MadJack

Hey rilex, What version of Vista and IE you have installed? Are latest patch installed? I've tested the gadget on two systems and it works fine. No bleeding. The text should cut at a certain point. It will not wrap mainly because of the text sent back to the gadget isn't exactly the way it is on github.com and because of that, it won't work as it does on the site. That's also the main reason the flyout has that size. If I shrik it, you won't see much. Three of four words at most and I would have to resize the fonts to an even smaller size. As for closing, click on the same source you used to close the opened flyout. I can take a look, but I'm not sure I can force the flyout to close with a button. Gadget aren't much. Javascript files, CSS and a XML file. It's not as customizable as a "real" app... Anyway, thanks for your input. I will check if there's something I can do to tweak the display. I'll get in touch with Github's mailing list to see if I can have them change the way the commits' text is sent so they can wrap properly. But even then, it would be at the same width as the web page... I guess I could come up with a scheme to manually wrap those lines, but I'm wondering if it's worth all the trouble. The gadget is something that is used sporadically... I'm not promising anything

I'm happy with the result, that's the most important I guess. Hopefully those that got it found it helpful. If not, well...

Huh? What's that supposed to mean?!

Got about 30 downloads but not a single comment? Does it suck that bad? Or is it that good? ROFL

Hey everyone! This message is to let you know that I've made a Windows Vista sidebar gadget that will monitor sources on Git. It was made with MaNGOS in mind and comes preconfigured with three sources. (MaNGOS, Mad's MaNGOS and 303-willcrashforsure) The gadget is beta stuff and I would ask you not to distribute if for now. Since I'm using some Git assets (images and CSS), I'd like to wait until I get the proper permission before letting this thing loose in the wild. What this thing won't do: - Log you on Git. - No abilities to "use" git in anyway whatsoever. What it will do: - When you have configured the gadget, it will fetch the configured username/project/object (master or branch) and display the commits history of the clicked-on item in a flyout window. - You can set up to 5 sources to "monitor". How does it work? - The gadget fetches commits based on GitHub's API and displays the results after it has been through an xls template for formatting. Known issues: - None that I'm aware of as of now. That's why it's in beta! Future development: I'm not sure I will implement the followings, but if people would like it, I might... - Autochecking based on a timer (some of the code is already there). -- An icon would change color to inform the user of a newer commit. - An original icon for the gadget, I'm no artist so unless someone is offering his services... Mostly the features I wanted are there, only the timed checking I'm unsure of. I don't want to end up spamming GitHub... At the start of the project my goal was to monitor Git/SVN but since svn doesn't really have a central repository (that I could find anyway), it's pretty hard to come up with a scheme. I know that Sourceforge offers one such place but I couldn't find an easy and rapid way to asses its API. If anyone is in the know, drop me a PM or an email and I'll take a look at it. Anyway... If you have any question, just ask. To download and install, simply click on the link at the bottom and open it (Vista should recognize the .gadget extension and prompt you for its installation) once you have downloaded it. Get the file here: YourSources.gadget If you have feedback or issues, either post it here on in a PM or by email. I'll get back to you ASAP. Thank you all.

You seem to know stuff about that Derex, why don't you give it a shot? If I had the knowledge and, mostly the time, I'd give it a go.

Well, let me just say that my intention was to have the possibility to load an alternate scripting library (whether on *nix or windows). That is the only goal here. Now, if loading/unloading/compiling doesn't work as intended, there is nothing I can do about it. My mod doesn't break any of those things and as such, I don't feel the need to fix what is ailing MaNGOS in the first place (related to that particularity). This is the work of someone who knows about that kind of stuff. I think the question here is: "Did the problem existed before my changes?" My answer is: I don't know. But if it doesn't work as intended now, I know it's not something in my code. If someone thinks otherwise, please give me clear explanations. On the other hand, if someone could test it out, I'd be grateful and I think it would put that whole thing to rest. Let me reiterate here: The patch is about loading an alternate scripting library independent of the platform if an entry is tagged into the configuration file (AltLib = ""). Any other issue should have a post in the BUGS forums unless _my_ code breaks something. I'm sorry if I sounds like I'm dismissing your point Derex, but I still don't see how my modifications make an impact on what you are suggesting.

LoL You make it sound like I broke the library loading code! Which is funny. So am I to understand that it didn't work before I made my changes? sLog.outString( "Initializing Scripts..." ); - if(!LoadScriptingModule()) - exit(1); - + /// Alternate script engine library loading + std::string AltLib = sConfig.GetStringDefault("AltLib", ""); + char const* sAltName = AltLib.c_str(); + + if(strlen(sAltName)) + { + /// Load alternate library + sLog.outString( "Loading script engine library: %s", sAltName ); + if(!LoadScriptingModule(sAltName)) + exit(1); + } + else + { + /// bailout w/ default lib + sLog.outString( "Loading default MaNGOS script engine."); + if(!LoadScriptingModule()) + exit(1); + } + ///- Initialize game time and timers NOTE: I used tag because it's easier to see the changes when colorized.Please explain how an IF/ELSE, a Screen Output (sLog.outString) and a Variable check would break the library loading/unloading/compiling? I'm not trying to be sarcastic, but I fail to understand how these changes would impact the way you say it would. These are the only changes I made in that file. The other modification is in mangosd.conf.dist.in The rest hasn't changed one bit (literally). I just can't connect the dots here... Please explain how my changes break MaNGOS library stuff.

Derex, I honestly don't know but it should. I don't have access to a linux partition so I can't test it out. I think it should work because all that patch does is check if the new tag is not empty. (If it's not empty, it tries to load the library named in the tag.) The rest is pretty much the same. So if the .so loaded before, it should still do with that mod. That patch doesn't change the way a library is compiled. That's unless I misunderstood what you meant.

Nice post TheLuda. Hopefully this will stop the madness!

Wouldn't <branch> be how you "defined" it? What I mean is, master here is set as: [branch "master"] remote = origin merge = refs/heads/master and origin is defined as: [remote "origin"] url = [email protected]:MadJack/mangos.git fetch = +refs/heads/*:refs/remotes/origin/* So, wouldn't that mean the same as using my remote repos? Since Git wants to have a branch, I just can't give it "origin" with pull, but that amounts to the same thing right? I'm not trying to argue or anything, just understand... Oh and thank you!

Wow! I'm impressed! 2hrs and a half and it's still not locked! QUICK! Got buy yourselves a lottery ticket!

For those interested in that patch, if any, you can use the followings: Commit e0deeef2cad92ecfb794a5cad538b99388a60b84 <-- world.cpp changes and Commit a64235050be66988f4dfd564213d2dbc1dcae69e <-- mangosd.conf.dist.in

I give it 5 mins... After that, it'll be locked. No, it doesn't. Not right now. When it will, it'll be for 3.0.3 but who know when...

Finally got it to work as I wanted! I did a lot of "mistake" in interpreting what I thought something was doing and its actual purpose... Anyway, all changes are now reflected in my fork'ed src of mangos and the main commits made by the team is also there. That way I can use my changes and still get the updates without suffering with a cut/paste syndrome Thank you freghar2 Wanted to add stuff that I learned along the many reads I've done... So after I have done my changes (in this example I had committed and pushed them before getting the main dev's commits) I did this: git checkout master <---- Which essentially means: USE master git pull git://github.com/mangos/mangos.git master <--- which means: get the changes FROM there and merge it THERE (master) git add . <--- add everything that has changed starting in this directory (changes from the previous step) git commit -a <--- commit ALL changes (locally) git push <--- send to origin (remotely to your forked version) Is that a good interpretation of how this works?

The above worked nice. Thanks. But before I had it working I decided to start with a clean slate and deleted the whole content of my local mangos directory. I was able to clone and download my forked branch (from mangos/mangos), made the changes again (I backed up World.cpp), committed them and finally pushed them. All is good and is reflected where I want it to be (ie: MadJack/mangos). But... (You knew that was coming right? ) My problem now is that, for some reason, when I try to "Git pull/merge/fetch" from mangos/mangos, it doesn't want to (so I can have the latest commits). MaNGOS is defined as: [remote "mangos"] url = git://github.com/mangos/master.git fetch = +refs/heads/*:refs/remotes/upstream/* So when I try: "Git merge mangos" I get: fatal: 'mangos' does not point to a commit... PS: I'm on Vista but I use Git Bash. I've used *nix before and I've been with computers most of my life... I have read a lot of stuff on Git but it's a learning process and some things are still blurred in my mind.

Yes jolan, that's exactly what I expected: my trunk/tree to reflect the changes I committed and pushed (and the merges I would eventually do with mangos's main tree) Let me take an example. If Vladimir makes an update and a commit, it is on mangos/mangos tree. Why isn't mine in my tree? Why is it only in that branch of mine, which incidentally, I never wanted in the first place. My understanding is that when you do a checkout, it simply downloads the checked-out tree/branch locally... I'm confused now... All forgiven. Don't worry. AFAIK only Vladimir or one of the other assigned persons on mangos/mangos can commit others changes to that tree and that's not what I expected from my commit/push. I don't want my changes to be in mangos/mangos, only in madjack/mangos... So, if I follow your instruction above: git push madjack mangos:madjack or git push madjack/mangos madjack:madjack ? LoL Sorry if I sound thick here but this is confusing the hell outta me!

http://github.com/mangos/mangos/tree/master <-- MaNGOS TREE http://github.com/MadJack/mangos/tree/master <--- This is MY TREE In mangos/mangos, you have one BRANCH. (303-willcrashforsure and master of course) The BRANCH above is reflected in my tree. That's fine and normal. Now, as I was saying, I committed my changes, THEN pushed them. From what I understand, my tree should reflect those changes but they are not. My problem is that all those changes I made are in that branch of my tree (http://github.com/MadJack/mangos/tree/madjack) and not in the "base tree", for lack of a better word... Before anyone asks, yes, I am sure I pushed those changes. Before I pushed them, they were not in my branch so I guess I made a typo or something along the way or, again as I asked earlier, maybe that's the way it works in the first place. Btw, stop the patronizing... It's not helpful.

I understand but I forgot to mention I did push the modifications... but those ended up in my branch in my tree... I wanted those in my tree, not in a branch...

I'm sorry!!! hehe Anyway, I need a hand here. Pretty please? Alright. Let's start with the obvious. I didn't follow any tutorials from here. I read GitHub's help files and one particular tutorial I found on the some corporate site which had a nice layout and explanations... Next time I'll know... Well, it might just be me too! Most probably is. LoL So... I forked mangos/mangos and got a nice MadJack/mangos, which is good. I make some changes and commit them to my branch after going through hoops and limboing across "git this here there"; but my commit ended up in a BRANCH of my fork... I succeeded to merge the mangos/mangos with madjack/mangos but MY changes are not in MY tree! It's in a branch of my tree... LoL What I'd like to do is have my fork, make my changes, commit them and merge the mangos/mangos to my tree. Pretty easy I think... OTOH, it might be working as intended, but I suspect I made a mistake along the way. Help very appreciated! Thanks.

The way I see it is, when the lib loads, it checks if its version is up to date in the table. If it is, no change, if not, it updates it. It's a bit hackish and would be preferable to poll it directly from the app to the lib but for now it could work.

Short answer: Yes. It would work with Linux. Long answer: This has no bearing whether it's Linux or Windows actually. It uses the same function as it was already using, all I did was add an entry in the config. If that entry (AltLib) is empty, it uses the default dll/so if not, it uses the one you defined in the .conf.

There's one thing I'd like to see when I ".server info": .server info MaNGOS/0.12.0 (Info STUFF)... MaNGOSScript version... or scriptsdev2 version or whatever script library used ... other stuff Anyway, I think you get the picture. But I'm wondering how could this be achieved. Iterating through the database looking for tables or other DBs for whatever "program" would be impractical or slow. So... I'm wondering if it would be possible that each "addon" could use a new table we could create with things like addon name, version, build #, etc Or we could define fields that would be updated by other programs as they are loaded. I'm suggesting that because sometimes I compile this or that stuff that I add up later but I'm never certain which version runs on the server. Would it be interesting/useful to have such a thing? If so I could work on that.

I juggled with the idea to load the default script engine if the AltLib failed to load but opted to leave it as is because it would be easier to "fix" the spelling in the conf file after it exited than to wait for the whole thing to load with the default lib then exit. The alternative would be to stop and query the user if they wanted to continue loading using the default lib...

I thought about it as a plugin connector, but IMO I'm not sure it would be relevant to have an entire "plugin" directory dedicated. From what I could see, I haven't spend a lot of time looking at that specifically but, wouldn't using two libs clash? Even if they do different things? If not then pardon my ignorance.