Archive

Archive for November, 2009

Bug unplugged

November 25th, 2009 14 comments

Since 1.23 is out a few people are experiencing a nasty map bug, or at least it manifestates in the map view – it exists elsewhere as well but is less noticeable. I crafted a small Javascriptlet to get around this bug, but it’s not a solution that can be applied on the productive system. So here’s a javascriptlet that can help you temporarily:

javascript:WMap.mapData = new MapData(); WMap.scroll_map_to_char(); void(0);

If you copy that line of code into your browser’s addressbar, you can bypass the bug for a while at least.

But now the more important question: What is the Matrix bug?

The answer is rather complicated and I would say that this is among one of the most notable bugs that I have experienced so far. Noteable because the reproduceabilty is extremely weird. The company does not have any computer where this bug can be reproduced, however I possess an Asus eee, a device that can reproduce the bug as I found out at home.

So on monday I took it to work (like always, I type a lot of stuff when I am on the train). I connected it with the company’s network and the first thing I noticed is that my local developer machine cannot reproduce the bug with my machine. Weird. But well, I am using a different webserver. So I connected to my collegue’s webserver and the bug worked as expected.

We tried the whole day to get behind this bug. We found out that it is not the fault of our PHP programs – PHP is never called. The client’s request simply vanishes at some point. We used wireshark and looked at everything – but we could not make any progress here. The requests arrive at the webserver but there, the requests simply die. At the end of the day my netbook’s battery was drained (they lastet for more than 6 hours).

The next day I brought my charger with me, just to get over the day a bit longer. I fired up the browser, requested the page in question – and it worked. Out of the box. We couldn’t figure out what was different. The server was unchanged, the software was and no matter how often we pressed F5, we couldn’t get the bug working again.

Until I unplugged the device and refreshed the browser. And there the bug surfaced again. I put the netbook on the charger and the bug vanished. I am pretty sure that this must be hard to believe, as I said, this is one of the weirdest bugs that I’ve ever faced, but this is true! It is only reproduceable if the netbook is running in battery mode where the CPU and other chips are running in battery-safe mode (which is of course influencing the timing in the network traffic, which can have side effects, in this case with grave effects)! You still don’t believe it? Watch the video! I am refreshing the browser (Internet Explorer 6 in this case) several times, sometimes unplugged, sometimes plugged in.


[youtube]HuR8XK393cc[/youtube]

Although this bug appears to be funny, I can reassure you that we are taking it very serious and that we are working on a solution to fix it.

Categories: Uncategorized Tags:

Upcoming feature: Custom Fortmaps

November 4th, 2009 23 comments

1.23 is going to introduce custom fort battle maps – as stated earlier. This post is going to explain this a bit more detailed what is going to be changed about the fortbattle maps.

Every fort is pretty individual because of its size and building levels. Well, of course unless the levels are maxed out, but that goes without saying. But the building levels are also not fixed for the forts – once we have custom maps, we’ll introduce destructable building structures in a following release, but that is not the topic for now. What matters is, that we have all these forts with different building levels and the battle map should reflect this fact.

So what’s going to be changed to make this happen?

  • Every building is present on the map, and the better the level is, the more advantages will be given to the sectors that belong to the building.
  • Character class sector bonus: Every tower will belong to a certain character class. Characters with the right character class that are located at such a building are going to have an additional bonus next to the sector bonus that the building structure is already providing.
  • The player limits are going to be lowered for the small and medium sized forts. There’s still a discussion going on about what the right values should be, but the initial version is most likely to set the limit to 30 players per side for the small and 60 players per side for the medium sized forts. Please note that there are unavoidable technical limitations that are limiting the number of players above the limit of 128 players, so please don’t even try to suggest to raise the limits instead of lowering them. Also, statistics of the battles show that most battles are played with about 60-70 players in average. So the limitation is mostly only hitting the small forts. The intention is to strengthen smaller alliances in the game, as the maximum limit does not involve so many players any more. Large battles can of course still happen at large forts, so the impact should be acceptable for the game.

And this is how it could look like (the map is done by me for testing purpose, please don’t start arguing about the bad sector layout):

devshot-48-map-2This is the map in a basic configuration

devshot-48-map-1This map is the same as before, except that three towers are now enlarged.

This introduces of course quite lot’s of changes to the tactics of the game play: The forts are most likely going to be not completely symmetric, then the topology might give an advantage to certain movements on the map. The smaller fort battles are going to be more tactical because of the lower player limits. All in all, battles are going to be, like the forts themselves, more individual and more unique. The motivation is, that we hope to see more battles in effect of these changes.

Categories: Uncategorized Tags: