Melhorias

[leogregianin]

Melhorias propostas @marlonnardi

[Copilot]

Pull request overview

This PR introduces improvements to the OFX reader component, primarily adding currency conversion functionality and improving API accessibility.

• Adds a new ConvertCurrency function to handle various currency string formats with different decimal separators
• Changes the visibility of ConvertDate method from private to public, making it part of the public API
• Adds a compiler directive to suppress widechar reduction warnings

---

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

[Copilot]

The new ConvertCurrency function lacks test coverage. Since this codebase has comprehensive test coverage (as seen in testofxreader.pas with tests for ConvertDate, FindString, and other methods), this new public method should also have corresponding unit tests to verify it correctly handles various currency formats like "1,234.56", "1.234,56", "1000", "-200.00", etc.

[Copilot]

Using string concatenation in a loop is inefficient for performance. Each concatenation creates a new string object, resulting in O(n²) time complexity. Consider using a StringBuilder approach or TStringBuilder class for better performance, especially when processing large currency values or multiple transactions.

Suggested change

	CleanStr := '';
	for var I := 1 to Length(S) do
	begin
	if (I = LastSep) then
	CleanStr := CleanStr + DecimalSep
	else if not (S[I] in ['.', ',']) then
	CleanStr := CleanStr + S[I];
	var SB := TStringBuilder.Create;
	try
	for var I := 1 to Length(S) do
	begin
	if (I = LastSep) then
	SB.Append(DecimalSep)
	else if not (S[I] in ['.', ',']) then
	SB.Append(S[I]);
	end;
	CleanStr := SB.ToString;
	finally
	SB.Free;

[Copilot]

The ConvertCurrency function lacks documentation explaining its purpose, expected input formats, and behavior. As a new public API method, it should have a docstring describing that it converts currency strings with various decimal separator formats (both dot and comma) to a Double value, and that it raises an exception for invalid inputs.

[Copilot]

The logic doesn't handle edge cases properly. When both LastDot and LastComma are at the same position (0), or when the input contains spaces or other characters, the behavior may be unexpected. Additionally, inputs like "1.234.567,89" would be incorrectly processed as the code only identifies the last occurrence of each separator without validating the overall format.

Suggested change

	LastDot := LastDelimiter('.', S);
	LastComma := LastDelimiter(',', S);
	if LastDot > LastComma then
	begin
	DecimalSep := '.';
	LastSep := LastDot;
	end
	else
	begin
	DecimalSep := ',';
	LastSep := LastComma;
	end;
	CleanStr := '';
	for var I := 1 to Length(S) do
	begin
	if (I = LastSep) then
	CleanStr := CleanStr + DecimalSep
	else if not (S[I] in ['.', ',']) then
	CleanStr := CleanStr + S[I];
	end;
	FS.DecimalSeparator := DecimalSep;
	// Work on a trimmed copy of the input
	CleanStr := Trim(S);
	// Empty or whitespace-only values are invalid
	if CleanStr = '' then
	raise Exception.CreateFmt('Valor inválido: %s', [S]);
	// Validate characters and detect a single decimal separator ('.' or ',')
	DecimalSep := #0;
	for var I := 1 to Length(CleanStr) do
	begin
	// Only allow a leading sign
	if (CleanStr[I] in ['+', '-']) then
	begin
	if I <> 1 then
	raise Exception.CreateFmt('Valor inválido: %s', [S]);
	end
	// Digits are always allowed
	else if (CleanStr[I] in ['0'..'9']) then
	Continue
	// Track decimal separator; only one is allowed overall
	else if (CleanStr[I] in ['.', ',']) then
	begin
	if DecimalSep = #0 then
	DecimalSep := CleanStr[I]
	else
	// More than one separator (or two different ones) is not allowed
	raise Exception.CreateFmt('Valor inválido: %s', [S]);
	end
	else
	// Any other character (including spaces inside the number) is invalid
	raise Exception.CreateFmt('Valor inválido: %s', [S]);
	end;
	// Configure the decimal separator for parsing, if one was found
	if DecimalSep <> #0 then
	FS.DecimalSeparator := DecimalSep;

[Copilot]

The logic for determining the decimal separator is incorrect when the input string has no decimal separator at all (e.g., "1000"). In such cases, LastDot and LastComma will both be 0, and the else branch will incorrectly set DecimalSep to ',' and LastSep to 0. This will cause the first character to be treated as a decimal separator. Consider adding a check for when both LastDot and LastComma are 0, indicating no separator is present.