Dot interpreted as literal string even when unquoted. Topic is solved

[Helgef]

Example,

Code: Select all

Msgbox, % .

See

---

v2.0-a081-cad307c detects these errors.

Indeed, thanks

[jNizM]

Code: Select all

MsgBox % .          ; -> .
MsgBox % .123e+1    ; -> .123e+1
MsgBox % .test      ; -> .test
Msgbox % .@#µ$3Z    ; -> .@#µ$3Z

Not sure.. maybe it is somewhere here -> https://github.com/Lexikos/AutoHotkey_L ... .cpp#L5597

[lexikos]

I'm not at all concerned that syntax errors are being interpreted as floating point literals.

Floating point literals in v1 are not converted to binary numbers until you use them (at runtime).

[Helgef]

Thank you both for your responses.
@jNizM, interesting observations.
@lexikos, imo, it would be a nice feature if syntax errors caused an error or showed a warning instead of being interpreted as floating point literals. I wish it showed something like

Code: Select all

Error:  Ambiguous or invalid use of "."

Cheers.

---

Spoiler

Floating point literals in v1 are not converted to binary numbers until you use them (at runtime).

That is a good information. Would that be the reason floatVar() is significantly faster than floatLiteral()?

Code: Select all

floatVar(x){
	static myFloat:=1.0001
	return myFloat+x
}
floatLiteral(x){
	return 1.0001+x
}

Code: Select all

Number of calls: 1000000
floatVar(), elapsed time is: 607ms.
floatLiteral(), elapsed time is: 1024ms.

[lexikos]

I have no plan to improve syntax validation in v1.

It will be detected in v2.

Note .test is, of course, valid when following a value (any kind of value), as in someObject .test (space is allowed before the dot). .@#µ$3Z raises an error in that context because of @, # and $.

[Helgef]

It will be detected in v2.

From my point of view, this topic is solved.

[jeeswg]

Is . viewed as a floating point literal (i.e. 0.0) by the syntax checker, but seen as the string . (i.e. ".") when the script runs?

Code: Select all

q::
vNum := .
if vNum is float
	MsgBox, % "y"
else
	MsgBox, % "n" ;n

vNum := .0
if vNum is float
	MsgBox, % "y" ;y
else
	MsgBox, % "n"

vNum := 0.
if vNum is float
	MsgBox, % "y" ;y
else
	MsgBox, % "n"

MsgBox, % (0 + .) ;blank
MsgBox, % (0 + .0) ;0.000000
MsgBox, % (0 + 0.) ;0.000000
MsgBox, % (0 + .1) ;0.100000
MsgBox, % (0 + 1.) ;1.000000
return

[lexikos]

Floating point literals in v1 are not converted to binary numbers until you use them (at runtime) **as numbers**, hence .123e+1 showing as is in jNizM's example. They are all just generic string values.

[lexikos]

v2.0-a081-cad307c detects these errors.