Beta version of MetaTrader 4 IDE including new MQL4 compiler and editor - page 8

New comment
 
TarasBY:
Your code is 'closer to the point'. There is rarely anything in the public domain that is 'not worth redoing'.

Wrong, oh wrong. There are good things, you just have to understand them
 
MetaDriver:

Vadik is bluntly angry at the inevitable competition. ;) // imha self-sabotaging.

It used to be a "panacea" with its libraries, but now they will be much less in demand - almost everything can be done without them, and much of what cannot be done can be found ready-made in kodobase at mql5.com

Vadim, look for the pros, there are orders of magnitude more.

Vladimir, what does this have to do with competition? The libraries will not be canceled. They will have to be redesigned because of incompatibilities. Not because they are in Code Base, but because I work with them.

True, we would have to change only one function in one library. But it's unpleasant, too.

This is another reason not to write anything else in MQL. I'm sick of all these changes and constant bugs!

 
Zhunko:

Vladimir, what does this have to do with competition? The libraries are not being cancelled. We'll have to redo them because of incompatibility. Not because they are in Code Base, but because I work with them.

True, we would have to change only one function in one library. But it's unpleasant too.

Geez, I was beginning to think that half of the library crashed there. :)

This is bullshit. Perhaps some real inconvenience-incompatibilities when working with DLL arise only in connection with the transition to UNICOD. But no more than when working with strings under other software [with similar transition].

And anyway, you have to switch to it sometime.

This is one more reason for me not to write anything in MQL. I'm fed up with all these changes and constant bugs!

You should not be like that. I like it. There are very few bugs now, and when they do appear they are fixed quickly. The changes, as a rule, are only pleasant.

There may be other reasons not to write TC in mql5(4), but not this one.

 
MetaDriver:

Man, I thought half the library had collapsed... :)

That's bullshit. Perhaps some real inconvenience-incompatibilities when working with DLL arise only in connection with transition to UNICOD. But no more than when working with strings in other software [with similar transition].

And anyway, you should switch to it at some point.

You shouldn't do that. I like it. There are very few bugs, and when they do pop up, they are quickly fixed. The changes, as a rule, are only pleasant.

Maybe there are other reasons not to write TS in mql5(4), but not that one.

There is no problem with unicode. I have libraries independent of the encoding. You can compile in different ways or add interfaces.

There are still plenty of bugs and more. I'm sick and tired of writing about them and don't do it anymore. It's easier to write them in a DLL in a normal language. From what I've reported nothing has been fixed. They just write that it was designed that way.

I remember one incident. When, apparently in the 406 build, the indicator buffer refresh was broken, Renat threatened to ban me, after I explained the reasoned nature of the problem (pictures and the code attached). And yet, it was fixed! After someone noticed that the undocumented ability to update the offline graphics stopped working! :-)) I.e. didn't care about the documented feature, but got worried about the undocumented ones.

Now there are bugs in some functions for arrays. They don't work correctly in libraries. It feels like they are working with the cache. The contents and size of the array have changed, and they keep working as if nothing has changed. The same code in the main module works correctly. Now it's easier to write it in DLL, than in Metacvot TP :-)) It's more reliable that way.

 
Vinin: Wrong, oh wrong. There are good things, you just have to understand them.
I didn't say "never", did I? ;)
 

I wanted to try the beta version today, but it won't start with an error:

when starting with start.bat:

stack overflow 1140 bytes in thread 0017 eip 7 bc400ca esp 00240 ebc stack 0x240000-0x241000-0x340000

When running MetaRditor.exe:

Unhandled illegal instruction at address 0x5959d0 (thread 0017), starting debugger...
Unhandled exception: illegal instruction in 32-bit code (0x005959d0).
Register dump:
 CS:0073 SS:007 b DS:007 b ES:007 b FS:0033 GS:003 b
 EIP:005959 d0 ESP:0033 fe3c EBP:0033 fe50 EFLAGS:00010206(  R- --  I   - -P- )
 EAX:0066 e6a0 EBX:00000000 ECX:00000000 EDX:00170 b58
 ESI:00670 f08 EDI:00670 f78
Stack dump:
0x0033fe3c:  00740 e78 0066 e6a5 00549487 008 b268e
0x0033fe4c:  00000001 0033 fe90 005477 dc 00000001
0x0033fe5c:  54441 f6b 008 b268e 7 ffdf000 7 b8855c0
0x0033fe6c:  ab881e04 c000001d 00000000 0033 fe5c
0x0033fe7c:  0033 f990 0033 ff10 0054 e2f0 54187 ebb
0x0033fe8c:  00000000 0033 fea8 7 b85726c 7 ffdf000
Backtrace:
=>0 0x005959d0 in metaeditor (+0x1959d0) (0x0033fe50)
  1 0x005477dc in metaeditor (+0x1477db) (0x0033fe90)
  2 0x7b85726c call_process_entry+0xb() in kernel32 (0x0033fea8)
  3 0x7b857e9f ExitProcess+0xc2e() in kernel32 (0x0033fee8)
  4 0x7bc72540 call_thread_func+0xb() in ntdll (0x0033fef8)
  5 0x7bc75010 call_thread_entry_point+0x6f() in ntdll (0x0033ffc8)
  6 0x7bc4b6fa call_dll_entry_point+0x659() in ntdll (0x0033ffe8)
0x005959d0: (bad)

Runs on:

processor : 0
vendor_id : AuthenticAMD
cpu family : 6
model : 7
model name : AMD Duron(tm) Processor
stepping : 1
cpu MHz : 1210.765
cache size : 64 KB
fdiv_bug : no
hlt_bug : no
f00f_bug : no
coma_bug : no
fpu : yes
fpu_exception : yes
cpuid level : 1
wp : yes
flags : fpu vme de pse tsc msr pae mce cx8 sep mtrr pge mca cmov pat pse36 mmx fxsr sse syscall mmxext 3dnowext 3dnow up
bogomips : 2421.53
clflush size : 32
cache_alignment : 32
address sizes : 34 bits physical, 32 bits virtual


Does the new version happen to require SSE2 like MT5?

 
ms502040:

I wanted to try the beta version today but it won't start with an error:

................

Does the new version require SSE2 like MT5?

Could be.
 
MetaDriver:

There are very few bugs now, and when they do pop up, they are quickly fixed. Changes, as a rule, are only pleasant.

There are very few bugs because MQL4 has been polished for many years without major changes, so the number of bugs is only reduced. But now it will be the other way around...

Undoubtedly, for those who coding is an end in itself (i.e. software development for money or just a hobby), modernization of the language is certainly a joy for them. And on compatibility problems and bugs occurring at that one can even make a fortune, forcing the customer to pay extra money and explaining to him that these metaquotes-robbers screwed up too much and now he is to rewrite half of the code :) Or earn money purely from rewriting codes to the new standard. All in all, they are nothing but pluses.

But for those who are interested in the final product, which they will use in trading (and subsequently improve), all these experiments with the language are unacceptable to them. The main thing here is stability, not all fancy features. And rewriting your code because of someone else's whims, and then spending years catching bugs again... no way, let's not.

So, yes, there's probably no point in arguing about it, because everyone has his own aims. I'm talking about why there are so few dissatisfied in this thread. And many do their conclusion in silence.

 
Meat:

Well, there are few bugs because MQL4 has been polished for years without any radical changes, so the number of bugs is only decreasing. But now it will be the other way around...

......... . .....

This all just should have been done a long time ago. We should have started mql5 on 4 first and started the fifth terminal after that.

// and don't whine, you'll like it in the end anyway. :))

// and they promised 99.9 percent compatibility, so there's not much you'll actually have to redo.

 
MetaDriver:

// and compatibility was promised to be 99.9 per cent, not much really has to be redone.

That's a lie.
123456789101112131415...23
New comment
Reason: