MCPE v0.16.1

The commands issue seems to be incomplete packets sent, where the packet that has the commands which will be displayed when typing /...
I'm not sure.

 

I'm sure someone from the PM team will/has already fix this issue

 

Some reflections... ...(of a man who has been where the PM team now is)...

Information is good.
Complaints are annoying.
So, would me reporting on the issue tracker be useful information, or be seen as an annoying complaint?

It certainly isn't meant as a complaint; after all, I deliberately loaded the most cutting edge alpha-not-released-yet version I could find, so I knew to expect crashes.

Over all, I think that it is information. It only happens after the system has been running for more than an hour, and in a development cycle, software loads under test rarely are allowed to go that long before rebooting. Perhaps the development team has not yet seen this behavior of Build 73.

 

Ok, now I have a GitHub account.

But...

A glance into the Jenkins shows that the PM team is now up to build 76, "successful".

I shall download 76 and see how it behaves before reporting anything formally.

 

So far so good.

I shall let 76 run for an hour or two and then I will know.

 

After four hours, Build 76 did the same thing. So now it is Issue #104, "PHP", "Unconfirmed".

 

I think in most cases with PM, as long as your bug reports are valid, the PMMP team is very responsive and understanding. The main annoying thing on the Github issues page is that people will post bugs that are their own mistake, or aren't valid. As long as your bugs are supported by good information, you should have your bug fixed soon without problem.

For the problem with crashing, I used The Auto Server Restart plugin (ASR) so that my server restarted every 2 hours to prevent crashing.

 

There is a github issue on a github issues repo that has a poll (started 2014) to be able to delete issues or not.

 

@SOFe, I see at GitHub (PMMP Issue #104) that @dktapps has reproduced the problem.
I agree with you that it is difficult to reproduce. The time needed, and the lack of visible evidence of the moment of break, is quite frustrating.

@Magicode, ASR sounds like a very interesting and useful workaround.

While others pursue other software issues which are quicker to evidence themselves, I am content to be the test technician on this issue, following suggestions and reporting results until the root problem is found and resolved.

 

The root problem is stdin breaking after some time. You can fix that by trying to detect when it breaks, and reopening it then.

 

Here is the link to ASR, you can just set it to restart before you notice the bug starting to occur.
https://forums.pocketmine.net/plugins/automatic-server-restarter.903/

 

what? that is like an awful idea
And, in addition, my experience is that stdin can break as fast as 10 minutes

Also, you can fix this with reopening stdin

 

Actually it isn't a bad idea. Restarting the server every now and then prevents memory leaks (chunks sometimes get reset) and clears the lag of the server.

 

Mathew 24:43

 

I find it's always best to give quotes from very old books in context; they are so easily misinterpreted.

As for the problem you are having, do you have an expected stable uptime for a PM server, and is it longer than the plugins you'll be using, and Windows?! Seriously though, I think you've done enough to help staff with this and can safely leave the problem with them now while you discover everything PM has to offer.

 

It doesn't prevent memory leak and should not clear lag out of the server. Please define 'lag'.

 

I have heard of many people (and have been able to reproduce it myself) that several chunks in the server get reset to an earlier time when having the server online for a very long time without restarts, being some kind of memory leak. Restarting usually does in fact clear lag, as it clears things such as entities, it unloads loaded worlds/chunks and more.

 

It does, but it should not. I never heard from chunks miraculously resetting back to earlier state. I doubt that happening. And that would not indicate a memory leak, it would indicate a huge mistake in the chunk management. Restarting just removes caches and entities, both things can be done during runtime. 'Turn the machine off and on again' is the worst practise you can ever find, it's just because someone made mess. I know it helps. But it shouldn't. Please do not see restarting as a normal thing to do. Restarting should only be needed for a software update. Auto-Restarting can be used as a temporary fix, but cannot be considered as a solution.