Petition to remove the /reload command

/Signed
Reload uses 15MB RAM on my server
However, if it dissapears, I would like a function to enable/disable only 1 plugin.

If you have one plugin written specifically for one server to do everything desired on that server, then sure, one plugin isn't a bad idea (though I still don't think it's a good idea). However, those plugins would not be worth releasing; chances are no-one other than that one server would ever want them.

The problem is that he is running "big" servers on those, it doesn't take that much to realize when you run the entire server on a plugin, there will be problems.

Yes a lot of plugins do handle reloads badly but developers should make sure to dispose of their data properly and instead of removing the command completely Bukkit should try to recode it to make it more efficient.

Reloads are actually very effective, I personally hate having to restart a server and I find that sometimes during testing or updates having the restart the server ever time is a pain.

By the same token, it's just as bad to make Spout a single plugin. Why not break it into 100 different plugins? Because I know that as an experienced developer, there is a very significant benefit to code simplicity. I have 3 servers - each has it's own jar file with code referencing a shared library compiled into each jar. When bukkit updates, I download it to my computer and run a quick test off my local versions of the jars. Takes about 5 minutes. If they don't work (which has never (yes never) happened), I take about 2 minutes and tweak the code to make it work. If I had 10 plugins for each of my 3 servers, instead of a 5 minute test for each bukkit update, it would take 30 minutes or longer. Simplicity. I maintain one code base. You are saying I'm inexperience because I should be maintaining many more?

True. Now if only my plugin was a "massive structure of easily breakable things" like your plugins that "were horrible, many bugs and glitches" (quoted directly from your single plugin, said by you).

Gah! Somebody tell me why I feel compelled to respond to crap like this!

Holy Crap! Somebody here has a brain! Thank you, thank you, thank you!

Yes, I did say that, and if you had the ability to read, you would see the plugin it's concerning was my first.

It spits out errors because it's not coded to support reloads. This is because the reload command does have a lot of issues with it. Surely you wouldn't complain to a Car Company about not having any milk or eggs for sale inside their facility, right? Spout chooses to ignore the reload command for the several reasons listed--they're not covering anything up.

Furthermore, this is not 'showing disrespect to the Bukkit team', it's giving constructive criticism. He's stating that there's an apparent flaw with the concept of a 'reload', and that if it's going to be done poorly, it'd be better off not being a feature of Bukkit at all.

It may be convenient on the outside, but it does a whole slew of things wrong internally.

My point was that you assumed plugins were "massive structures of easily breakable things" because you are (in the best of light) an inexperienced programmer. I am now assuming that you are a script kiddie who has been dabbling in java code for a few months, writing bukkit plugins. I've coded for a living for the past 7(ish) years. Let's hear your credentials.

Certified in HTML5, CSS3, used PHP for 3 years now, and been "dabbling" it Java.
It's easy to say, no matter what platform, it's not a good idea to base your entire thing of a central point that is easily breakable.

/Signed

Is it normal that static variables do no seem to be reload-persistent?

Thats says that plugins are unloaded at /reload, and not what they said at the begin

Personally I find it really feckin' ennoying to have to restart the server because of spout.. need to sort myself out a test server

As a server administrator myself I've come to realize one important thing. /reload is very broken and never works right on quite a bit of plugins. If you are going to maintain a server you should learn to use a test server, put together an update on the test server and when its ready take down your main server to push the update.

The /reload command, and the way it works, is necessary! I use it to update plugins I've made without restarting my server. I can make a quick bugfix, then upload and not worry about players raging at 45 restarts in 10 minutes. I think if people really care this much about it, then get a plugin that disables /reload. It's not very hard. 5 lines at the most. Come on, people. Do something better with your time.

I'm very against this. Reloads help me fix bugs rapidly. If I'm really stuck on a bug I'll try something, compile, install, reload, repeat. That means obviously all my plugins are reload compatible. It's a really really handy developer tool for test servers, and probably works for small servers with few plugins as well.

Also look at it from the bukkit staff's view. When you are as big as bukkit, you simply just can't remove something with out providing a better alternative. So the only thing this thread would accomplish, if the bukkit staff were to fulfill its demands, is tick off a portion of the community by removing a small feature of bukkit that not everyone uses.

Yes, I agree it does not work for the "big server" who has 100+ plugins, and 50+ members online, however it is not a problem for them at all if they choose not to use it. However, the bukkit community isn't solely composed of big servers. The big servers know not to use it. The people who cannot use it really gain nothing from reload getting removed, and everyone else loses something useful.

/signed

This should be made as an option in bukkit.yml with a default of false.

@EvilSeph

What's the point in that? If people don't want to use it, just don't use the command, adding a toggle doesn't achieve anything.

It could potentially be dangerous with plugins that don't support /reload so only developers would really use it and server owners who have decided to take the risk.

Indeed, obscuring the option would scare off most future server owners from it while giving current ones their features

Against it

While the points layed out are all valid, the reaction to remove or hide or obfuscate the command is rather damaging.

/reload is and should be purely considered as a "development command". In that hindsight I agree, DO NOT USE IT FOR PRODUCT/LIVE SERVERS. Easy as that.

It is the server owners responsibility to know the tools they are using, to know the commands they are using. Ignorance is not an excuse. Hence, if You know, the command is damaging DO NOT USE. If You do NOT KNOW what the command does DO NOT USE IT.

Also if You find, that the plugin You are using, is damaging or any other way "damaged" You do not have to use it. Alternative is You notify the plugin maintainer of its shortcomings and get it fixed. Poor plugin design or implementation is no excuse to remove the /reload command. It is the plugin maintainers/developers responsibility to have a high quality product. It is the server owners/operators responsibility to know what they are running.

Lag? Shouldn't be in this topic anyway.

For me as a dev, /reload is the most used and most helpful command of them all. However, I wouldn't dream of using /reload in a final running populated server. That wouldn't be just stupid, wouldn't it ?

/signed

I have 3 plugins that do not work after a reload, it's annoying and I've contacted the developer. Reloads have caused a lot of problems and lag for me.

I don't think they should take it away, I think they should recode it to be more efficient.

/signed

/signed

Memory leaks as well as numerous other problems makes this unnecessary.

The point is: it doesn't work properly.
It should either be removed or replaced by an improved alternative, but since I don't see them recoding the command any time soon they might aswell just remove it.
Server owners are supposed to know that reload doesn't work properly, but the truth is that many simply don't.

To those that say the plugin developers should just take care of the problem:

Although it wouldn't fix everything, we should at least get a boolean indicating whether it is a reload or stop in the onDisable function.

Something tells me you don't quite understand the meaning of the word "unnecessary"...

For my server, in the most... Reloading everything can be done by just restarting, and that seems to be a lot more safe.

Mm, maybe it's something else you didn't understand... let me clarify properly.

I won't confirm or deny that /reload is unnecessary (that's really a matter of opinion); however, it's not unnecessary as a result of memory leaks as your post implies. Memory leaks don't make it unnecessary; they might make it unsafe, but not unnecessary.

That is my point. I would rather restart the server properly than subject it to problems.