Errors, bugs, questions - page 1572

[A100]

MT4/950/32. Loss of digits when changing profile

When I change the profile using the toolbar icon, I immediately lose digits on the price scale (picture on the left). Then when you change the chart by selecting another tab - the digits are restored (picture on the right). Windows 8.1/32. Resolution 1024x768, I also tried 1280x1024. The scale is 125%. One digit was lost on 4 characters, two on five.

It must have started right after I increased font size in MT4 to MT5

[A100]

CHART_SHIFT_SIZE does not work with trend

void OnStart()
{
        ::ChartSetInteger( 0, CHART_SHIFT, true );
        for ( int i = 50; i >= 10; i-- )
        {
                ::ChartSetDouble( 0, CHART_SHIFT_SIZE, (double)i );
                ::ChartRedraw();
                ::Sleep( 100 );
        }
}

Dynamics as in test391.ex5 was expected

[Vladimir Karputov]

Can't download my files from the Warehouse via the MT4 editor - I get an error

2016.05.05 15:11:05.427 Storage failed to read http data (storage.mql5.com:443 read failed [12152])

[Vladimir Karputov]

Karputov Vladimir:

Can't download my files from the Warehousevia the MT4 editor- I get an error

And the error is only on one editor. On the same computer in another folder MT4 and its editor easily downloaded from the store the codes.

[coderex]

Dear developers, please introduce a namespace like in C languages.

[Alexey Navoykov]

How long will it go on that every time we update the build, the code stops compiling! And if it does compile, it doesn't work the same way (which is even worse). Who needs such a programming language?

I admire A100's patience while scrupulously scrutinizing those bugs, I'm so disgusted.

Someone suggested above that A100 should build tests to verify the compiler, but it's funny how it's the users who have to deal with this problem, not the compiler developers.

Most importantly, all this is a monkey's work. To spend years of hard work by a team of programmers (and therefore a lot of money) and years of hard work by users who must rewrite their code many times, and all this for what?To reinvent the wheel called "the C++ compiler" (with slight modifications) instead of simply using some open source compiler (or even buying one) and modifying it to suit one's needs in a couple of months.

But no, simple ways are not for us... It's much more important to proudly boast that we have good moustache ourselves, and with each new build we are able to recreate our bicycle a bit by bit.

And speaking of concrete things, I fully support A100's idea concerning the possibility to turn off the optimization, for instance, to make Debug and Release modes like in many real compilers.

Personally, because of this optimization you are praising, I'm still sticking with build 1159 because my projects compile in 2 seconds with it and compile in 20 seconds with later builds. A slight performance boost doesn't solve anything. Most of my time is spent on development and program editing.

[[Deleted]]

Alexey Navoykov:

Personally, I'm still on build 1159 because of this optimization you're praising, because my projects compile in 2 seconds there and compile in 20 seconds on later builds. A small performance boost doesn't solve anything for me. Most of the time is spent on development and program editing.

A project with 100Kb of source code compiles in less than a second in 1325 builds. Solid OOP, a lot of virtual functions and overloads, templates, pointers, const modifier (wherever possible). No DLL and OpenCL.

I would like to find out the reason for your lags. Maybe it's const that helps the compiler to optimize quickly. I have never encountered lags. Please provide me the source code from kodobase which is slowing down.

About your own bicycle in the form of a compiler. Taking someone else's project for revision has its pros and cons. I think, having weighed all the pros and cons I would have leaned towards my own bike initially. Of course, when such a decision was made, no one thought there would be such an ambush on the timing and capabilities of the language/compiler. Some overestimation of their powers, or perhaps underestimation of the complexity of the task. Of course, a lot of money was invested in developing the bike.

[Renat Fatkhullin]

Anton Zverev:

I would like to find out the reason for your lags. Perhaps it is const that helps the compiler to optimize quickly. I've never encountered lags. Please, please, give me a source code from kodobase which slows down.

It most likely has gigantic functions in the form of spools of text.

The optimizer has to make many passes at such code fragments and improve it over and over again. It's enough to reduce the size of functions for the optimizer to speed up dramatically.

Well, you should definitely switch to the latest builds since we are constantly improving both quality and speed in them.

[[Deleted]]

Renat Fatkhullin:

It is likely to have giant functions in the form of spools of text.

An optimizer has to make many passes on such pieces, improving the code time after time. It's enough to reduce the size of functions for the optimizer to speed up dramatically.

Well, you should definitely switch to the latest builds since we are constantly improving both quality and speed in them.

It's probably also a matter of scraps of text. I, at least, don't have them.

I once heard from winners of international programming olympiads that functions should be a maximum of 20 lines (conditional). If more, then architecturally/algorithmically not optimal.

When looking through Roman Yelizarov's sources, there are huge amount of simple functions with wild nesting. And almost all of them are up to five lines long. I myself am a caterpillar compared to this intellectual clump.... That's why it is not so cool, no matter how hard I tried in my time.

[A100]

When hovering the cursor over overlapping objects, the description of the background object is displayed instead of the top object. This is pronounced on OBJ_EVENT objects. I see red, but the description is from blue.