Bug in StringReplace() already solved?

[funkey]

Hi,

 

I use v3.3.14.5 here and have a bug.

Local $sBug = StringReplace("123456789", 2, "a")
ConsoleWrite($sBug & " - replacements: " & @extended & " <-- expected 1" & @CRLF)

This is the result

1a3456789 - replacements: 0 <-- expected 1

 

Is this already solved in v3.3.16.1

I don't want to update now...

 

Thanks for your answers!

[Andreik]

Not yet.

PS: has this been reported?

Edited December 29, 2023 by Andreik

[funkey]

Don't know. Just found it.

Thanks for testing.

[Andreik]

I'm not sure if this is by design or a bug. Since the function it's named StringReplace() the search string should be a string not an integer. If it's a string it works fine. Yet since AutoIt uses a variant implementation to represent variables and does all kind of conversions probably should deal internally and make sure the parameters are converted to strings.

The interesting part it's that in this code only the first 2 it's replaced with an a:

Local $sBug = StringReplace("1234567892", 2, "a")
ConsoleWrite($sBug & " - replacements: " & @extended & " <-- expected 1" & @CRLF)

and the result

Quote

1a34567892 - replacements: 0 <-- expected 2

I would like to know more about what AutoIt does internally but for now even if it's not enforced I would say it's up to programmer to be sure that the search string it's really a string.

Edited December 29, 2023 by Andreik

[funkey]

I wanted it to be an integer for replacing characters at position. This is the second way StringReplace() works.

[Nine]

Why do you need @extended to report 1 when you ask to change 1 char (or string) with another based on position ?  Check @error instead !

[funkey]

I don't need it. But the behaviour is not consistently. I just expected there to be a '1' in @extended. 

[Nine]

Well, since it is useless, I don't think any dev in here will put 1ms to please this eccentricity...

[argumentum]

14 minutes ago, Nine said:

Well, since it is useless,

based on https://www.autoitscript.com/autoit3/docs/functions/StringReplace.htm we either fix the code or fix the help.

[Nine]

I agree to fix help as @error is not discussed there as it should be...

[junkew]

When you use it with value start which would be a number at a certain position it becomes vague to give input as number or as string

When you want to replace 1 char at a certain position result seems to be ok but @extended not filled with 1 (would be most logical)

;~ StringReplace ( "string", "searchstring/start", "replacestring" [, occurrence = 0 [, casesense = 0]] )

Local $sBug = StringReplace("001234567892", 2, "a")
ConsoleWrite($sBug & " - replacements: " & @extended & " <-- expected 1" & @CRLF)

;~  **** ambiguous ***
;~ unclear what to expect as input is string but could be interpreted as start
Local $sBug = StringReplace("001234567892", "2", "a")
ConsoleWrite($sBug & " - replacements: " & @extended & " <-- expected ?" & @CRLF)

 

[water]

I suggest to replace the syntax description

StringReplace ( "string", "searchstring/start", "replacestring" [, occurrence = 0 [, casesense = 0]] )

with

StringReplace ( "string", "searchstring"/start, "replacestring" [, occurrence = 0 [, casesense = 0]] )

This makes clear that parameter 2 is either a string to search for or an integer to be used as the starting position.

[junkew]

And it works by design🙂

If the start method is used the occurrence and casesense parameters are ignored. The function will replace the characters in "string", starting at the requested position, with the characters in "replacestring" - as many characters will be replaced as are in "replacestring". However, if there are not enough characters in "string" for the entire "replacestring" to be inserted an empty string is returned and @error is set to 1.