 June 2010 - 07:45 PM

I am an emule-web.de member and fan!

[Imagine there was a sarcasm meter right here!]

No, there will not be a new version of my mods. No, I do not want your PM. No, I am certain, use the board and quit sending PMs. No, I am not kidding, there will not be a new version of my mods just because of YOU asking for it!
0

#12 User is offline   James R. Bath 

  • Golden eMule
  • PipPipPipPipPipPipPip
  • Group: Members
  • Posts: 790
  • Joined: 02-August 04

Posted 24 June 2010 - 01:18 AM

View PosttHeWiZaRdOfDoS, on 23 June 2010 - 06:48 PM, said:

View PostJames R. Bath, on 22 June 2010 - 10:18 PM, said:

My skull match for today is Xtreme vs. Morph using the CA in SharkX 1.7.
  • Xtreme: 5 skulls out of 54 total (9.26%) :flowers:
  • Morph: 19 skulls out of 43 total (44.19%) :-1:

While it is interesting to get some data, it's of no use if you don't supply us with additional details. The main point in this case would have been the reason why those clients got marked as bad guys. If you find the Xtreme clients as sources for
...
personally, I dislike Xtreme for its chunk-kick-behaviour (and of course DLP) but it's LEGAL to use.

I'm not arguing for Xtreme. You don't see me recommending it in my signature or on my page here. It's simply that relative to Morph, by CA standards, it's the better behaved mod.

As to the reason, most of the Morphs were marked FastXS or Reask@. It's a good point about the details. I'll make it a point to exclude from future counts those that are marked because of bad UL/DL ratios (but not failed UL/DL). I'm adding to the top post the number of active DLs at the time as part of the reference info. There were very few and by far I was doing more uploading.
Currently recommending and using: eMule beba 2.63
For slot control only, currently recommending: Tombstone Xtended 1.0 (or higher) if you absolutely must have slot control


Posted Image

Quote

Where there is a mule there is fuel. Where there is a stool sits a fool. - Winston Churchill

-1

#13 User is offline   James R. Bath 

  • Golden eMule
  • PipPipPipPipPipPipPip
  • Group: Members
  • Posts: 790
  • Joined: 02-August 04

Posted 24 June 2010 - 05:29 AM

See first message in this topic for more background.

Tonight's skull match features yesterday's loser versus a non-emule client with the CA results courtesy of eMule beba v2.60:
  • Morph: 12 skulls out of 31 total (38.71%) :-1:
  • Shareaza: 7 skulls out of 38 total (18.42%)

Testing conditions:
Fakes/thieves: 0 Morph, 0 Shareaza. Excluded from the skull and total counts because they are different, unofficial mods.
Bad UL/DL ratio: 0 Morph, 0 Shareaza. Excluded from counts.
Queue/Known Clients at match time: 2236/2417
Number of active Downloads: 2
Currently recommending and using: eMule beba 2.63
For slot control only, currently recommending: Tombstone Xtended 1.0 (or higher) if you absolutely must have slot control


Posted Image

Quote

Where there is a mule there is fuel. Where there is a stool sits a fool. - Winston Churchill

-2

#14 User is offline   tHeWiZaRdOfDoS 

  • Man, what a bunch of jokers...
  • PipPipPipPipPipPipPip
  • Group: Members
  • Posts: 5630
  • Joined: 28-December 02

Posted 24 June 2010 - 05:43 AM

View PostStulle, on 23 June 2010 - 09:33 PM, said:

PS: @ Wiz: I take it you meant the Xtreme Full Chunk upload code, right? It's of course not exactly perfect if we stop a upload transfer but trying to complete chunks in as many remote clients as possible is also a desirable aim. To my mind it's a matter of taste what should be preferred.

That's something to discuss... IMHO, if 2 clients wait for their turn and finally get it, they should get the same as they waited long enough. It's not okay to kick client A after 2.5MB just because he already completed a chunk that was close to completion and/or forcing him to ask for a new chunk so he isn't treated in that unfair way - IMHO Xtreme is cheating on those users.

View PostJames R. Bath, on 24 June 2010 - 03:18 AM, said:

As to the reason, most of the Morphs were marked FastXS or Reask@. It's a good point about the details. I'll make it a point to exclude from future counts those that are marked because of bad UL/DL ratios (but not failed UL/DL).

Well, that's interesting and maybe it's worth a look into Morph sources to see whether there are some changes (and/or bugs) around that code parts.
1

#15 User is offline   Stulle 

  • [Enter Mod] Dev
  • PipPipPipPipPipPipPip
  • Group: Members
  • Posts: 5804
  • Joined: 07-April 04

Posted 24 June 2010 - 08:58 AM

no there is not. at least not around the reask. i just reviewed the code the other day. as a matter of fact do we use the official GetTimeUntilReask function. SX should in fact be slower because we use SourceCache which buffers sources we received in excess for a while. CreateSrcInfoPacket is unmodified for the most part, the only changes done are required to work with the ICS code but are fine. IsSourceRequestAllowed has also a number of changes but all of them conclude with the statement "return false;" which means we certainly do not ask earlier, rather later. the one only change that might cause faster reasks is in SetDownloadState where I readded the below code. note that this is still official code which was removed in 0.49a or b that i needed to readd to stop swapped sources to be stuck for 20 minutes:
            case DS_NONEEDEDPARTS:
                // Since tcp asks never sets reask time if the result is DS_NONEEDEDPARTS
                // If we set this, we will not reask for that file until some time has passed.
                SetLastAskedTime();
                //DontSwapTo(reqfile);

			//MORPH START - Changed by Stulle, Ensure we reset m_dwLastTriedToConnect for A4AF and more
			//Note: I don't know why but this code is required for Morph... some more research may show why but I am not up for that right now.
			default:
				switch( m_nDownloadState )
				{
					case DS_WAITCALLBACK:
					case DS_WAITCALLBACKKAD:
						break;
					default:
						m_dwLastTriedToConnect = ::GetTickCount()-20*60*1000;
						break;
				}
				break;
			//MORPH END   - Changed by Stulle, Ensure we reset m_dwLastTriedToConnect for A4AF and more
		}

		if (reqfile){

did i miss anything? btw, i figured out why it was required when i dug deeper into it at a later time. turns out we set download state to NNP, which will reset dwLastTriedToConnect. then we swap the source to another file and if this was successful hope that the download continues for another file. not resetting this variable, however would stop eMule from sending AskForDownload so we timed out because we never sent the DownloadCancel opcode and never sent a new block request, either.

anyhow, i would also like to point out - AGAIN - that there is no proof provided that any of the clients marked as bad are indeed legit MorphXT clients. as of now there exist at least two leecher mods which are based on MorphXT and still bear that name! all we do have is James' word for that and i really don't think any of that is quite sufficient. you, Wiz, pointed it out yourself, there is just not enough information!

reading your last post, James, also shows that you don't understand how the fakes are detected. Wiz or zz_fly had the idea that we could detect fake mods if they suggest they were, say, MorphXT but did not send the name attached to the nick, which is mandatory for these clients. this will not catch bad mods based on MorphXT - in this case - but only bad clients with custom modstring where users set MorphXT as the modstring, while nothing is added to the nick. however, this is only done by few stupid leechers, as usually the leechers adjusted to this practice and started sending the name tag, too. other fakes may be nick or mod thieves, i figure. IMO, there is no other way to identify any client as bad, unless you probably know which protocol extensions should be sent but are not. i doubt, however, that wiz went to the extent of checking this because this requires updates once it changes.

wiz, as for the chunks, IIRC this could also happen when disabling full chunks in official, given another client is able to accumulate enough score to exceed the client in upload. anyway, it would be interesting to see how much clients do actually suffer from this. as a matter of fact will the queue turn over quicker due to the shorter upload transfers. so the disadvantage they gain from being kicked earlier might just as well balanced by the next transfer which starts quicker.

This post has been edited by Stulle: 24 June 2010 - 09:02 AM

I am an emule-web.de member and fan!

[Imagine there was a sarcasm meter right here!]

No, there will not be a new version of my mods. No, I do not want your PM. No, I am certain, use the board and quit sending PMs. No, I am not kidding, there will not be a new version of my mods just because of YOU asking for it!
0

#16 User is offline   tHeWiZaRdOfDoS 

  • Man, what a bunch of jokers...
  • PipPipPipPipPipPipPip
  • Group: Members
  • Posts: 5630
  • Joined: 28-December 02

Posted 24 June 2010 - 10:45 AM

View PostStulle, on 24 June 2010 - 10:58 AM, said:

did i miss anything? btw, i figured out why it was required when i dug deeper into it at a later time. turns out we set download state to NNP, which will reset dwLastTriedToConnect. then we swap the source to another file and if this was successful hope that the download continues for another file. not resetting this variable, however would stop eMule from sending AskForDownload so we timed out because we never sent the DownloadCancel opcode and never sent a new block request, either.

Did you check if that happens with official, too? Would be worth a bug report, them.

Quote

wiz, as for the chunks, IIRC this could also happen when disabling full chunks in official, given another client is able to accumulate enough score to exceed the client in upload. anyway, it would be interesting to see how much clients do actually suffer from this. as a matter of fact will the queue turn over quicker due to the shorter upload transfers. so the disadvantage they gain from being kicked earlier might just as well balanced by the next transfer which starts quicker.

Yes, though you won't be kicked if you still have a better score. Xtreme however, simply checks if your block requests exceeded the chunk boundaries and kicks you (if you got at least 2.5MB IIRC - was changed after I ranted a bit) - however, this is NO "full chunk" option and as described above I think it's pretty bad for the clients waiting for their turn. The client looses all of its waiting time after his upload session thus it may take a LONG time until he gets into upload again (might be even worse when combined with Xtremes credit system though I didn't check that) - however, this discussion shouldn't take place here.
1

#17 User is offline   Stulle 

  • [Enter Mod] Dev
  • PipPipPipPipPipPipPip
  • Group: Members
  • Posts: 5804
  • Joined: 07-April 04

Posted 24 June 2010 - 11:02 AM

no, it does not happen with official, because official does not try to. official sets NNP, cancels download and then swaps. SiRoB thought it would be better like it is now and i think he's right! for common swap or stop/resume scenario i verified that sources are stalled unknown for 20 minutes with official. changing it the way i did has the possibility of creating fast reasks but two things need to be considered here, firstly we should usually not swap too often. iirc there is also a check for that but i don't know for sure. anyway, like i said, it should not happen a lot either way! secondly, this is official code i just readded or rather un-commented. so this is official behaviour until 0.49a or b!

i am not sure about the loosing all waiting time thing. i know for sure that i changed it in a way the client looses only a percentage of the waiting time but IIRC something similar is in Xtreme, although, like i said, i am not really sure about that.

edit: okay, confirmed it has this code. note i left the ScarAngel tagging there but you get the code. ScarAngel handles this differently.
//Xtreme Full Chunk
/*
void CUpDownClient::SendOutOfPartReqsAndAddToWaitingQueue()
*/
// ==> SUQWT [Moonlight/EastShare/ MorphXT] - Stulle
/*
void CUpDownClient::SendOutOfPartReqsAndAddToWaitingQueue(bool givebonus)
*/
void CUpDownClient::SendOutOfPartReqsAndAddToWaitingQueue(bool /*givebonus*/)
// <== SUQWT [Moonlight/EastShare/ MorphXT] - Stulle
//Xman end
{
        //OP_OUTOFPARTREQS will tell the downloading client to go back to OnQueue..
        //The main reason for this is that if we put the client back on queue and it goes
        //back to the upload before the socket times out... We get a situation where the
        //downloader thinks it already sent the requested blocks and the uploader thinks
        //the downloader didn't send any request blocks. Then the connection times out..
        //I did some tests with eDonkey also and it seems to work well with them also..
        if (thePrefs.GetDebugClientTCPLevel() > 0)
                DebugSend("OP__OutOfPartReqs", this);
        Packet* pPacket = new Packet(OP_OUTOFPARTREQS, 0);
        theStats.AddUpDataOverheadFileRequest(pPacket->size);
        SendPacket(pPacket, true);
        m_fSentOutOfPartReqs = 1;
    
        //Xtreme Full Chunk
        // ==> SUQWT [Moonlight/EastShare/ MorphXT] - Stulle
        /*
        //if we gave less than 9 MB payload, we give back some waiting time in relation to payload
        uint32 bonus=0;
        uint32 waitingtime= (uint32)(GetWaitTime() ); //only half will be counted 
        if(givebonus)
        {
                //calculate the time
                if(GetQueueSessionPayloadUp() < 9*1024*1024)
                {
                        bonus = (uint32)(((PARTSIZE) - GetQueueSessionPayloadUp()) / (double)(PARTSIZE) * (waitingtime/2));
                }
        }
        */
        // <== SUQWT [Moonlight/EastShare/ MorphXT] - Stulle
        //Xman end

    theApp.uploadqueue->AddClientToQueue(this, true);

        //Xtreme Full Chunk
        // ==> SUQWT [Moonlight/EastShare/ MorphXT] - Stulle
        /*
        if(bonus>0)
        {
                if(credits)
                {
                        credits->SetWaitStartTimeBonus(GetIP(),::GetTickCount()-bonus);
                        AddDebugLogLine(false, _T("giving client bonus. old waitingtime: %s, new waitingtime: %s, client: %s"), CastSecondsToHM(waitingtime/1000), CastSecondsToHM((::GetTickCount() - GetWaitStartTime())/1000),DbgGetClientInfo()); 
                }
        }
        */
        // <== SUQWT [Moonlight/EastShare/ MorphXT] - Stulle
        //Xman end
}

This post has been edited by Stulle: 24 June 2010 - 11:07 AM

I am an emule-web.de member and fan!

[Imagine there was a sarcasm meter right here!]

No, there will not be a new version of my mods. No, I do not want your PM. No, I am certain, use the board and quit sending PMs. No, I am not kidding, there will not be a new version of my mods just because of YOU asking for it!
0

#18 User is offline   James R. Bath 

  • Golden eMule
  • PipPipPipPipPipPipPip
  • Group: Members
  • Posts: 790
  • Joined: 02-August 04

Posted 24 June 2010 - 04:35 PM

View PosttHeWiZaRdOfDoS, on 24 June 2010 - 11:45 AM, said:

Did you check if that happens with official, too? Would be worth a bug report, them.

I did a quick sample check, and there is a surprisingly large number of 50a with fastXS and reask@. Maybe I'll give the the loser mod a night off and do a match between 50a and 47c or something like that.

This post has been edited by James R. Bath: 24 June 2010 - 04:35 PM

Currently recommending and using: eMule beba 2.63
For slot control only, currently recommending: Tombstone Xtended 1.0 (or higher) if you absolutely must have slot control


Posted Image

Quote

Where there is a mule there is fuel. Where there is a stool sits a fool. - Winston Churchill

1

#19 User is offline   Stulle 

  • [Enter Mod] Dev
  • PipPipPipPipPipPipPip
  • Group: Members
  • Posts: 5804
  • Joined: 07-April 04

Posted 24 June 2010 - 07:19 PM

he was referring to something entirely different... thank you...
I am an emule-web.de member and fan!

[Imagine there was a sarcasm meter right here!]

No, there will not be a new version of my mods. No, I do not want your PM. No, I am certain, use the board and quit sending PMs. No, I am not kidding, there will not be a new version of my mods just because of YOU asking for it!
0

#20 User is offline   James R. Bath 

  • Golden eMule
  • PipPipPipPipPipPipPip
  • Group: Members
  • Posts: 790
  • Joined: 02-August 04

Posted 25 June 2010 - 04:01 AM

See first message in this topic for more background.

Tonight's skull match, since no objections were voiced, will give Morph some time off to wipe tears and fix things. The focus is on official versions 50a versus 47c. CA results courtesy of eMule beba v2.60:
  • 50a: 44 skulls out of 67 total (65.67%) :-1: (there were lots more, I just sampled the first 67 and paged quickly through the rest, which looked a little less skully, probably enough to get under 60%, but not 50%)
  • 47c: 8 skulls out of 37 total (21.62%)

Most of the skulls for both were FastXS. 50a also had a few Failed UL/DL.

Testing conditions:
Fakes/thieves: 0 50a, 0 47c. Excluded from the skull and total counts because they are different, unofficial mods.
Bad UL/DL ratio: 0 50a, 0 47c. Excluded from counts.
Queue/Known Clients at match time: 2452/2721
Number of active Downloads: 0
Currently recommending and using: eMule beba 2.63
For slot control only, currently recommending: Tombstone Xtended 1.0 (or higher) if you absolutely must have slot control


Posted Image

Quote

Where there is a mule there is fuel. Where there is a stool sits a fool. - Winston Churchill

-3

  • Member Options

  • (8 Pages)
  • +
  • 1
  • 2
  • 3
  • Last »

1 User(s) are reading this topic
0 members, 1 guests, 0 anonymous users