On 2025-12-16 05:13, Corinna Vinschen wrote:
On Dec 10 10:52, Takashi Yano wrote:
This patch fixes the bug in ESC sequence parser used when pseudo
console is enabled in pty_master_fwd_thread(). Previously, if
multiple ESC sequences exist in a fowarding chunk, the later one
might not be processed appropriately. In addition, the termination
ST was not supported, that is, only BEL was supported.
What's ST? I only know STX, 0x02 in the C0 codeblock. There's an ST in
the C1 codeblock, 0x9c, "String Terminator", but I don't see this in the
code, nor are the other C1 controls even recognized here.
Some symbolic character function names would help a lot!
ST \E\\ is an alternative to \a <BEL> another ANSI X3.64/FIPS-86/ECMA-48/DEC STD
070/ISO 6429 control sequence with 7/8 bit alternatives like APC, DCS, CSI, OSC, PM:
https://invisible-island.net/xterm/ctlseqs/ctlseqs.html#h3-Operating-System-Commands
"Operating System Commands
OSC Ps ; Pt BEL
OSC Ps ; Pt ST"
https://invisible-island.net/xterm/ctlseqs/ctlseqs.html#h3-Control-Bytes_-Characters_-and-Sequences
"o Some controls (such as OSC) introduce a string mode, which is ended on a ST
(string terminator)."
https://invisible-island.net/xterm/ctlseqs/ctlseqs.html#h3-C1-lparen-8-Bit-rparen-Control-Characters
"...
ESC P Device Control String (DCS is 0x90).
...
ESC [ Control Sequence Introducer (CSI is 0x9b).
ESC \ String Terminator (ST is 0x9c).
ESC ] Operating System Command (OSC is 0x9d).
ESC ^ Privacy Message (PM is 0x9e).
ESC _ Application Program Command (APC is 0x9f).
..."
ANSI X3.64-1979 FIPS-86 Additional Controls for Use with ANS Code for
Information Interchange 1979-07-18
https://nvlpubs.nist.gov/nistpubs/Legacy/FIPS/fipspub86.pdf
ECMA-48 Control Functions for Coded Character Sets 5th.ed 1991-06
https://ecma-international.org/wp-content/uploads/ECMA-48_5th_edition_june_1991.pdf
EL-SM070-00 DEC STD 070 Video Systems Reference Manual 1991-12-03
http://www.bitsavers.org/pdf/dec/standards/EL-SM070-00_DEC_STD_070_Video_Systems_Reference_Manual_Dec91.pdf
ISO/IEC 6429:1992 Control functions for coded character sets Ed.3 1992-12
https://www.iso.org/standard/12782.html
Fixes: 10d083c745dd ("Cygwin: pty: Inherit typeahead data between two input
pipes.")
Reviewed-by:
Signed-off-by: Takashi Yano <[email protected]>
---
winsup/cygwin/fhandler/pty.cc | 13 +++++++++----
1 file changed, 9 insertions(+), 4 deletions(-)
diff --git a/winsup/cygwin/fhandler/pty.cc b/winsup/cygwin/fhandler/pty.cc
index 679068ea2..3b0b4f073 100644
--- a/winsup/cygwin/fhandler/pty.cc
+++ b/winsup/cygwin/fhandler/pty.cc
@@ -2680,7 +2680,7 @@ fhandler_pty_master::pty_master_fwd_thread (const
master_fwd_thread_param_t *p)
int state = 0;
int start_at = 0;
for (DWORD i=0; i<rlen; i++)
- if (outbuf[i] == '\033')
+ if (state == 0 && outbuf[i] == '\033')
^ ESC '\x1b'
As a long former RAD50 and 08 fan can we please move on to hex \x1b now? ;^>
{
start_at = i;
state = 1;
@@ -2688,12 +2688,14 @@ fhandler_pty_master::pty_master_fwd_thread (const
master_fwd_thread_param_t *p)
}
else if ((state == 1 && outbuf[i] == ']') ||
^ OSC
(state == 2 && outbuf[i] == '0') ||
^ OSC_ICN_NM_WIN_TTL
- (state == 3 && outbuf[i] == ';'))
+ (state == 3 && outbuf[i] == ';') ||
^ PARAM_SEP
+ (state == 4 && outbuf[i] == '\033'))
^ ESC '\x1b'
{
state ++;
continue;
}
- else if (state == 4 && outbuf[i] == '\a')
+ else if ((state == 4 && outbuf[i] == '\a')
^ BEL
+ || (state == 5 && outbuf[i] == '\\'))
^ ST
{
const char *helper_str = "\\bin\\cygwin-console-helper.exe";
if (memmem (&outbuf[start_at], i + 1 - start_at,
@@ -2701,11 +2703,14 @@ fhandler_pty_master::pty_master_fwd_thread (const
master_fwd_thread_param_t *p)
{
memmove (&outbuf[start_at], &outbuf[i+1], rlen-i-1);
rlen = wlen = start_at + rlen - i - 1;
+ i = start_at - 1;
}
state = 0;
continue;
}
- else if (outbuf[i] == '\a')
+ else if (state == 4)
+ continue;
+ else
{
state = 0;
continue;
--
2.51.0
[IMHO literals should never appear in code - except *maybe sometimes* 0!
Define all your constants with meaningful names in context before all code,
or in a meaningfully named header with those definitions, if you can't steal
them from somewhere with a history, say mintty or xterm sources, or ancient
assembler macros, convert them from a document table, say man pages `man -m
linux console_codes`, Wikipedia, or xterm (above), or from references therein.]
--
Take care. Thanks, Brian Inglis Calgary, Alberta, Canada
La perfection est atteinte Perfection is achieved
non pas lorsqu'il n'y a plus rien à ajouter not when there is no more to add
mais lorsqu'il n'y a plus rien à retrancher but when there is no more to cut
-- Antoine de Saint-Exupéry
