In Defence of Magic Numbers

[funkey]

MB_OK = 0

[BrewManNH]

Don't forget it also includes $MB_DEFBUTTON1 = 0 and $MB_APPLMODAL = 0.

[czardas]

What's this all about?

  Reveal hidden contents

MB_ICONHAND = MB_ICONSTOP = MB_ICONERROR = 16

Edited December 13, 2013 by czardas

[Richard Robertson]

  On 12/13/2013 at 4:37 AM, czardas said:

What's this all about?

  Reveal hidden contents

MB_ICONHAND = MB_ICONSTOP = MB_ICONERROR = 16

It means that all three are documented to be the same result. It's consistent with Win32 programming.

I prefer compile-time constants that can be optimized. Calling them global variables is asking for trouble and inefficiency.

[Exit]

MsgBox("", "", "No magic numbers :-)")

[water]

LOL

[guinness]

  On 12/13/2013 at 1:25 PM, Exit said:

MsgBox("", "", "No magic numbers :-)")

Wrong! "" isn't a valid parameter that can be passed to flag.

[water]

Why? The parameter is accepted by AutoIt so it can't be wrong.

[guinness]

  On 12/13/2013 at 1:38 PM, water said:

Why? The parameter is accepted by AutoIt so it can't be wrong.

Is it documented in the remarks?

[water]

Is it accepted by AutoIt?

[BrewManNH]

As we should all know an empty string when compared to a number equals 0, so MsgBox("", "", "Text) is equivalent to MsgBox(0, "", "Text").

[Mat]

There are such thing as magic strings as well. $GUI_RUNDEFMSG for example. 

[iamtheky]

  Quote

The day I have to type MB_DUNNOWHAT while debugging, instead using a quick zero, is the day I stop coding. Publishing code is a different matter.

 

 

If a defense for variables is later maintenance of the code. Then i think the same case could be made for production use of magic numbers commonly used in debugging.

[czardas]

  On 12/13/2013 at 1:39 PM, guinness said:

Is it documented in the remarks?

;

MsgBox(Not "", "", "" = Not "documented in the remarks")

Edited December 14, 2013 by czardas

[guinness]

  On 12/14/2013 at 2:13 PM, czardas said:

 

;

MsgBox(Not "", "", "" = Not "documented in the remarks")

I don't get it. In the MsgBox() remarks for supported flag values it doesn't mention anything about a blank string, only numerical values and Default.

[czardas]

  On 12/14/2013 at 2:15 PM, guinness said:

I don't get it. In the MsgBox() remarks for supported flag values it doesn't mention anything about a blank string, only numerical values and Default.

 

But that is not a blank string.

[guinness]

  On 12/14/2013 at 2:06 PM, boththose said:

If a defense for variables is later maintenance of the code. Then i think the same case could be made for production use of magic numbers commonly used in debugging.

How would you differentiate between a numerical value and one that is a magic number?

  On 12/14/2013 at 2:17 PM, czardas said:

But that is not a blank string.

I thought "" was a blank string?

[czardas]

  On 12/14/2013 at 2:19 PM, guinness said:

How would you differentiate between a numerical value and one that is a magic number?

I thought "" was a blank string?

 

It's a joke. If you run the example, it gives the right answer.

[czardas]

  On 12/14/2013 at 2:06 PM, boththose said:

If a defense for variables is later maintenance of the code. Then i think the same case could be made for production use of magic numbers commonly used in debugging.

I'm not quite sure what you mean by this.

---

@guinness, I think it would be impossible to document all expected behaviour in every situation. Perhaps that's not so important here. The fact the interpreter is capable of doing such magic tricks makes the language extremely user friendly and flexible. It's one of the endearing features of AutoIt IMO, albeit perhaps one of a multitude of consequential side effects.

Edited December 14, 2013 by czardas

[AZJIO]

in C++ constants are converted to numbers?
#define