opensource/HeidiSQL
1
0
Fork 0
mirror of https://github.com/HeidiSQL/HeidiSQL.git synced 2026-03-13 09:24:25 +08:00
Code Issues Packages Projects Releases Wiki Activity

Cosmetic: Keep the 'Bytes' variable counting length without the null terminator (shuffle around some +1's).

Browse Source 
Bugfix: Delphi 2007's Utf8ToUnicode malfunctions and causes an access violation when given a 1-byte UTF8 buffer containing only the null terminator.
This commit is contained in:
rosenfield.albert
2008-06-13 13:54:10 +00:00
parent 327f4c041d
commit 505cf7b443
1 changed files with 18 additions and 6 deletions
Download Patch File Download Diff File Expand all files Collapse all files

24
components/zeosdbo/src/dbc/ZDbcMySqlStatement.pas
View File

@@ -338,11 +338,11 @@ var
begin
// UTF-8 strings are at most 4 bytes per character.
// Include room for a null terminator.
Bytes := Length(Value) * 4 + 1;
GetMem(Input, Bytes);
Bytes := Length(Value) * 4;
GetMem(Input, Bytes + 1);
// Convert WideString input to UTF-8 PChar.
Bytes := UnicodeToUtf8(Input, Bytes, PWideChar(Value), Length(Value)) - 1;
Bytes := UnicodeToUtf8(Input, Bytes + 1, PWideChar(Value), Length(Value)) - 1;
// Make room for escaped UTF-8 string.
GetMem(Buffer, Bytes * 2 + 1);
@@ -355,10 +355,22 @@ begin
Bytes := FPlainDriver.GetEscapeString(FHandle, Buffer, Input, Bytes);
// Make room for wide version of escaped string.
GetMem(Output, Length(Value) * 4 + 2);
GetMem(Output, Length(Value) * 2 * 2 + 2);
// There's a big nasty BUG in Utf8ToUnicode:
// ------------------------------------------
// If it is called with parameters: (2-byte buf), (0 chars), (1-byte buf), (1 byte),
// the function will skip a loop, then decrease an unsigned counter from 0 to 4294967295,
// then use that variable to access the 2-byte buffer way out of bounds. Workaround here:
if Length(Value) = 0 then begin
// Skip Utf8ToUnicode() to avoid bug.
Output[0] := #0;
Output[1] := #0;
end else begin
// Convert back to WideChars, then compose result.
Utf8ToUnicode(Output, Length(Value) * 2, Buffer, Bytes + 1);
end;
// Convert back to WideChars, then compose result.
Utf8ToUnicode(Output, Length(Value) * 2, Buffer, Bytes + 1);
Result := '''' + WideString(Output) + '''';
// Free temporary buffers.