fankes/termux-app
2
0
Fork 0
mirror of https://github.com/fankes/termux-app.git synced 2025-09-06 02:35:19 +08:00
Code Issues Actions Packages Projects Releases Wiki Activity

Replace surrogate chars with U+FFFD

Browse Source 
Also add some more unicode input tests.
This commit is contained in:
Fredrik Fornwall
2016-01-19 23:11:20 +01:00
parent 38dd99e827
commit 54857d5fd4
2 changed files with 47 additions and 2 deletions
Download Patch File Download Diff File Expand all files Collapse all files

8
app/src/main/java/com/termux/terminal/TerminalEmulator.java
View File

@@ -218,7 +218,7 @@ public final class TerminalEmulator {
*/
private int mScrollCounter = 0;
private int mUtf8ToFollow, mUtf8Index;
private byte mUtf8ToFollow, mUtf8Index;
private final byte[] mUtf8InputBuffer = new byte[4];
public final TerminalColors mColors = new TerminalColors();
@@ -424,7 +424,11 @@ public final class TerminalEmulator {
processCodePoint(/* escape (hexadecimal=0x1B, octal=033): */27);
processCodePoint((codePoint & 0x7F) + 0x40);
} else {
if (Character.UNASSIGNED == Character.getType(codePoint)) codePoint = UNICODE_REPLACEMENT_CHAR;
switch (Character.getType(codePoint)) {
case Character.UNASSIGNED:
case Character.SURROGATE:
codePoint = UNICODE_REPLACEMENT_CHAR;
}
processCodePoint(codePoint);
}
}

41
app/src/test/java/com/termux/terminal/UnicodeInputTest.java
View File

@@ -12,6 +12,47 @@ public class UnicodeInputTest extends TerminalTestCase {
withTerminalSized(5, 5);
mTerminal.append(new byte[]{(byte) 0b11101111, (byte) 'a'}, 2);
assertLineIs(0, ((char) TerminalEmulator.UNICODE_REPLACEMENT_CHAR) + "a ");
// https://code.google.com/p/chromium/issues/detail?id=212704
byte[] input = new byte[]{
(byte) 0x61, (byte) 0xF1,
(byte) 0x80, (byte) 0x80,
(byte) 0xe1, (byte) 0x80,
(byte) 0xc2, (byte) 0x62,
(byte) 0x80, (byte) 0x63,
(byte) 0x80, (byte) 0xbf,
(byte) 0x64
};
withTerminalSized(10, 2);
mTerminal.append(input, input.length);
assertLinesAre("a\uFFFD\uFFFD\uFFFDb\uFFFDc\uFFFD\uFFFDd", " ");
// Surrogate pairs.
withTerminalSized(5, 2);
input = new byte[]{
(byte) 0xed, (byte) 0xa0,
(byte) 0x80, (byte) 0xed,
(byte) 0xad, (byte) 0xbf,
(byte) 0xed, (byte) 0xae,
(byte) 0x80, (byte) 0xed,
(byte) 0xbf, (byte) 0xbf
};
mTerminal.append(input, input.length);
assertLinesAre("\uFFFD\uFFFD\uFFFD\uFFFD ", " ");
// https://bugzilla.mozilla.org/show_bug.cgi?id=746900: "with this patch 0xe0 0x80 is decoded as two U+FFFDs,
// but 0xe0 0xa0 is decoded as a single U+FFFD, and this is correct according to the "Best Practices", but IE
// and Chrome (Version 22.0.1229.94) decode both of them as two U+FFFDs. Opera 12.11 decodes both of them as
// one U+FFFD".
withTerminalSized(5, 2);
input = new byte[]{(byte) 0xe0, (byte) 0xa0, ' '};
mTerminal.append(input, input.length);
assertLinesAre("\uFFFD ", " ");
// withTerminalSized(5, 2);
// input = new byte[]{(byte) 0xe0, (byte) 0x80, 'a'};
// mTerminal.append(input, input.length);
// assertLinesAre("\uFFFD\uFFFDa ", " ");
}
public void testUnassignedCodePoint() throws UnsupportedEncodingException {