AutoIt v3.3.15.2 Beta

[Jon]

AutoIt v3.3.15.2 Beta

View File

---

3.3.15.2 (13th May, 2020) (Beta)
AutoIt:
- Fixed: Map elements being lost due to a deep-copy logic fail.

---

• Submitter
  Jon
• Submitted
  05/13/2020
• Category
  Beta

 

[Iczer]

It seems UDF-s has been reworked, especially _GUI* - control functions not use _WinAPI_InProcess() any more... Is there a no functionality loss in regard interaction with non-autoit GUI-s?

[jpm]

it is stil luse by an intrenal UDF no change in some case fixing 

[wyf]

Added UDF _GUICtrlHeader_GetFilterText () did not enter the help document.

[mdanielm]

Hello,
There is always the error:
   -1 ^ 2 = 1
In all algebraic systems the result is -1.
Cordially.

[Nine]

"Where two operators have the same precedence the expression is evaluated left to right.".  Tested with Excel (same approach) = same result, ie +1.

[mdanielm]

Yes, but who said that - and ^ have the same priority?
For me, the operator ^ has higher priority than the unary operator -.
With AutoIt:
3-1^2 = 4 !!!!!!!!!!

[mdanielm]

With Excel, the result is 2.

[Melba23]

mdanielm,

While appreciating that the results do not appear to be what you would expect - especially in your second example - I would argue that you make a good argument for the use of parentheses to explicitly force the operators to be used in the order you require. 

M23

P.S. And you do realise that the latest version is actually 3.3.15.3? 

Edited June 28, 2020 by Melba23

[BrewManNH]

On 6/27/2020 at 5:00 PM, mdanielm said:

 -1 ^ 2 = 1

In this example, you're evaluating -1 * -1, that's what the exponent is telling you to do. Two negative numbers multiplied together always yields a positive number. I'm not sure what the precedence is in algebra because I did terribly in it, but I do know that the exponent is just telling you how many times to multiply a number by itself. If you had used -1^3, the answer would be negative.

Quote

Negative numbers with exponents
If the base is negative and the exponent is an even number, the final product will always be a positive number. If the base is negative and the exponent is an odd number, the final product will always be a negative number.

 

[jchd]

I agree with @mdanielm about the existence of a real bug, examplified with his expression 3-1^2.

Look at the expression 1 - 1 ^ 2 which yields 2 under AutoIt.
If you expect it to be evaluated as 1 - (1 ^ 2) the result should be 0
If you expect it to be evaluated as (1 - 1) ^ 2 the result should also be 0

The only possible nonsensical evaluation path that can give 2 is 1 + ((-1) ^ 2) but that means a ghost addition has magically appeared out of thin air inside the guts of the core.

 

[jpm]

On 6/28/2020 at 8:25 AM, mdanielm said:

Yes, but who said that - and ^ have the same priority?
For me, the operator ^ has higher priority than the unary operator -.
With AutoIt:
3-1^2 = 4 !!!!!!!!!!

in this case - is not unary but a substraction. The  ^ operator as a same  priority as - operator so the 4 is logical !!! different from Excel.

Just use the () to avoid such behavior

[jchd]

Sorry @jpm wrong explanation!

My previous post shows that the interpreter sees thing differently and does 3 + ((-1) ^ 2) just as it interprets 1-1^2 as 1 + ((-1) ^ 2)
If the interpreter would interpret the red minus in 1-1^2 as a substraction, it would yield 0, not 2, whatever grouping or precedence it uses.

[argumentum]

...interesting theme y'all have. I never got/understood clearly, all these arithmetic stuff, so I google.
When I enter "(-1^2)" in google calculator,  it reorganize it to "-(1^2)". link
Same with "3-1^2"  it reorganize it to "3-(1^2)". link
"1-1^2" to "1-(1^2)"

Did I clarify the subject with a irrefutable solution ?. Nope, nowhere near.
We as peers, should look for peer reviewed info. to fall back to, in case of disagreement. That way the matter at hand can be based in a cited previous work and errors in opinion or the presentation of an idea, not get personal. ( I'm efficient that way ) 

[Nine]

Help file is our friend and it clearly states that :

 

Precedence

When more than one operator is used in an expression the order in which things happen is controlled by operator precedence. The precedence used in AutoIt is given below. Where two operators have the same precedence the expression is evaluated left to right.

From highest precedence to lowest:

Not
^
* /
+ -
&
< > <= >= = <> ==
And Or

Unary operators are not mentioned, so I thought it would be at the highest level.  But with the above examples ppl gave, I tend to believe something is wrong with precedence of ^.

[jchd]

Indeed precedence of unary minus is not mentionned, but it should be.
Anyway that still doesn't explain by which perverted magic the interpretor invents a wrong & ghost + sign in the middle of the operations in 1 - 1 ^ 2.

If ^ precedence is higher than both unary minus and substraction, we get 1 - (1 ^ 2) which is 1 - 1 = 0
If ^ precedence is lower than both unary minus and substraction, we get (1 - 1) ^ 2 which is 0 ^ 2 = 0

The problem is that in this case the interpreter illogically considers the minus sign as a unary minus and then, to correct its mistake, invents a + sign to convert the now invalid syntax 1 (-1) ^2 into 1 + ((-1) ^ 2) = 1 + (1) = 2

 

[mdanielm]

Hello,

I suggest, as a preprocessing of an algebraic expression, the replacement of the - unary by  _
with an instruction like:

              $in = StringRegExpReplace ($in, "(^|(\*|\/|^|\())-", "$1_")

Then just set the priority of _ less than ^

- greater or equal than * / and _2*3 will be calculated as (_2) * 3

or
- less than * / and _2*3 will be calculated as _ (2 * 3)

In both cases the priority of _ must remain greater than + -

Edited July 24, 2020 by mdanielm

[jchd]

Operators' precedence and associativity must be formally defined and documented for all operators and strictly applied to any expression, not only algebraic.

[mdanielm]

I want to give _ _ 2 the meaning _ (_ 2) so:
associativity = RIGHT