Re: L2J Changes
Posted: Sun May 04, 2014 7:28 pm
Damn, it looks so much like the query builder I've made in PHP... I've been thinking about making something similar in Java.
Your Javolution "customs" are irrelevant and unnecessary for an application conceived without threading issues... oh wait, you are right. Sorry about that. Still, I suppose none of you actually knows that javolution 6.1 is officially out since December 2013? I suppose that neither of you actually tried it...right?Zoey76 wrote:About Javolution, I do know the project lead there (we are facebook buddies ) and we are waiting for the stable 6.x version due we have customs in our outdated 5.x version
Bone-CP wasn't really what I was thinking about, but since you went up there, I can also say that Bone-CP performs if applications are correctly designed, and can keep their threading issues in check. Bone-CP access is so fast, mostly because the access to the database is that much closer to "raw". There is no "fat" to serve as a buffer or crash cushion for applications of sub-par design. c3p0 is, in that sense almost idiot-proof. So it is easy to understand why it is so much easier to stick with it.Zoey76 wrote:About c3p0, we had bonecp for a little while and it didn't work very well, so we didn't invest much time, there are other alternatives, maybe in the future we will change it, for now c3p0 is an stable, active, clean and reliable api that we have come a long time
So you have a roach infestation in the house, and you wish to use a nuke to solve it. You can't be bothered to use normal roach spray because it needs too much button pressing, while the nuke is just one press and you can sit back and watch it go down...? But, it is true. When (if...?) L2J manages to get that implemented, I'm sure it will go up by a thousand points in it's bragging value. The way it stands now, most of the people that use it are even ashamed to admit to use L2J. Private servers lie about not using it. It's a nice selling point to attract players. And before you think about it: I use L2J ... but as a plugin library for the purposes I see fit for it.UnAfraid wrote: The better option is to implement hibernate and use JPA approach but that requires hell of a changes to be implemented right.
So either we do it correct or we don't do it at all.
Excuse me, how many years ago that was...? Can you claim to do the same with current L2J clean beta/stable (take your pick they are the same, the way they are now)? How about specs. of the server where that ran...? No wait, I can guess. CPU load probably ought to have been through the roof on all cores, if you tell me otherwise I will not believe a word of it, because I know how it used to work back then. Obviously with a big enough system, you can hold/run whatever you want in it without problems. To use that as backup to claim that "everything is fine", is at least laughable. Servers running efficient software last longer, and keep running expenses to a minimum, representing overall more ROI (Return Of Investment). But I guess this might not be up your alley, of course, L2J is in the end just a nice educational project... among other things.UnAfraid wrote:Do you have performance issues with c3p0?
Weird i had over 3k players with c3p0 and i didn't had issues.
Neither nBd he had 9k (almost) with c3p0 and it was working pretty good are there nowdays such servers?
I don't think that answers like this will help at all, specially from autor of the thread, who I suppose is the most interested on it's answers in order to take the next decisions about the future of the project.Zoey76 wrote:It's like if jurchiks had duplicated itself
Neither this, because basically is said that since L2j can handle 3k/9k players, it needs no more improvement. It contradicts the fact of the project being still active, also.UnAfraid wrote:Do you have performance issues with c3p0?jurchiks wrote:There will be a difference in performance.
Weird i had over 3k players with c3p0 and i didn't had issues.
Neither nBd he had 9k (almost) with c3p0 and it was working pretty good are there nowdays such servers?
I agree, but in my point of view, such things happen because people have in mind that the Dev. Team work only for the project. This is why I think proposing their works to avaliation and being more open to discussions and decisions with the community should change this.Zoey76 wrote:My answer was short and trivial, because I see that people concentrate so much effort on stuff like:
X contribution(s) I made wasn't committed because a team member(s) didn't care or didn't had time, instead of thinking why other contributions were committed and what did they have better than that X contribution I use any chances to remind everyone that developers don't commit stuff!
Other stuff that is being mentioned a lot, and doesn't bring anything to the project is:
We did X feature or change and for L2j it's impossible to do because <insert any reason here>, that's why people makes and use forks!
Agreed. However, these are only examples of proposals that seem to be "required" by the community but constantly denied or not given so much attention as people want.Using Javolution 6, BoneCP, Hibernate or any other new implementation won't save the project, it'll be part of the improvements we have been doing for years.
This is your case and I totally agree with you, but not everyone. I see in many places people "bumping" or asking "is this going to be committed" or "anyone tested it" and so on... Perhaps because you are very skilled and may have had some other "reward" for your work with stuff non related to L2j. But think as a not-so-good developer, will it be encouraged to do something else or be "ashamed" thinking his work is useless?When I started in L2j, none cared about what I though or I did, after I worked for months and my work was ignored, some team member noticed some small parts of this work and gave me some room to work, after getting commit rights I ended up committing my own work shared months ago.
This means that anyone that keeps trying will eventually get it's reward.
That's another thing... Your part is done and it's clear, at least it should be, for everyone. Your proposal of "renewing" the team accordingly to their work fits very well here.I also pushed to include and defended many users that for different reasons weren't welcome on the team and they did a great work and some still do.
I have to say that many members wasted the chance they got when we brought them closer to the team, but of course none wants to talk about that.
That's why I say talks should be less personal (for this cases, there's skype, facebook and etc). People usually aren't grateful for the help they get, but very hateful if the don't get it as desired. Nothing specific to Open-source, in my opinion.Open-source is hard, I barely get a thanks for my work once a year, mostly people complain and demand stuff to be done or fixed, some even hate me (without knowing me), not everyone is up to the tasks.
Actually, nobody has said anything about that in this thread yet, but since you brought it up -- we all know that their contributions were committed because they were personally closer to the team leadership, with some rare exceptions (minor HTML fixes and the like).Zoey76 wrote:X contribution(s) I made wasn't committed because a team member(s) didn't care or didn't had time, instead of thinking why other contributions were committed and what did they have better than that X contribution I use any chances to remind everyone that developers don't commit stuff!
@Zoey - I think you might be misunderstanding a few things people are trying to voice on this topic. Just this sentence alone evidences it. The sentence itself isn't false, per-se, because I know the majority of volunteers of open-source projects do this at least for "fame". Like, "I did <whatever> and this <insert popular project here> uses it!". There are cases, however, that couldn't care less for this and their "reward" is to see the project they are interested in evolve in a better path. This may not even have anything to do with writing code. The way I believe you are understanding what you are commenting with this sentence, is as if you are narrowing it down to people seeing their commits getting accepted, and have their names "immortalized" in the L2J repository logs. I've done a bunch of past snippets and fixes of easy integration into L2J in the past, I didn't see my name on any credits, and I still recognize them on today's sources. What I was actually disappointed, was that these snippets were not used in the way I designed them to, and as such, they may be actually causing borderline-case problems. I couldn't care less about my name on the credits, so I would appreciate if people don't start throwing that argument into discussion.Zoey76 wrote:[...] This means that anyone that keeps trying will eventually get it's reward. [...]
This is, spot on, part of the image the community has of higher ranked L2J developers (there is a lot more, but I'm doing my best to keep things civil...). This should be changed, for real, with true action. Not just opening "pretend-we-care" topics, and then stuff stays the same. Even assuming the intent of changing is true, hence why I broke my silence here, if no tangible moving action come of this within a reasonable time frame, this will hurt L2J... no, this is already hurting L2J enough to cause "ranked" people to show mild "interest" in topics such as this. Unfortunately, despite the good momentum the topic is gaining, I haven't seen any real sign of any intention on taking advantage of this. For the time being, it seems this topic is raising the collective awareness of the community. It would not be smart to keep disregarding everyone's comments as "rants", which IMO is the idea that is rather getting across from the developers to the people commenting on the topic. I am trusting this is not the intention of the L2J team, but this might actually be aggravating things. Taking care about this is also a must!jurchiks wrote:[...] we all know that their contributions were committed because they were personally closer to the team leadership, with some rare exceptions (minor HTML fixes and the like).
[...]
Also, watching from aside reveals who is being preferred over whom quite fast. You don't see that so well when you're the one doing the committing.
L2j will never be paid. It's open-source. It's forks or mods can be payable but will not be developed by L2j. Customs will be made even by me and shared as soon as we get some stuff cleared in core so customization will be able to be done fully DP side.t0p3a wrote:Customs would be nice and even a l2jserver paid version with many custom stuffs which should be cheap ;p to get.