Devlog update (2014-07-09) Bonus devlog

[Shoozza]

Hey again,

posting the devlog one day late, but you had a devlog less than a week ago so I hope it's ok

Anyways, we had our private meeting and it seems like a good thing.
Not everyone was there though >:[ but at least the people there gave an status update and committed to tasks (or made clear they will do nothing).

I can only tell what I did since the meeting though.
I continued to work on the virtual machine scripts and they are mostly done, I can install most of the tools only configuring them is not ready yet as that cannot be done via commandline for delphi (which is sad).
For some installers a silent install didn't work. I tried to figure out the correct parameters and lost some time there. In the end I had to use AutoHotKey and make it check for window titles and press keys :/ but it works now.
I also wrote a script for converting lazarus deb files so we can use them with ubuntu 10.04 (converts data.tar.xz into data.tar.gz for those who are interested). It still needs testing but though.

And that's it for now till next time (next Tuesday)!

[Zakumy]

I miss the developer so badly   
I checked the forum weekly in this Summer(WEEKLY, IT'S TRUE, Beside playing other PC games), and I don't see any devlogs since 1.6.7 out, It makes me a little sad though
And I thought everyone dumbed Soldat because I don't see anyone complaining about the dev log lack. The Forum was nearly dead, I felt like that.

Anyway, I'm just waiting for the "Supa Secret Project" that you guys are up to 1.7.0

 

[jrgp]

I'm unable to attend the private meetings because unfortunately all of the soldat devs except me are in europe and the timing is impossible for me. It isn't fair.

And shoozza I'm glad that trick I suggested with the debs works.

[Shoozza]

I'm unable to attend the private meetings because unfortunately all of the soldat devs except me are in europe and the timing is impossible for me. It isn't fair.

And shoozza I'm glad that trick I suggested with the debs works.

Yeah when i switched to 12.04 I thought it would be more complicated to repack debs but after you suggested it I decided to give it a try. Thanks for that.

I miss the developer so badly   
I checked the forum weekly in this Summer(WEEKLY, IT'S TRUE, Beside playing other PC games), and I don't see any devlogs since 1.6.7 out, It makes me a little sad though
And I thought everyone dumbed Soldat because I don't see anyone complaining about the dev log lack. The Forum was nearly dead, I felt like that.

Anyway, I'm just waiting for the "Supa Secret Project" that you guys are up to 1.7.0

It's true there were less devlogs (also nothing really happened)
The main reason was that we were stuck (because of the SVN->GIT migration which I didn't finish for a long time).
But now stuff should change. I plan to give some information on what we are doing without spoiling the 1.7 release though

[Zakumy]

GIT ? Well I searched on Wiki, Are you guy planing to let Soldat run on Linux and Mac ?

[Shoozza]

Git is a Version Control System. I posted more about it in the devlog before this one.
Basically it is used to keep track of the changes we make to the Soldat code and it allows us to work on Soldat at the same time.
SVN is similar but works different and we want to use some of the advantages GIT has.

Linux and Mac support is planned for a long time already, it is just a ton of work. But I really want to make it happen.
The server runs on Windows, Mac and Linux already. The client runs on Windows and with wine on Linux.

[jrgp]

The client runs on Windows and with wine on Linux.

It's worth mentioning that the client also runs perfectly under VMWare on Mac.

[nosejj]

I've to say; LAW and grenades eats have increased vastly in the last ver.

[Shoozza]

Hmm I wish we had some way to actually test this stuff (the eats and hits). That would make it easier to decide on how to change the WM.
What do you think skoskav?

[15th_account]

Umm, magic?

Even though eats are perhaps the most pervasive problem in Soldat, we still don't know enough about them to know what we can do about them. Is the main culprit packet loss, packet corruption, lag spikes, sync issues or some other weird bug?

I think what makes this type of bug such an ass is that it requires two people to test. If it was possible to - on compile - immediately toss up a server and dummy client that joins the server and starts jumping in place or something, then it might be possible to see what caused the eat.

[jrgp]

I think what makes this type of bug such an ass is that it requires two people to test. If it was possible to - on compile - immediately toss up a server and dummy client that joins the server and starts jumping in place or something, then it might be possible to see what caused the eat.

I'm sure Shoozza and I could set something up that does this with fancy commit hooks. Maybe on a branch just used for scratch testing like this.

[Bistoufly]

I think what makes this type of bug such an ass is that it requires two people to test. If it was possible to - on compile - immediately toss up a server and dummy client that joins the server and starts jumping in place or something, then it might be possible to see what caused the eat.

I'm sure Shoozza and I could set something up that does this with fancy commit hooks. Maybe on a branch just used for scratch testing like this.

Nice!

Exhunter told me basically the same thing. It's a hard and tedious task that scare the team because it requires a lot of testing.

Your plan sounds like it could make the task more manageable.

[Shoozza]

I remember that MM used some tools to simulate some scenarios like packet loss.
Does anyone know about such tools?

Ideally we would have a programmable commandline soldat client which joins a server with some simulated packet loss.
This client would then run some actions and record the packet loss and hit stats.

The only thing we can simulate now is adding sleep to the apponidle script function and make the server hang :3 (did that long time ago by accident).

I'm not sure this is suitable for post commit hooks though. As it will not finish quickly (unless someone figures out how to test it quickly).

Before we add such a thing though we probably need a post commit hook for the commit message bot

[Szaman]

afaik, iptables can simulate packet loss. Or this one - netem - http://www.linuxfoundation.org/collaborate/workgroups/networking/netem

[Shoozza]

Thanks for the link!
That would cover it serverside on linux but clientside uses windows and it seems like netem is only for linux.

Is there something similar for windows?

[Szaman]

You can enable packet loss for packets that are sending from the server to the client - that's the same as packet loss on client side.