mainnika
/
weston
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity

terminal: Escape sequence handling fixes

Browse Source 
Upgrade and refactor terminal_data to properly handle non-csi escape
codes, control characters in escape codes, and invalid escape sequences.
Also fix a buffer overflow in the escape sequence buffer.

Signed-off-by: Callum Lowcay <callum@callumscode.com>
dev
Callum Lowcay 14 years ago committed by Kristian Høgsberg
parent a0ee21c7dc
commit b8609ada50
1 changed files with 156 additions and 57 deletions
Show all changes Ignore whitespace when comparing lines Ignore changes in amount of whitespace Ignore changes in whitespace at EOL
Show Stats Download Patch File Download Diff File
  1. 201
      clients/terminal.c

201
clients/terminal.c
Unescape View File

@ -52,6 +52,10 @@ static int option_fullscreen;
#define ATTRMASK_BLINK 0x04
#define ATTRMASK_INVERSE 0x08
/* Buffer sizes */
#define MAX_RESPONSE 11
#define MAX_ESCAPE 64
union utf8_char {
unsigned char byte[4];
uint32_t ch;
@ -189,9 +193,10 @@ struct terminal {
int fd, master;
GIOChannel *channel;
uint32_t modifiers;
char escape[64];
char escape[MAX_ESCAPE];
int escape_length;
int state;
int qmark_flag;
struct utf8_state_machine state_machine;
int margin;
int fullscreen;
@ -513,10 +518,15 @@ redraw_handler(struct window *window, void *data)
#define STATE_NORMAL 0
#define STATE_ESCAPE 1
#define STATE_ESCAPE_SPECIAL 2
#define STATE_ESCAPE_CSI 3
static void
terminal_data(struct terminal *terminal, const char *data, size_t length);
static void
handle_char(struct terminal *terminal, union utf8_char utf8);
static void
handle_sgr(struct terminal *terminal, int code);
@ -629,13 +639,21 @@ handle_escape(struct terminal *terminal)
}
break;
default:
terminal_data(terminal,
terminal->escape + 1,
terminal->escape_length - 2);
fprintf(stderr, "Unknown CSI escape: %c\n", *p);
break;
}
}
static void
handle_non_csi_escape(struct terminal *terminal, char code)
{
}
static void
handle_special_escape(struct terminal *terminal, char special, char code)
{
}
static void
handle_sgr(struct terminal *terminal, int code)
{
@ -698,58 +716,25 @@ handle_sgr(struct terminal *terminal, int code)
}
}
static void
terminal_data(struct terminal *terminal, const char *data, size_t length)
/* Returns 1 if c was special, otherwise 0 */
static int
handle_special_char(struct terminal *terminal, char c)
{
int i;
union utf8_char utf8;
enum utf8_state parser_state;
union utf8_char *row;
struct attr *attr_row;
for (i = 0; i < length; i++) {
parser_state =
utf8_next_char(&terminal->state_machine, data[i]);
switch(parser_state) {
case utf8state_accept:
utf8.ch = terminal->state_machine.s.ch;
break;
case utf8state_reject:
/* the unicode replacement character */
utf8.byte[0] = 0xEF;
utf8.byte[1] = 0xBF;
utf8.byte[2] = 0xBD;
utf8.byte[3] = 0x00;
break;
default:
continue;
}
row = terminal_get_row(terminal, terminal->row);
attr_row = terminal_get_attr_row(terminal, terminal->row);
if (terminal->state == STATE_ESCAPE) {
terminal->escape[terminal->escape_length++] = utf8.byte[0];
if (terminal->escape_length == 2 && utf8.byte[0] != '[') {
/* Bad escape sequence. */
terminal->state = STATE_NORMAL;
goto cancel_escape;
}
if (isalpha(utf8.byte[0])) {
terminal->state = STATE_NORMAL;
handle_escape(terminal);
}
continue;
}
cancel_escape:
switch (utf8.byte[0]) {
switch(c) {
case '\r':
terminal->column = 0;
break;
case '\n':
terminal->column = 0;
/* fallthrough */
case '\v':
case '\f':
if (terminal->row + 1 < terminal->height) {
terminal->row++;
} else {
@ -768,11 +753,6 @@ terminal_data(struct terminal *terminal, const char *data, size_t length)
attr_init(&attr_row[terminal->column], terminal->curr_attr, -terminal->column & 7);
terminal->column = (terminal->column + 7) & ~7;
break;
case '\e':
terminal->state = STATE_ESCAPE;
terminal->escape[0] = '\e';
terminal->escape_length = 1;
break;
case '\b':
if (terminal->column > 0)
terminal->column--;
@ -781,14 +761,133 @@ terminal_data(struct terminal *terminal, const char *data, size_t length)
/* Bell */
break;
default:
if (terminal->column < terminal->width)
if (utf8.byte[0] < 32) utf8.byte[0] += 64;
return 0;
}
return 1;
}
static void
handle_char(struct terminal *terminal, union utf8_char utf8)
{
union utf8_char *row;
struct attr *attr_row;
if (handle_special_char(terminal, utf8.byte[0])) return;
/* There are a whole lot of non-characters, control codes,
* and formatting codes that should probably be ignored,
* for example: */
if (strncmp((char*) utf8.byte, "\xEF\xBB\xBF", 3) == 0) {
/* BOM, ignore */
return;
}
/* Some of these non-characters should be translated, e.g.: */
if (utf8.byte[0] < 32) {
utf8.byte[0] = utf8.byte[0] + 64;
}
/* handle right margin effects */
if (terminal->column >= terminal->width) {
terminal->column--;
}
row = terminal_get_row(terminal, terminal->row);
attr_row = terminal_get_attr_row(terminal, terminal->row);
row[terminal->column] = utf8;
attr_row[terminal->column++] = terminal->curr_attr;
if (utf8.ch != terminal->last_char.ch)
terminal->last_char = utf8;
}
static void
terminal_data(struct terminal *terminal, const char *data, size_t length)
{
int i;
union utf8_char utf8;
enum utf8_state parser_state;
for (i = 0; i < length; i++) {
parser_state =
utf8_next_char(&terminal->state_machine, data[i]);
switch(parser_state) {
case utf8state_accept:
utf8.ch = terminal->state_machine.s.ch;
break;
case utf8state_reject:
/* the unicode replacement character */
utf8.byte[0] = 0xEF;
utf8.byte[1] = 0xBF;
utf8.byte[2] = 0xBD;
utf8.byte[3] = 0x00;
break;
default:
continue;
}
/* assume escape codes never use non-ASCII characters */
if (terminal->state == STATE_ESCAPE) {
terminal->escape[terminal->escape_length++] = utf8.byte[0];
if (utf8.byte[0] == '[') {
terminal->state = STATE_ESCAPE_CSI;
continue;
} else if (utf8.byte[0] == '#' || utf8.byte[0] == '(' ||
utf8.byte[0] == ')')
{
terminal->state = STATE_ESCAPE_SPECIAL;
continue;
} else {
terminal->state = STATE_NORMAL;
handle_non_csi_escape(terminal, utf8.byte[0]);
continue;
}
} else if (terminal->state == STATE_ESCAPE_SPECIAL) {
terminal->escape[terminal->escape_length++] = utf8.byte[0];
terminal->state = STATE_NORMAL;
if (isdigit(utf8.byte[0]) || isalpha(utf8.byte[0])) {
handle_special_escape(terminal, terminal->escape[1],
utf8.byte[0]);
continue;
}
} else if (terminal->state == STATE_ESCAPE_CSI) {
if (handle_special_char(terminal, utf8.byte[0]) != 0) {
/* do nothing */
} else if (utf8.byte[0] == '?') {
terminal->qmark_flag = 1;
} else {
/* Don't overflow the buffer */
if (terminal->escape_length < MAX_ESCAPE)
terminal->escape[terminal->escape_length++] = utf8.byte[0];
if (terminal->escape_length >= MAX_ESCAPE)
terminal->state = STATE_NORMAL;
}
if (isalpha(utf8.byte[0]) || utf8.byte[0] == '@' ||
utf8.byte[0] == '`')
{
terminal->state = STATE_NORMAL;
handle_escape(terminal);
continue;
} else {
continue;
}
}
/* this is valid, because ASCII characters are never used to
* introduce a multibyte sequence in UTF-8 */
if (utf8.byte[0] == '\e') {
terminal->state = STATE_ESCAPE;
terminal->escape[0] = '\e';
terminal->escape_length = 1;
terminal->qmark_flag = 0;
} else {
handle_char(terminal, utf8);
} /* if */
} /* for */
window_schedule_redraw(terminal->window);
}

Write Preview
Loading…
Cancel
Save