New MetaTrader 5 Platform Build 5260: Enhancements in Algo Forge, extended OpenBLAS support, and new inheritance rules in MQL5 - page 3

New comment
 

Forum on trading, automated trading systems, and simulation of trading strategies.

Calendar events older than 1 week do not load their current values.

Nickel Mendoza , 2025.09.07 17:09

Hello everyone,

I'm working with a news EA that requires historical economic events to simulate them in the strategy tester . During testing, I noticed some strange results, such as the appearance of the LONG_MIN value.

This reminded me that members of the MqlCalendarValue structure, such as actual_value and others, are set to LONG_MIN when that value isn't defined. For this reason, I checked the CSV file from which the bot reads events and confirmed that some records did indeed contain that value.

When looking directly at the calendar, I noticed that certain events appear marked as ND (Not Defined) , for example: " dallas-fed-services-revenues" from the USA . However, when checking other sources, I found that this event does have an updated value.

Everything indicates that this is a recent error.

Does anyone know if this problem can be solved in the future?


 
Miguel Angel Vico Alba #:
There is a clear difference between constructive criticism and just throwing negativity.

Pointing out a real issue with explanations and proposals is valuable and helps improve the product.

Using sarcasm without offering alternatives only adds noise, derails the bug report, and looks like pure hostility. That is not feedback, it is just trash in the discussion.

The worst part is that MQ still allows them to be moderators. They log in once a week or even once a month and use any bug report or suggestion to throw trash. What do we call that? Pure toxicity.



Referring to me? - Hmm.

Toxicity goes both ways. Always sugarcoating whatever savage we users (and moderators) are being exposed to by failed implementation is as toxic as not providing details and just throwing out any context missing messages. - I agree on that.

Within my daily work with ME5, one base rule stands out to me: "Never trust the debugger". Let that sink in.

I am confident, I have reported in great detail quite a number of bugs, have invested a lot of work to make the documentation better, and none of it got ever actually fixed, with a few exceptions.

Beyond that, I would say, I do have a good understanding of the limitations within MQL5, and I have no issue admitting when I am wrong. Very contrary to you as can be seen in quite some discussions. - Just think about: " dont sell us this as a feature".

I have no incentive to fight your hostilities against any if your derails while trying to shut someone up, I just walk away, as you seem to have no need to better yourself, and therefore I have no engagement in trying to support your efforts, as it is either your way or highway. And that is a much worse toxic behavior, because it damaged s actually everyone in the long term, company and consumers.

Anyways, have a good day and please keep supporting all the users on the forum as you do.
 
Rashid Umarov #:
We also have moderators who are company haters. This is an interesting curiosity.

Hate and pointing out are two different things, like cloning and forking. 

Are you a microsoft lover? - With disappearing SSds, they dont admit to? - Putting out a patch that fixes the problem, without admitting to be at fault with their previous patch?
Should we users just accept the fault, not even reporting the issue?

The effort of users and moderators to point out a flaw is actually far away from hate, but a desperate try to point out fields of potential improvment. 


So your comment actually shows the problem we users and moderators are fighting, and why MQ is having the issues, it has....  

But thank you for pointing this out.

 
Miguel Angel Vico Alba #:

I don't have exactly this issue but I have a lot of problems with classes and structs

I not a proffesional but previously this was working 

    CInterval Intvls[3];

    Intvls[2]   = CInterval(smbl,PERIOD_M3);
    Intvls[1]   = CInterval(smbl,PERIOD_M2);
    Intvls[0]   = CInterval(smbl,PERIOD_M1); 

    CInterval intvl_1m = Intvls[0];

Now i had to change it to :

CInterval* intvl_1m = &Intvls[0];

delete intvl_1m;

and thats fine but thw whole day I'm trying to figure out what else causes memeory leak, where it was ok before the update;

I suspect structs they stay somewhere in the memory but I don't have idea how to check it;

I think I will wait until new update;

 
Ryan L Johnson #:
Drag your vertical price scale down so far that your candles become a flat line before you enable Scale fix, and make sure Chart auto scroll is disabled.
No, I already tried doing that, but it always automatically returns to the previous zoom level. MT4 handles this very well, but MT5 handles it very poorly. Just try it on both MT4 and MT5 and you’ll understand.
 
This update was very useful for me. I could not use all of my CPU Thread for strategy tester. Now its fixed and the performace with 265K is very good.
 
Update: All compiler-related issues mentioned in this topic up to the time of this post will be resolved in the next beta release.
 
Luu Tuan Trung #:
No, I already tried doing that, but it always automatically returns to the previous zoom level. MT4 handles this very well, but MT5 handles it very poorly. Just try it on both MT4 and MT5 and you’ll understand.

This actually is how I do it in MT5. Unfortunately, I can't reproduce your problem in order to help you.

 
Miguel Angel Vico Alba #:

The calendar has never worked in Tester. Use third-party solutions to enable working with the calendar in Tester.

 
Przemysław Piotrowiak #:

I suspect structs they stay somewhere in the memory but I don't have idea how to check it;

Need source code to reproduce. So that you can see different behavior in different builds.

1234
New comment