paweld/lazarus
1
0
Fork 0
mirror of https://gitlab.com/freepascal.org/lazarus/lazarus.git synced 2025-09-19 08:59:42 +02:00
Code Issues Packages Projects Releases Wiki Activity

LazUtf8: implement custom escape strings for Utf8EscapeStrings.

Browse Source 
This also allows to remove unwanted control characters from the output.
Remove unneeded exception (if a user typecasts something else into the EscapeMode param, (s)he's on is own).
MaxGrowFactors to Cardinal (disallow them to be negative)
This commit is contained in:
Bart 2025-07-13 14:46:29 +02:00
parent 6495e72c2c
commit 8e168bc2f5
1 changed files with 31 additions and 11 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

40
components/lazutils/lazutf8.pas
View File

@ -168,10 +168,15 @@ type
emC, // C-style character encoding: "\n" (if present, otherwise hex) emC, // C-style character encoding: "\n" (if present, otherwise hex)
emHexC, // C-style character encoding: "\0x0D" emHexC, // C-style character encoding: "\0x0D"
emAsciiControlNames, // abbreviation of ASCII control characters in square brackets: "[CR]" emAsciiControlNames, // abbreviation of ASCII control characters in square brackets: "[CR]"
emPercent // percent encoding (control characters only): "%0D" emPercent, // percent encoding (control characters only): "%0D"
emCustom // user defined scheme
); );
TControlChars = #0..#31;
TEscapedStrings = array[TControlChars] of string;
PEscapedStrings = ^TEscapedStrings;
function Utf8EscapeControlChars(S: String; EscapeMode: TEscapeMode = emPascal): String;
function Utf8EscapeControlChars(S: String; EscapeMode: TEscapeMode = emPascal; CustomStrings: PEscapedStrings = nil; AMaxGrowFactor: Cardinal = 0): String;
type type
TUTF8TrimFlag = ( TUTF8TrimFlag = (
@ -3023,12 +3028,13 @@ end;
Mainly used for logging purposes. Mainly used for logging purposes.
Parameters: Parameters:
S : Input string. Must be UTF8 encoded. S : Input string. Must be UTF8 encoded.
EscapeMode: controls the human readable format for escape characters. EscapeMode : controls the human readable format for escape characters.
CustomStrings : only applicable if EscapeMode=emCustom: pointer to a fixed array of strings
AMaxGrowFactor: max growfactor for EscapeMode=emCustom.
If non-zero, no error checking is done on it's value
If set to zero, it will be calculated on the fly (which is time consuming)
} }
function Utf8EscapeControlChars(S: String; EscapeMode: TEscapeMode = emPascal): String; function Utf8EscapeControlChars(S: String; EscapeMode: TEscapeMode; CustomStrings: PEscapedStrings; AMaxGrowFactor: Cardinal): String;
type
TEscapedStrings = array[#0..#31] of string;
PEscapedStrings = ^TEscapedStrings;
const const
// lookup tables are almost 2x faster than using IntToStr or IntToHex // lookup tables are almost 2x faster than using IntToStr or IntToHex
EscapedStringsOfPascal: TEscapedStrings = ( EscapedStringsOfPascal: TEscapedStrings = (
@ -3067,7 +3073,7 @@ var
i, ResLen: Integer; i, ResLen: Integer;
SLen, SubLen: SizeInt; SLen, SubLen: SizeInt;
const const
MaxGrowFactor: array[TEscapeMode] of integer = (3, 4, 5, 5, 5, 3); MaxGrowFactor: array[TEscapeMode] of Cardinal = (3, 4, 5, 5, 5, 3, 0);
begin begin
Result := ''; Result := '';
case EscapeMode of case EscapeMode of
@ -3077,8 +3083,22 @@ begin
emHexC : EscapedStrings := @EscapedStringsOfCHex; emHexC : EscapedStrings := @EscapedStringsOfCHex;
emAsciiControlNames: EscapedStrings := @EscapedStringsOfAscii; emAsciiControlNames: EscapedStrings := @EscapedStringsOfAscii;
emPercent : EscapedStrings := @EscapedStringsOfPercent; emPercent : EscapedStrings := @EscapedStringsOfPercent;
else emCustom : begin
raise Exception.Create('Invalid EscapeMode is specified'); if not Assigned(EscapedStrings) then
Raise EInvalidOp.Create('Utf8EscapeControlChars: CustomStrings cannot be nil when EscapeMode=emCustom.');
EscapedStrings := CustomStrings;
MaxGrowFactor[emCustom] := AMaxGrowFactor;
if (MaxGrowFactor[emCustom] = 0) then
begin
MaxGrowFactor[emCustom] := 1;
for Ch := #0 to #31 do
begin
SLen := Length(EscapedStrings^[Ch]);
if (SLen > MaxGrowFactor[emCustom]) then
MaxGrowFactor[emCustom] := SLen;
end;
end;
end;
end; end;
if FindInvalidUTF8Codepoint(PChar(S), Length(S)) >= 0 then if FindInvalidUTF8Codepoint(PChar(S), Length(S)) >= 0 then