Html5 Client Preview

[beaver]

I will be releasing an update later today or tomorrow that will fix one of the causes behind the disconnects.

[beaver]

I released it this morning. Please let me know if it is any better.

[beaver]

An update on the current status of 2.5:

There are 4 main problems that still exist in the html5 client:
1) Frequently new updates are broken
2) I updated the 3rd party library I use to parse server messages in the middle of January, and since then the client hasn't worked on Safari or Firefox.
3) Sometimes after leaving a game it stops being able to parse server messages until the page is reloaded. So far I haven't found a reliable way to reproduce this.
4) Disconnections. In January I released many fixes and improvements. Disconnections in the beta are automatically recorded, and none have been recorded in the last 5 days! Though there could be a bug in recording disconnections.

My focus next week will be finding the root cause behind those issues, and fixing them.
1) To fix #1, I will be adding automated tests that run before the client is updated. If the tests fail, the update will be canceled. I should have done this a long time ago.
4) If you use chrome or Microsoft Edge, please test the beta, and let me know if you are disconnected.

[GivinUDaHimelik]

4) If you use chrome or Microsoft Edge, please test the beta, and let me know if you are disconnected

I'll try to check on this when I get home today.

[GivinUDaHimelik]

I used Beta on Chrome and it worked fine. The rolls however as mentioned in Bug reports were still pretty bad. (I won 7v8 twice without a boost, held off 8v5 and 8v6).

[beaver]

The rolls however as mentioned in Bug reports were still pretty bad. (I won 7v8 twice without a boost, held off 8v5 and 8v6).

Did you notice anything unusual about the numbers that were rolled?

[GivinUDaHimelik]

The rolls however as mentioned in Bug reports were still pretty bad. (I won 7v8 twice without a boost, held off 8v5 and 8v6).

Did you notice anything unusual about the numbers that were rolled?

No, they were fairly close (within 4) each time. That was about it.

[beaver]

This afternoon I looked through the code, and I didn't find anything. That code hasn't changed in several years. We've seen battles like what you had, but the difference this time is it is happening more frequently. If I can't find the cause before Monday, I will try switching the algorithm for random numbers next week.

[BeatlesFan]

Based on my previous experience with Dragons in certain modes I stopped using them.

Could I ask that you look at the code for the following four situations and see if you find ANY difference in what you expect to happen:

Dragons vs Reptiles in HC mode

Dragons vs Reptiles in Att mode

Reptiles vs Dragons in HC mode

Reptiles vs Dragons in Att mode

If those all work out with expected probabilities then I withdraw my concern, however based on statistically unlikely results, I contend that somewhere in the code when a positive attacking race faces a positive defensive race in Att or HC you will find an error.

[beaver]

I looked through the code earlier this week, and it looks correct. You are not the first person to say that something seems wrong, so when I have time I might run some simulations.

I've been sick for most of the week, so progress has been much slower. I finished adding automated tests. If the tests fail, the new version will not be released. Currently the tests are pretty basic, but will catch a few mistakes I occasionally make when deploying an update.