Amebis/wxWidgets
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

Fix wxTextInputStream incorrectly decoding multibyte fallback encodings

Browse Source
This commit is contained in:
Pavel Tyunin
2020-09-28 21:58:52 +03:00
parent 857950c626
commit 1e435d2347
1 changed files with 7 additions and 6 deletions
Download Patch File Download Diff File Expand all files Collapse all files

13
src/common/txtstrm.cpp
View File

@@ -134,12 +134,13 @@ wxChar wxTextInputStream::GetChar()
// one extra byte, the only explanation is that we were using a // one extra byte, the only explanation is that we were using a
// wxConvAuto conversion recognizing the initial BOM and that // wxConvAuto conversion recognizing the initial BOM and that
// it couldn't detect the presence or absence of BOM so far, // it couldn't detect the presence or absence of BOM so far,
// but now finally has enough data to see that there is none. // but now finally has enough data to see that there is none, or
// As we must have fallen back to Latin-1 in this case, return // it was trying to decode the data as UTF-8 sequence, but now
// just the first byte and keep the other ones for the next // recognized that it's not valid UTF-8 and switched to fallback.
// time. // We don't know how long is the first character or if it's decoded
m_validBegin = 1; // as 1 or 2 wchar_t characters, so we need to start with 1 byte again.
return wbuf[0]; inlen = -1;
break;
#if SIZEOF_WCHAR_T == 2 #if SIZEOF_WCHAR_T == 2
case 2: case 2: