Code Update - Discussion

[Nard]

Roll back or not?
Poll runs till Sun Sep 29, 2013 1:29 am

As there will be no rollback, topic locked.

[AnonDuck]

First let me ask that everyone keep the topic in News (http://forums.themanaworld.org/viewtopi ... =1&t=17729) free of discussion. That topic should be used for bug reports and issues only.

So as you may be aware, TMW has upgraded the server to a new version. There are some bugs and issues you should be aware of. See the news on http://themanaworld.org/

I know that you, as players, are wondering why the devs went and broke all sorts of things! Are they insane? Do they like torturing players? Did a cat jump on their keyboard? No!

Well believe it or not.. these changes were actually needed! Without getting too technical, the old game server was a MESS. The code was very hard to understand, and absolutely riddled with subtle bugs and improper behavior. Rather than add new features and make it even MORE of a mess, o11c decided to clean up the mess first.

That's what this update is all about - cleaning up and live-testing the code so we can eventually add new features and behaviors that will make TMW better. If your car was held together with nothing but tape and glue.. Wouldn't you want to remove all that and bolt it together properly before putting sweet rims on it? Same deal.

About the bugs: They happen! When I wrote the anti-spam system I tested it for an entire month with 20+ players.. And when we put it live on TMW it randomly banned people, doh! For the huge amount of changes that have been made to the server it's nothing short of AMAZING there are so few problems! Compared to previous server-code updates from years ago this one has actually gone very, very well.

I think o11c deserves some major recognition and credit for spending literally months of his time working on this and doing such a great job. These bugs will eventually get fixed and everything will be back to normal, and with the cleaned up code we'll be able to make TMW better than ever.

About the poll and the closing: It was very premature to make a poll on such a thing without informing the players WHAT is going on and WHY. The poll was therefore closed. Please note that the poll was tied 50/50 between wanting a rollback and moving forward with what we have. This whole thing could have been handled better by TMWC and we apologize for that.

[melkior]

Nard's post moved into this thread.

[Nard]

About the poll and the closing: It was very premature to make a poll on such a thing without informing the players WHAT is going on and WHY. The poll was therefore closed. Please note that the poll was tied 50/50 between wanting a rollback and moving forward with what we have. This whole thing could have been handled better by TMWC and we apologize for that.

If you don't care to have real players on the server, I can agree with you. If you want them, take care about them.
Your others arguments would stand if themanaworld.org was not a production server. This kind of bug appear only when there is no team working, but this is no news.

[WildX]

First let me ask that everyone keep the topic in News (http://forums.themanaworld.org/viewtopi ... =1&t=17729) free of discussion. That topic should be used for bug reports and issues only.

So as you may be aware, TMW has upgraded the server to a new version. There are some bugs and issues you should be aware of. See the news on http://themanaworld.org/

I know that you, as players, are wondering why the devs went and broke all sorts of things! Are they insane? Do they like torturing players? Did a cat jump on their keyboard? No!

Well believe it or not.. these changes were actually needed! Without getting too technical, the old game server was a MESS. The code was very hard to understand, and absolutely riddled with subtle bugs and improper behavior. Rather than add new features and make it even MORE of a mess, o11c decided to clean up the mess first.

That's what this update is all about - cleaning up and live-testing the code so we can eventually add new features and behaviors that will make TMW better. If your car was held together with nothing but tape and glue.. Wouldn't you want to remove all that and bolt it together properly before putting sweet rims on it? Same deal.

About the bugs: They happen! When I wrote the anti-spam system I tested it for an entire month with 20+ players.. And when we put it live on TMW it randomly banned people, doh! For the huge amount of changes that have been made to the server it's nothing short of AMAZING there are so few problems! Compared to previous server-code updates from years ago this one has actually gone very, very well.

I think o11c deserves some major recognition and credit for spending literally months of his time working on this and doing such a great job. These bugs will eventually get fixed and everything will be back to normal, and with the cleaned up code we'll be able to make TMW better than ever.

About the poll and the closing: It was very premature to make a poll on such a thing without informing the players WHAT is going on and WHY. The poll was therefore closed. Please note that the poll was tied 50/50 between wanting a rollback and moving forward with what we have. This whole thing could have been handled better by TMWC and we apologize for that.

Can someone make MadCamel the new TMWC PR agent? Seriously, this is how you talk to players.
_{Yes, you heard me.}

[o11c]

Can someone make MadCamel the new TMWC PR agent? Seriously, this is how you talk to players.

Well, since MadCamel if no longer wiling to make any sort of contribution to TMW since he might get this sort of backlash directed at him ... I guess he's a safe spokesman.

Please, tell me what more I could have done. I released this code 2 months ago, announced it on the forums, and tested a *huge* number of things myself, including, I might point out, trading.

If you want to make this kind of thing less likely, be on the test server and do all the things. But please realize, though more testing by people who follow different patterns, it is impossible for any amount of testing to prove the complete absence of bugs.

Frankly, given how bad TMWA's code is, I find this code update a miraculous success. Remember, last time we restarted the server using the *old* code it crashed in less than an hour.

TMW will never become great if we can't ever change anything. Particularly, it will *never*:

1. have native guild support
2. implement professions
3. be rebalanced
4. be able to have more than about 100 players online at a time without major performance problems
5. provide the content developers with the tools necessary to make their jobs easier
6. be able to add any serious features at all
7. offer *any* sort of guarantee about data corruption or avoiding crashes

[WildX]

Can someone make MadCamel the new TMWC PR agent? Seriously, this is how you talk to players.

Well, since MadCamel if no longer wiling to make any sort of contribution to TMW since he might get this sort of backlash directed at him ... I guess he's a safe spokesman.

Please, tell me what more I could have done. I released this code 2 months ago, announced it on the forums, and tested a *huge* number of things myself, including, I might point out, trading.

If you want to make this kind of thing less likely, be on the test server and do all the things. But please realize, though more testing by people who follow different patterns, it is impossible for any amount of testing to prove the complete absence of bugs.

Frankly, given how bad TMWA's code is, I find this code update a miraculous success. Remember, last time we restarted the server using the *old* code it crashed in less than an hour.

TMW will never become great if we can't ever change anything. Particularly, it will *never*:

1. have native guild support
2. implement professions
3. be rebalanced
4. be able to have more than about 100 players online at a time without major performance problems
5. provide the content developers with the tools necessary to make their jobs easier
6. be able to add any serious features at all
7. offer *any* sort of guarantee about data corruption or avoiding crashes

Well, at this point everything's sorted out. Once people know what's happening and what is going to happen they should just let you do your job and stop freaking out calling for rollbacks like a bunch of angry chickens.

[o11c]

I might also point out that the new server code automatically makes backups, which should make a rollback - if it turns out to be really necessary - much easier.

[Hello=)]

Well believe it or not.. these changes were actually needed!

Needed for whom exactly? o11c himself? Sure, its nice if programmer's life getting more convenient. But its not okay if it requires to screw up whole bunch of players in process for some odd reasons.

And this game exists for about 10 years so there is working gameplay, alive players community, some content which is OK and so on. Failure to recognize this dumb. Treating all players as alpha testers? They would hate you for 100% valid reason. Demanding to test something without obvious motivation? Going to FAIL. Claiming it's alpha? Futile. It exists for about 10 years, works and playable. So nope, it's not alpha. Its alive and running. Its not ok to break it in arbitrary ways "for good".

absolutely riddled with subtle bugs and improper behavior.

I bet new code would not be seriously better in this regard. Sure, if you delete all code, bugs would disappear as well. But anyway there were already some numerous new bugs and most of old bugs remained on their places I guess. At least those I know and can (ab)use.

Rather than add new features and make it even MORE of a mess, o11c decided to clean up the mess first.

Some years ago, Andrew Tanenbaum decided that internal OS architecture outweights feature set and whatever else. Then Linus Torvalds used his Minix to write some strange and weird thing known as Linux. Quickly facing comments from Tanenbaum about the fact Linux arch suxx and somesuch. Time passed. So, why the heck you dont use Minix with it superior architecture these days? But quite many people use Linux with "old" and "bad" architecture. Maybe it's because features and overall usability count as well?

That's what this update is all about - cleaning up and live-testing the code so we can eventually add new features and behaviors that will make TMW better.

The only issue here is that different people can have different ideas on what "good" and "better" means for them. Getting rid of players who would be unable to log in does not looks "good" for me, for example. It looks like some very minor gain on one side and chances to lose players on other.

If your car was held together with nothing but tape and glue.. Wouldn't you want to remove all that and bolt it together properly before putting sweet rims on it? Same deal.

At the end of day it's getting from city A to city B what really counts. For most people car just some tool to do so, they dont care about internals as long as it getting them from A to B.

About the bugs: They happen! When I wrote the anti-spam system I tested it for an entire month with 20+ players..

Sure, they are. But some things could be foreseen and some could be caught by tests.

I think o11c deserves some major recognition and credit for spending literally months of his time working on this and doing such a great job.

Sure, impressive job. But his ways of thinking and doing things are strange and often happen to be hostile or unpleasant from point of view of ingame players. While I dont have anything personal against o11c, I really dislike some of his actions.

These bugs will eventually get fixed and everything will be back to normal, and with the cleaned up code we'll be able to make TMW better than ever.

Except that number of players would fail to log in to game. While gain from login system screwing is unobvious to say the least.

About the poll and the closing: It was very premature to make a poll on such a thing without informing the players WHAT is going on and WHY. The poll was therefore closed. Please note that the poll was tied 50/50 between wanting a rollback and moving forward with what we have. This whole thing could have been handled better by TMWC and we apologize for that.

You would dislike to hear what some players think about how things were conducted and their private opinions about some devs and/or TMWC. I'm still think that failure to evaluate how your certain actions would look after landing your actions on players heads BEFORE backslash happens is not very smart approach and really weird lack of foresight.

I'm personally would call TMW's bug #1 lack of proper interaction with players community, failure to foresee how players would accept "good" changes, ignorance of feedback, lack of gameplay experience. Overall, TMW devs tend to have bad knowledge of gameplay and/or community. And it happens to be a problem for both. Players happen to feel pissed off by some actions. Devs face backslash then. Maybe it could be better?

[o11c]

FWIW, the case-sensitivity made it possible to reduce the startup time of the server from about 1 minute 15 seconds to about a second, and the startup time for the char-server (no player visible change there, but same basic effect) from over 5 minutes to about 5 seconds.

[WildX]

I'm personally would call TMW's bug #1 lack of proper interaction with players community, failure to foresee how players would accept "good" changes, ignorance of feedback, lack of gameplay experience. Overall, TMW devs tend to have bad knowledge of gameplay and/or community. And it happens to be a problem for both. Players happen to feel pissed off by some actions. Devs face backslash then. Maybe it could be better?

This is a major problem for most games, actually. While it would be nice to see more effort put into avoiding it, it's kind of inevitable to some extent.

[var]

FWIW, the case-sensitivity made it possible to reduce the startup time of the server from about 1 minute 15 seconds to about a second, and the startup time for the char-server (no player visible change there, but same basic effect) from over 5 minutes to about 5 seconds.

You could lowercase user ids data in the data file (as you shown you can edit datafiles) with a script and use that data,
then lowercase user id data coming from client and check that in the lowercased user ids datafile having the same result and keeping case insensitivity.

Temporarily hashmap user ids information (until data will be stored into a real database) into a file or memory would speed up log in as well.

You may have lost players but at least they will be happy to know the server start faster.

[o11c]

You could lowercase user ids data in the data file (as you shown you can edit datafiles) with a script and use that data,
then lowercase user id data coming from client and check that in the lowercased user ids datafile having the same result and keeping case insensitivity.

That's exactly what I do for the character database, and I considered it for .

The difference is that character names have always been forced to be case-insensitive. Accounts have not been. For the login database, that would mean outright deletion for 117 accounts and there is not realistic way to avoid that.

Are you suggesting that I should have deleted those 117 accounts instead?

[var]

Are you suggesting that I should have deleted those 117 accounts instead?

Inform the users using forums and game and any possible way to reach them, talk with them and tell them their accounts would be renamed, you love to delete stuff dont you?

Use the thing called comunication.

[tux9th]

Are you suggesting that I should have deleted those 117 accounts instead?

Inform the users using forums and game and any possible way to reach them, talk with them and tell them their accounts would be renamed, you love to delete stuff dont you?

Use the thing called comunication.

This has been discussed at full length here http://forums.themanaworld.org/viewtopi ... =1&t=17620