L2J Server

Im just wondering. L2J has become pretty big project, and I guess many developers are working on a couple of patches at one time. How do you do that locally? I mean, when you have two different things you are working on, and then you want to commit changes, do you manually remember which files belong to which patch and needs to be committed? Or do you use branches locally to keep several patches away from each other?

We use subversion, all of us have a local version of L2j on their computers and work with it, once we have something ready to be commited to the main SVN we check if our changes affect files modified since the last time that we updated our local version, if those files has been modified by another developer, we update our local version and check those changes and see if the patch that we were working on need a code review, if it doesnt we just commit our new modified file.

On collaborative work the coordination is essential usually L2j Developers know what others are working on, and keep on contact, if a huge change is going to do main core SVN we ask other developers to hold their commits for a while once the big change is done.

Aha. I see. But what I mean is how do developers do it personally. I guess any developer is working on few patches at one time. He (or she) started to do something, then got into some problem, started another patch. Isn't that true? If so, how do you not get lost between all different patches you are working on.

I am asking this, because I keep my own live server. I applied few patches I found here that were not committed to L2J. Then I worked on my own patch and wanted to post it here for testing/maybe commit, but I got lost which files are modified by me, and which are modified by older patches that I applied.

So do you use SVN branches locally for this kind of situations?

what I do if I really get stuck and have to start working on something different, is make a patch of what I've done so far, store in a .diff and revert it from my project, then start working on the next thing

when I want to get back to it, I apply the patch again(manually if necessary for line changes) and get back to work on it.
abit tedious I know, but avoids having to worry about committing wrong things

for small changes i just save the code in a .txt file
for bigger, or patches by others i do like Probe does

i store it on external hdd in rar/patch and forgot about it

for bigger things a clean checkout sometimes specially if changes are all over the place