Errors, bugs, questions - page 1987

 
fxsaber:

A lot of entries. See the log file.

"Lots of entries" is a strong word... Most of the entries just aren't there. But the question is, what exactly could be preventing the output to the log file?

For example, an EA that's not very sophisticated:

void OnTick ()
{
  Print (TimeCurrent());
}

In "Real ticks" mode it shows these logs:

2017.09.05 09:38:56.994 Tester EURUSD: history data begins from 2014.07.29 00:00
2017.09.05 09:38:56.996 Tester EURUSD: preliminary downloading of history ticks started, it may take quite a long time
2017.09.05 09:38:58.001 Tester EURUSD: preliminary downloading of history ticks completed, 35.52 Kb in 0:01.030 (34.48 Kb/sec)
2017.09.05 09:38:58.001 Tester EURUSD: ticks data begins from 2017.06.01 00:00
2017.09.05 09:38:58.002 Core 1 agent process started
2017.09.05 09:38:58.346 Core 1 connecting to 127.0.0.1:3000
2017.09.05 09:38:58.347 Core 1 connected
2017.09.05 09:38:58.350 Core 1 authorized (agent build 1653)
2017.09.05 09:38:58.352 Tester EURUSD,M15 (MetaQuotes-Demo): testing of experts\Test.ex5 from 2017.08.30 00:00 to 2017.08.31 00:00
2017.09.05 09:38:58.404 Core 1 common synchronization completed
2017.09.05 09:38:58.455 Core 1 EURUSD: ticks synchronized already [43 bytes]
2017.09.05 09:39:04.609 Core 1 2017.08.30 00:36:55 2017.08.30 00:36:55
2017.09.05 09:39:10.760 Core 1 2017.08.30 00:48:27 2017.08.30 00:48:27
2017.09.05 09:39:16.866 Core 1 2017.08.30 01:02:20 2017.08.30 01:02:20
2017.09.05 09:39:22.976 Core 1 2017.08.30 01:24:08 2017.08.30 01:24:08
2017.09.05 09:39:27.668 Core 1 disconnected
2017.09.05 09:39:27.668 Core 1 connection closed
2017.09.05 09:39:27.681 Tester stopped by user

With that said, the tester is terribly slow, as you can see skips in the tens of minutes...

Machine on VPS.

 
Andrey Dik:

"Lots of records" is a very strong word... Most of the entries are simply not there. But the question is, what exactly could be preventing the logging?

Look at the log file.

For example, an EA that's not very sophisticated:

in "Real ticks" mode outputs such logs:

2017.09.05 09:38:56.994 Tester EURUSD: history data begins from 2014.07.29 00:00
2017.09.05 09:38:56.996 Tester EURUSD: preliminary downloading of history ticks started, it may take quite a long time
2017.09.05 09:38:58.001 Tester EURUSD: preliminary downloading of history ticks completed, 35.52 Kb in 0:01.030 (34.48 Kb/sec)
2017.09.05 09:38:58.001 Tester EURUSD: ticks data begins on 2017.06.01 00:00
2017.09.05 09:38:58.002 Core 1 agent process started
2017.09.05 09:38:58.346 Core 1 connecting to 127.0.0.1:3000
2017.09.05 09:38:58.347 Core 1 connected
2017.09.05 09:38:58.350 Core 1 authorized (agent build 1653)
2017.09.05 09:38:58.352 Tester EURUSD,M15 (MetaQuotes-Demo): testing of experts\Test.ex5 from 2017.08.30 00:00 to 2017.08.31 00:00
2017.09.05 09:38:58.404 Core 1 common synchronization completed
2017.09.05 09:38:58.455 Core 1 EURUSD: ticks synchronized already [43 bytes]
2017.09.05 09:39:04.609 Core 1 2017.08.30 00:36:55 2017.08.30 00:36:55
2017.09.05 09:39:10.760 Core 1 2017.08.30 00:48:27 2017.08.30 00:48:27
2017.09.05 09:39:16.866 Core 1 2017.08.30 01:02:20 2017.08.30 01:02:20
2017.09.05 09:39:22.976 Core 1 2017.08.30 01:24:08 2017.08.30 01:24:08
2017.09.05 09:39:27.668 Core 1 disconnected
2017.09.05 09:39:27.668 Core 1 connection closed
2017.09.05 09:39:27.681 Tester stopped by user

but the tester is slowing down terribly, as you can see skips in the tens of minutes...

Well so many thousands of log entries, that's why it's slowing down.

 
Andrey Dik:

"Lots of records" is a very strong word... Most of the entries are simply not there. But the question is, what exactly could be hindering the logging?

For example, an EA that's not very complex:

Real ticks" mode displays these logs:


but the tester is terribly slow, as you can see skips in the tens of minutes...

Machine on VPS.

The skipping consists of a huge number of records, the log cuts half of them when they are not needed, which slows down the Expert Advisor

 
fxsaber:

See the log file.

So many thousands of log entries, that's why it's slow.

Vitaly Muzichenko:

The skips are from the huge number of entries, the log chops up half of them unnecessarily, which also makes it slow.

Have you both tried what you are talking about?

On the other machine the log print is output on every tick as it should be and there are no lags, but on this particular machine not all logs are output, with big skips and lags.

The tester's amateurishness just amazes me sometimes, can't you make custom log output (errors, print, etc.) according to user's choice? If there is a problem with outputting full logs and tester cuts them off then why he doesn't say about it, like "Listen, I can't output all logs, only in several times, the reason ......".

So, what can be the reason for incomplete output to the log with skipping (not with cutting, namely with skipping)?

 
Andrey Dik:

Have you both tried what you are talking about?

On the other machine the log print is printed on every tick as it should be and there are no brakes, but on this particular one not all.

I didn't just try it, I started looking for a bug in the program that wasn't there. It turned out that it was just that the logs were being printed a lot and that's why there were skips.

 
Vitaly Muzichenko:

I didn't just try it, I started looking for a bug in the program that wasn't there. It turned out that it was just that the logs were being printed a lot, and so there were omissions.

No, Vitaly, in the previous post I finished. On the other machine the logs are printed without skips on every tick as they should be.
 
Andrey Dik:
No, Vitaly, I finished it in the previous post. On the other machine the logs are output without skips on every tick as they should be.

Well, maybe the operating system is different, I have win7 x64

 
Vitaly Muzichenko:

Well maybe the operating system is different, I have win7 x64

On the problem machine Server 2008, there is free memory on disk and in RAM.

On Win 10 64 and Win7 64 the logs are displayed completely, without any skips. Try to run EA only for one day in "ticks" or "real ticks" mode, if there are skips it means you have the same problem as on my problematic machine. This is a specific bug in the tester, if it is not able to output full logs then it must inform the user about this and the reason, logs have no meaning if they are not full. Who needs logs with gaps at all, how to do debugging of the program?

 
Andrey Dik:

Have you both tried what you're talking about?

On the other machine print is logged on every tick as it should be and there are no brakes, but on this particular machine not all logs are logged, with big skips and brakes.

The tester's amateurishness just amazes me sometimes, can't you make custom log output (errors, print, etc.) according to user's choice? If there is a problem with outputting full logs and the tester cuts them off, why doesn't it say something like "Listen, I can't output all the logs, only in several times, the reason ......"?

So what could be the reason for the incomplete log output (not trimming, but skipping)?

Apart from the tester logs, there are also the tester agent logs. It's all there for sure
 
Slava:
Besides the tester logs, there are also the tester agent logs. You can find everything there.


That's right, there are, here they are:

2017.09.05
2017.09.05 11:42:04:04 Logger log was cleaned
2017.09.05 2017.09.05 00:00:00.000 Server MetaTester 5 stopped

There's nothing else out there.

So why there is no full information about Print() function in the tester logs?

Reason: