2023-03-21 14:26:02 +08:00
|
|
|
#include "git-compat-util.h"
|
add generic terminal prompt function
When we need to prompt the user for input interactively, we
want to access their terminal directly. We can't rely on
stdio because it may be connected to pipes or files, rather
than the terminal. Instead, we use "getpass()", because it
abstracts the idea of prompting and reading from the
terminal. However, it has some problems:
1. It never echoes the typed characters, which makes it OK
for passwords but annoying for other input (like usernames).
2. Some implementations of getpass() have an extremely
small input buffer (e.g., Solaris 8 is reported to
support only 8 characters).
3. Some implementations of getpass() will fall back to
reading from stdin (e.g., glibc). We explicitly don't
want this, because our stdin may be connected to a pipe
speaking a particular protocol, and reading will
disrupt the protocol flow (e.g., the remote-curl
helper).
4. Some implementations of getpass() turn off signals, so
that hitting "^C" on the terminal does not break out of
the password prompt. This can be a mild annoyance.
Instead, let's provide an abstract "git_terminal_prompt"
function that addresses these concerns. This patch includes
an implementation based on /dev/tty, enabled by setting
HAVE_DEV_TTY. The fallback is to use getpass() as before.
Signed-off-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2011-12-10 18:41:01 +08:00
|
|
|
#include "compat/terminal.h"
|
2023-03-21 14:25:54 +08:00
|
|
|
#include "gettext.h"
|
add generic terminal prompt function
When we need to prompt the user for input interactively, we
want to access their terminal directly. We can't rely on
stdio because it may be connected to pipes or files, rather
than the terminal. Instead, we use "getpass()", because it
abstracts the idea of prompting and reading from the
terminal. However, it has some problems:
1. It never echoes the typed characters, which makes it OK
for passwords but annoying for other input (like usernames).
2. Some implementations of getpass() have an extremely
small input buffer (e.g., Solaris 8 is reported to
support only 8 characters).
3. Some implementations of getpass() will fall back to
reading from stdin (e.g., glibc). We explicitly don't
want this, because our stdin may be connected to a pipe
speaking a particular protocol, and reading will
disrupt the protocol flow (e.g., the remote-curl
helper).
4. Some implementations of getpass() turn off signals, so
that hitting "^C" on the terminal does not break out of
the password prompt. This can be a mild annoyance.
Instead, let's provide an abstract "git_terminal_prompt"
function that addresses these concerns. This patch includes
an implementation based on /dev/tty, enabled by setting
HAVE_DEV_TTY. The fallback is to use getpass() as before.
Signed-off-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2011-12-10 18:41:01 +08:00
|
|
|
#include "sigchain.h"
|
|
|
|
#include "strbuf.h"
|
2020-01-15 02:43:48 +08:00
|
|
|
#include "run-command.h"
|
|
|
|
#include "string-list.h"
|
built-in add -p: handle Escape sequences more efficiently
When `interactive.singlekey = true`, we react immediately to keystrokes,
even to Escape sequences (e.g. when pressing a cursor key).
The problem with Escape sequences is that we do not really know when
they are done, and as a heuristic we poll standard input for half a
second to make sure that we got all of it.
While waiting half a second is not asking for a whole lot, it can become
quite annoying over time, therefore with this patch, we read the
terminal capabilities (if available) and extract known Escape sequences
from there, then stop polling immediately when we detected that the user
pressed a key that generated such a known sequence.
This recapitulates the remaining part of b5cc003253c8 (add -i: ignore
terminal escape sequences, 2011-05-17).
Note: We do *not* query the terminal capabilities directly. That would
either require a lot of platform-specific code, or it would require
linking to a library such as ncurses.
Linking to a library in the built-ins is something we try very hard to
avoid (we even kicked the libcurl dependency to a non-built-in remote
helper, just to shave off a tiny fraction of a second from Git's startup
time). And the platform-specific code would be a maintenance nightmare.
Even worse: in Git for Windows' case, we would need to query MSYS2
pseudo terminals, which `git.exe` simply cannot do (because it is
intentionally *not* an MSYS2 program).
To address this, we simply spawn `infocmp -L -1` and parse its output
(which works even in Git for Windows, because that helper is included in
the end-user facing installations).
This is done only once, as in the Perl version, but it is done only when
the first Escape sequence is encountered, not upon startup of `git add
-i`; This saves on startup time, yet makes reacting to the first Escape
sequence slightly more sluggish. But it allows us to keep the
terminal-related code encapsulated in the `compat/terminal.c` file.
Signed-off-by: Johannes Schindelin <johannes.schindelin@gmx.de>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2020-01-15 02:43:52 +08:00
|
|
|
#include "hashmap.h"
|
add generic terminal prompt function
When we need to prompt the user for input interactively, we
want to access their terminal directly. We can't rely on
stdio because it may be connected to pipes or files, rather
than the terminal. Instead, we use "getpass()", because it
abstracts the idea of prompting and reading from the
terminal. However, it has some problems:
1. It never echoes the typed characters, which makes it OK
for passwords but annoying for other input (like usernames).
2. Some implementations of getpass() have an extremely
small input buffer (e.g., Solaris 8 is reported to
support only 8 characters).
3. Some implementations of getpass() will fall back to
reading from stdin (e.g., glibc). We explicitly don't
want this, because our stdin may be connected to a pipe
speaking a particular protocol, and reading will
disrupt the protocol flow (e.g., the remote-curl
helper).
4. Some implementations of getpass() turn off signals, so
that hitting "^C" on the terminal does not break out of
the password prompt. This can be a mild annoyance.
Instead, let's provide an abstract "git_terminal_prompt"
function that addresses these concerns. This patch includes
an implementation based on /dev/tty, enabled by setting
HAVE_DEV_TTY. The fallback is to use getpass() as before.
Signed-off-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2011-12-10 18:41:01 +08:00
|
|
|
|
2013-05-03 03:26:08 +08:00
|
|
|
#if defined(HAVE_DEV_TTY) || defined(GIT_WINDOWS_NATIVE)
|
2012-12-04 16:10:41 +08:00
|
|
|
|
|
|
|
static void restore_term_on_signal(int sig)
|
|
|
|
{
|
|
|
|
restore_term();
|
2022-02-23 02:53:33 +08:00
|
|
|
/* restore_term calls sigchain_pop_common */
|
2012-12-04 16:10:41 +08:00
|
|
|
raise(sig);
|
|
|
|
}
|
|
|
|
|
add generic terminal prompt function
When we need to prompt the user for input interactively, we
want to access their terminal directly. We can't rely on
stdio because it may be connected to pipes or files, rather
than the terminal. Instead, we use "getpass()", because it
abstracts the idea of prompting and reading from the
terminal. However, it has some problems:
1. It never echoes the typed characters, which makes it OK
for passwords but annoying for other input (like usernames).
2. Some implementations of getpass() have an extremely
small input buffer (e.g., Solaris 8 is reported to
support only 8 characters).
3. Some implementations of getpass() will fall back to
reading from stdin (e.g., glibc). We explicitly don't
want this, because our stdin may be connected to a pipe
speaking a particular protocol, and reading will
disrupt the protocol flow (e.g., the remote-curl
helper).
4. Some implementations of getpass() turn off signals, so
that hitting "^C" on the terminal does not break out of
the password prompt. This can be a mild annoyance.
Instead, let's provide an abstract "git_terminal_prompt"
function that addresses these concerns. This patch includes
an implementation based on /dev/tty, enabled by setting
HAVE_DEV_TTY. The fallback is to use getpass() as before.
Signed-off-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2011-12-10 18:41:01 +08:00
|
|
|
#ifdef HAVE_DEV_TTY
|
|
|
|
|
2012-12-04 16:10:41 +08:00
|
|
|
#define INPUT_PATH "/dev/tty"
|
|
|
|
#define OUTPUT_PATH "/dev/tty"
|
|
|
|
|
2022-03-17 02:54:03 +08:00
|
|
|
static volatile sig_atomic_t term_fd_needs_closing;
|
add generic terminal prompt function
When we need to prompt the user for input interactively, we
want to access their terminal directly. We can't rely on
stdio because it may be connected to pipes or files, rather
than the terminal. Instead, we use "getpass()", because it
abstracts the idea of prompting and reading from the
terminal. However, it has some problems:
1. It never echoes the typed characters, which makes it OK
for passwords but annoying for other input (like usernames).
2. Some implementations of getpass() have an extremely
small input buffer (e.g., Solaris 8 is reported to
support only 8 characters).
3. Some implementations of getpass() will fall back to
reading from stdin (e.g., glibc). We explicitly don't
want this, because our stdin may be connected to a pipe
speaking a particular protocol, and reading will
disrupt the protocol flow (e.g., the remote-curl
helper).
4. Some implementations of getpass() turn off signals, so
that hitting "^C" on the terminal does not break out of
the password prompt. This can be a mild annoyance.
Instead, let's provide an abstract "git_terminal_prompt"
function that addresses these concerns. This patch includes
an implementation based on /dev/tty, enabled by setting
HAVE_DEV_TTY. The fallback is to use getpass() as before.
Signed-off-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2011-12-10 18:41:01 +08:00
|
|
|
static int term_fd = -1;
|
|
|
|
static struct termios old_term;
|
|
|
|
|
2022-03-17 02:54:05 +08:00
|
|
|
static const char *background_resume_msg;
|
|
|
|
static const char *restore_error_msg;
|
|
|
|
static volatile sig_atomic_t ttou_received;
|
|
|
|
|
|
|
|
/* async safe error function for use by signal handlers. */
|
|
|
|
static void write_err(const char *msg)
|
|
|
|
{
|
|
|
|
write_in_full(2, "error: ", strlen("error: "));
|
|
|
|
write_in_full(2, msg, strlen(msg));
|
|
|
|
write_in_full(2, "\n", 1);
|
|
|
|
}
|
|
|
|
|
|
|
|
static void print_background_resume_msg(int signo)
|
|
|
|
{
|
|
|
|
int saved_errno = errno;
|
|
|
|
sigset_t mask;
|
|
|
|
struct sigaction old_sa;
|
|
|
|
struct sigaction sa = { .sa_handler = SIG_DFL };
|
|
|
|
|
|
|
|
ttou_received = 1;
|
|
|
|
write_err(background_resume_msg);
|
|
|
|
sigaction(signo, &sa, &old_sa);
|
|
|
|
raise(signo);
|
|
|
|
sigemptyset(&mask);
|
|
|
|
sigaddset(&mask, signo);
|
|
|
|
sigprocmask(SIG_UNBLOCK, &mask, NULL);
|
|
|
|
/* Stopped here */
|
|
|
|
sigprocmask(SIG_BLOCK, &mask, NULL);
|
|
|
|
sigaction(signo, &old_sa, NULL);
|
|
|
|
errno = saved_errno;
|
|
|
|
}
|
|
|
|
|
|
|
|
static void restore_terminal_on_suspend(int signo)
|
|
|
|
{
|
|
|
|
int saved_errno = errno;
|
|
|
|
int res;
|
|
|
|
struct termios t;
|
|
|
|
sigset_t mask;
|
|
|
|
struct sigaction old_sa;
|
|
|
|
struct sigaction sa = { .sa_handler = SIG_DFL };
|
|
|
|
int can_restore = 1;
|
|
|
|
|
|
|
|
if (tcgetattr(term_fd, &t) < 0)
|
|
|
|
can_restore = 0;
|
|
|
|
|
|
|
|
if (tcsetattr(term_fd, TCSAFLUSH, &old_term) < 0)
|
|
|
|
write_err(restore_error_msg);
|
|
|
|
|
|
|
|
sigaction(signo, &sa, &old_sa);
|
|
|
|
raise(signo);
|
|
|
|
sigemptyset(&mask);
|
|
|
|
sigaddset(&mask, signo);
|
|
|
|
sigprocmask(SIG_UNBLOCK, &mask, NULL);
|
|
|
|
/* Stopped here */
|
|
|
|
sigprocmask(SIG_BLOCK, &mask, NULL);
|
|
|
|
sigaction(signo, &old_sa, NULL);
|
|
|
|
if (!can_restore) {
|
|
|
|
write_err(restore_error_msg);
|
|
|
|
goto out;
|
|
|
|
}
|
|
|
|
/*
|
|
|
|
* If we resume in the background then we receive SIGTTOU when calling
|
|
|
|
* tcsetattr() below. Set up a handler to print an error message in that
|
|
|
|
* case.
|
|
|
|
*/
|
|
|
|
sigemptyset(&mask);
|
|
|
|
sigaddset(&mask, SIGTTOU);
|
|
|
|
sa.sa_mask = old_sa.sa_mask;
|
|
|
|
sa.sa_handler = print_background_resume_msg;
|
|
|
|
sa.sa_flags = SA_RESTART;
|
|
|
|
sigaction(SIGTTOU, &sa, &old_sa);
|
|
|
|
again:
|
|
|
|
ttou_received = 0;
|
|
|
|
sigprocmask(SIG_UNBLOCK, &mask, NULL);
|
|
|
|
res = tcsetattr(term_fd, TCSAFLUSH, &t);
|
|
|
|
sigprocmask(SIG_BLOCK, &mask, NULL);
|
|
|
|
if (ttou_received)
|
|
|
|
goto again;
|
|
|
|
else if (res < 0)
|
|
|
|
write_err(restore_error_msg);
|
|
|
|
sigaction(SIGTTOU, &old_sa, NULL);
|
|
|
|
out:
|
|
|
|
errno = saved_errno;
|
|
|
|
}
|
|
|
|
|
|
|
|
static void reset_job_signals(void)
|
|
|
|
{
|
|
|
|
if (restore_error_msg) {
|
|
|
|
signal(SIGTTIN, SIG_DFL);
|
|
|
|
signal(SIGTTOU, SIG_DFL);
|
|
|
|
signal(SIGTSTP, SIG_DFL);
|
|
|
|
restore_error_msg = NULL;
|
|
|
|
background_resume_msg = NULL;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2022-03-17 02:54:03 +08:00
|
|
|
static void close_term_fd(void)
|
|
|
|
{
|
|
|
|
if (term_fd_needs_closing)
|
|
|
|
close(term_fd);
|
|
|
|
term_fd_needs_closing = 0;
|
|
|
|
term_fd = -1;
|
|
|
|
}
|
|
|
|
|
2021-10-05 15:46:47 +08:00
|
|
|
void restore_term(void)
|
add generic terminal prompt function
When we need to prompt the user for input interactively, we
want to access their terminal directly. We can't rely on
stdio because it may be connected to pipes or files, rather
than the terminal. Instead, we use "getpass()", because it
abstracts the idea of prompting and reading from the
terminal. However, it has some problems:
1. It never echoes the typed characters, which makes it OK
for passwords but annoying for other input (like usernames).
2. Some implementations of getpass() have an extremely
small input buffer (e.g., Solaris 8 is reported to
support only 8 characters).
3. Some implementations of getpass() will fall back to
reading from stdin (e.g., glibc). We explicitly don't
want this, because our stdin may be connected to a pipe
speaking a particular protocol, and reading will
disrupt the protocol flow (e.g., the remote-curl
helper).
4. Some implementations of getpass() turn off signals, so
that hitting "^C" on the terminal does not break out of
the password prompt. This can be a mild annoyance.
Instead, let's provide an abstract "git_terminal_prompt"
function that addresses these concerns. This patch includes
an implementation based on /dev/tty, enabled by setting
HAVE_DEV_TTY. The fallback is to use getpass() as before.
Signed-off-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2011-12-10 18:41:01 +08:00
|
|
|
{
|
|
|
|
if (term_fd < 0)
|
|
|
|
return;
|
|
|
|
|
|
|
|
tcsetattr(term_fd, TCSAFLUSH, &old_term);
|
2022-03-17 02:54:03 +08:00
|
|
|
close_term_fd();
|
2022-02-23 02:53:33 +08:00
|
|
|
sigchain_pop_common();
|
2022-03-17 02:54:05 +08:00
|
|
|
reset_job_signals();
|
add generic terminal prompt function
When we need to prompt the user for input interactively, we
want to access their terminal directly. We can't rely on
stdio because it may be connected to pipes or files, rather
than the terminal. Instead, we use "getpass()", because it
abstracts the idea of prompting and reading from the
terminal. However, it has some problems:
1. It never echoes the typed characters, which makes it OK
for passwords but annoying for other input (like usernames).
2. Some implementations of getpass() have an extremely
small input buffer (e.g., Solaris 8 is reported to
support only 8 characters).
3. Some implementations of getpass() will fall back to
reading from stdin (e.g., glibc). We explicitly don't
want this, because our stdin may be connected to a pipe
speaking a particular protocol, and reading will
disrupt the protocol flow (e.g., the remote-curl
helper).
4. Some implementations of getpass() turn off signals, so
that hitting "^C" on the terminal does not break out of
the password prompt. This can be a mild annoyance.
Instead, let's provide an abstract "git_terminal_prompt"
function that addresses these concerns. This patch includes
an implementation based on /dev/tty, enabled by setting
HAVE_DEV_TTY. The fallback is to use getpass() as before.
Signed-off-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2011-12-10 18:41:01 +08:00
|
|
|
}
|
|
|
|
|
2022-03-17 02:54:02 +08:00
|
|
|
int save_term(enum save_term_flags flags)
|
2021-10-05 15:46:47 +08:00
|
|
|
{
|
2022-03-17 02:54:05 +08:00
|
|
|
struct sigaction sa;
|
|
|
|
|
2021-10-05 15:46:47 +08:00
|
|
|
if (term_fd < 0)
|
2022-03-17 02:54:03 +08:00
|
|
|
term_fd = ((flags & SAVE_TERM_STDIN)
|
|
|
|
? 0
|
|
|
|
: open("/dev/tty", O_RDWR));
|
2022-02-23 02:53:33 +08:00
|
|
|
if (term_fd < 0)
|
|
|
|
return -1;
|
2022-03-17 02:54:03 +08:00
|
|
|
term_fd_needs_closing = !(flags & SAVE_TERM_STDIN);
|
|
|
|
if (tcgetattr(term_fd, &old_term) < 0) {
|
|
|
|
close_term_fd();
|
2022-02-23 02:53:33 +08:00
|
|
|
return -1;
|
2022-03-17 02:54:03 +08:00
|
|
|
}
|
2022-02-23 02:53:33 +08:00
|
|
|
sigchain_push_common(restore_term_on_signal);
|
2022-03-17 02:54:05 +08:00
|
|
|
/*
|
|
|
|
* If job control is disabled then the shell will have set the
|
|
|
|
* disposition of SIGTSTP to SIG_IGN.
|
|
|
|
*/
|
|
|
|
sigaction(SIGTSTP, NULL, &sa);
|
|
|
|
if (sa.sa_handler == SIG_IGN)
|
|
|
|
return 0;
|
|
|
|
|
|
|
|
/* avoid calling gettext() from signal handler */
|
|
|
|
background_resume_msg = _("cannot resume in the background, please use 'fg' to resume");
|
|
|
|
restore_error_msg = _("cannot restore terminal settings");
|
|
|
|
sa.sa_handler = restore_terminal_on_suspend;
|
|
|
|
sa.sa_flags = SA_RESTART;
|
|
|
|
sigemptyset(&sa.sa_mask);
|
|
|
|
sigaddset(&sa.sa_mask, SIGTSTP);
|
|
|
|
sigaddset(&sa.sa_mask, SIGTTIN);
|
|
|
|
sigaddset(&sa.sa_mask, SIGTTOU);
|
|
|
|
sigaction(SIGTSTP, &sa, NULL);
|
|
|
|
sigaction(SIGTTIN, &sa, NULL);
|
|
|
|
sigaction(SIGTTOU, &sa, NULL);
|
2021-10-05 15:46:47 +08:00
|
|
|
|
2022-02-23 02:53:33 +08:00
|
|
|
return 0;
|
2021-10-05 15:46:47 +08:00
|
|
|
}
|
|
|
|
|
2022-03-17 02:54:02 +08:00
|
|
|
static int disable_bits(enum save_term_flags flags, tcflag_t bits)
|
2012-12-04 16:10:39 +08:00
|
|
|
{
|
|
|
|
struct termios t;
|
|
|
|
|
2022-03-17 02:54:02 +08:00
|
|
|
if (save_term(flags) < 0)
|
2022-03-17 02:54:03 +08:00
|
|
|
return -1;
|
2012-12-04 16:10:39 +08:00
|
|
|
|
2021-10-05 15:46:47 +08:00
|
|
|
t = old_term;
|
2012-12-04 16:10:39 +08:00
|
|
|
|
2020-01-15 02:43:47 +08:00
|
|
|
t.c_lflag &= ~bits;
|
2022-02-23 02:53:34 +08:00
|
|
|
if (bits & ICANON) {
|
|
|
|
t.c_cc[VMIN] = 1;
|
|
|
|
t.c_cc[VTIME] = 0;
|
|
|
|
}
|
2012-12-04 16:10:39 +08:00
|
|
|
if (!tcsetattr(term_fd, TCSAFLUSH, &t))
|
|
|
|
return 0;
|
|
|
|
|
2022-02-23 02:53:33 +08:00
|
|
|
sigchain_pop_common();
|
2022-03-17 02:54:05 +08:00
|
|
|
reset_job_signals();
|
2022-03-17 02:54:03 +08:00
|
|
|
close_term_fd();
|
2012-12-04 16:10:39 +08:00
|
|
|
return -1;
|
|
|
|
}
|
|
|
|
|
2022-03-17 02:54:02 +08:00
|
|
|
static int disable_echo(enum save_term_flags flags)
|
2020-01-15 02:43:47 +08:00
|
|
|
{
|
2022-03-17 02:54:02 +08:00
|
|
|
return disable_bits(flags, ECHO);
|
2020-01-15 02:43:47 +08:00
|
|
|
}
|
|
|
|
|
2022-03-17 02:54:02 +08:00
|
|
|
static int enable_non_canonical(enum save_term_flags flags)
|
2020-01-15 02:43:49 +08:00
|
|
|
{
|
2022-03-17 02:54:02 +08:00
|
|
|
return disable_bits(flags, ICANON | ECHO);
|
2020-01-15 02:43:49 +08:00
|
|
|
}
|
|
|
|
|
2022-03-17 02:54:04 +08:00
|
|
|
/*
|
|
|
|
* On macos it is not possible to use poll() with a terminal so use select
|
|
|
|
* instead.
|
|
|
|
*/
|
|
|
|
static int getchar_with_timeout(int timeout)
|
|
|
|
{
|
|
|
|
struct timeval tv, *tvp = NULL;
|
|
|
|
fd_set readfds;
|
|
|
|
int res;
|
|
|
|
|
2022-03-17 02:54:05 +08:00
|
|
|
again:
|
2022-03-17 02:54:04 +08:00
|
|
|
if (timeout >= 0) {
|
|
|
|
tv.tv_sec = timeout / 1000;
|
|
|
|
tv.tv_usec = (timeout % 1000) * 1000;
|
|
|
|
tvp = &tv;
|
|
|
|
}
|
|
|
|
|
|
|
|
FD_ZERO(&readfds);
|
|
|
|
FD_SET(0, &readfds);
|
|
|
|
res = select(1, &readfds, NULL, NULL, tvp);
|
2022-03-17 02:54:05 +08:00
|
|
|
if (!res)
|
2022-03-17 02:54:04 +08:00
|
|
|
return EOF;
|
2022-03-17 02:54:05 +08:00
|
|
|
if (res < 0) {
|
|
|
|
if (errno == EINTR)
|
|
|
|
goto again;
|
|
|
|
else
|
|
|
|
return EOF;
|
|
|
|
}
|
2022-03-17 02:54:04 +08:00
|
|
|
return getchar();
|
|
|
|
}
|
|
|
|
|
2013-05-03 03:26:08 +08:00
|
|
|
#elif defined(GIT_WINDOWS_NATIVE)
|
2012-12-04 16:10:41 +08:00
|
|
|
|
|
|
|
#define INPUT_PATH "CONIN$"
|
|
|
|
#define OUTPUT_PATH "CONOUT$"
|
|
|
|
#define FORCE_TEXT "t"
|
|
|
|
|
2020-01-15 02:43:48 +08:00
|
|
|
static int use_stty = 1;
|
|
|
|
static struct string_list stty_restore = STRING_LIST_INIT_DUP;
|
2012-12-04 16:10:41 +08:00
|
|
|
static HANDLE hconin = INVALID_HANDLE_VALUE;
|
2021-10-05 15:46:47 +08:00
|
|
|
static HANDLE hconout = INVALID_HANDLE_VALUE;
|
|
|
|
static DWORD cmode_in, cmode_out;
|
2012-12-04 16:10:41 +08:00
|
|
|
|
2021-10-05 15:46:47 +08:00
|
|
|
void restore_term(void)
|
2012-12-04 16:10:41 +08:00
|
|
|
{
|
2020-01-15 02:43:48 +08:00
|
|
|
if (use_stty) {
|
|
|
|
int i;
|
|
|
|
struct child_process cp = CHILD_PROCESS_INIT;
|
|
|
|
|
|
|
|
if (stty_restore.nr == 0)
|
|
|
|
return;
|
|
|
|
|
2020-07-29 04:24:53 +08:00
|
|
|
strvec_push(&cp.args, "stty");
|
2020-01-15 02:43:48 +08:00
|
|
|
for (i = 0; i < stty_restore.nr; i++)
|
2020-07-29 04:24:53 +08:00
|
|
|
strvec_push(&cp.args, stty_restore.items[i].string);
|
2020-01-15 02:43:48 +08:00
|
|
|
run_command(&cp);
|
|
|
|
string_list_clear(&stty_restore, 0);
|
|
|
|
return;
|
|
|
|
}
|
|
|
|
|
2022-02-23 02:53:33 +08:00
|
|
|
sigchain_pop_common();
|
|
|
|
|
2012-12-04 16:10:41 +08:00
|
|
|
if (hconin == INVALID_HANDLE_VALUE)
|
|
|
|
return;
|
|
|
|
|
2021-10-05 15:46:47 +08:00
|
|
|
SetConsoleMode(hconin, cmode_in);
|
|
|
|
CloseHandle(hconin);
|
|
|
|
if (cmode_out) {
|
|
|
|
assert(hconout != INVALID_HANDLE_VALUE);
|
|
|
|
SetConsoleMode(hconout, cmode_out);
|
|
|
|
CloseHandle(hconout);
|
|
|
|
}
|
|
|
|
|
|
|
|
hconin = hconout = INVALID_HANDLE_VALUE;
|
|
|
|
}
|
|
|
|
|
2022-03-17 02:54:02 +08:00
|
|
|
int save_term(enum save_term_flags flags)
|
2021-10-05 15:46:47 +08:00
|
|
|
{
|
|
|
|
hconin = CreateFileA("CONIN$", GENERIC_READ | GENERIC_WRITE,
|
|
|
|
FILE_SHARE_READ, NULL, OPEN_EXISTING,
|
|
|
|
FILE_ATTRIBUTE_NORMAL, NULL);
|
|
|
|
if (hconin == INVALID_HANDLE_VALUE)
|
|
|
|
return -1;
|
|
|
|
|
2022-03-17 02:54:02 +08:00
|
|
|
if (flags & SAVE_TERM_DUPLEX) {
|
2021-10-05 15:46:47 +08:00
|
|
|
hconout = CreateFileA("CONOUT$", GENERIC_READ | GENERIC_WRITE,
|
|
|
|
FILE_SHARE_WRITE, NULL, OPEN_EXISTING,
|
|
|
|
FILE_ATTRIBUTE_NORMAL, NULL);
|
|
|
|
if (hconout == INVALID_HANDLE_VALUE)
|
|
|
|
goto error;
|
|
|
|
|
|
|
|
GetConsoleMode(hconout, &cmode_out);
|
|
|
|
}
|
|
|
|
|
|
|
|
GetConsoleMode(hconin, &cmode_in);
|
|
|
|
use_stty = 0;
|
2022-02-23 02:53:33 +08:00
|
|
|
sigchain_push_common(restore_term_on_signal);
|
2021-10-05 15:46:47 +08:00
|
|
|
return 0;
|
|
|
|
error:
|
2012-12-04 16:10:41 +08:00
|
|
|
CloseHandle(hconin);
|
|
|
|
hconin = INVALID_HANDLE_VALUE;
|
2021-10-05 15:46:47 +08:00
|
|
|
return -1;
|
2012-12-04 16:10:41 +08:00
|
|
|
}
|
|
|
|
|
2022-03-17 02:54:02 +08:00
|
|
|
static int disable_bits(enum save_term_flags flags, DWORD bits)
|
2012-12-04 16:10:41 +08:00
|
|
|
{
|
2020-01-15 02:43:48 +08:00
|
|
|
if (use_stty) {
|
|
|
|
struct child_process cp = CHILD_PROCESS_INIT;
|
|
|
|
|
2020-07-29 04:24:53 +08:00
|
|
|
strvec_push(&cp.args, "stty");
|
2020-01-15 02:43:48 +08:00
|
|
|
|
|
|
|
if (bits & ENABLE_LINE_INPUT) {
|
|
|
|
string_list_append(&stty_restore, "icanon");
|
2022-02-23 02:53:34 +08:00
|
|
|
/*
|
|
|
|
* POSIX allows VMIN and VTIME to overlap with VEOF and
|
|
|
|
* VEOL - let's hope that is not the case on windows.
|
|
|
|
*/
|
|
|
|
strvec_pushl(&cp.args, "-icanon", "min", "1", "time", "0", NULL);
|
2020-01-15 02:43:48 +08:00
|
|
|
}
|
|
|
|
|
|
|
|
if (bits & ENABLE_ECHO_INPUT) {
|
|
|
|
string_list_append(&stty_restore, "echo");
|
2020-07-29 04:24:53 +08:00
|
|
|
strvec_push(&cp.args, "-echo");
|
2020-01-15 02:43:48 +08:00
|
|
|
}
|
|
|
|
|
|
|
|
if (bits & ENABLE_PROCESSED_INPUT) {
|
|
|
|
string_list_append(&stty_restore, "-ignbrk");
|
|
|
|
string_list_append(&stty_restore, "intr");
|
|
|
|
string_list_append(&stty_restore, "^c");
|
2020-07-29 04:24:53 +08:00
|
|
|
strvec_push(&cp.args, "ignbrk");
|
|
|
|
strvec_push(&cp.args, "intr");
|
|
|
|
strvec_push(&cp.args, "");
|
2020-01-15 02:43:48 +08:00
|
|
|
}
|
|
|
|
|
|
|
|
if (run_command(&cp) == 0)
|
|
|
|
return 0;
|
|
|
|
|
|
|
|
/* `stty` could not be executed; access the Console directly */
|
|
|
|
use_stty = 0;
|
|
|
|
}
|
|
|
|
|
2022-03-17 02:54:02 +08:00
|
|
|
if (save_term(flags) < 0)
|
2012-12-04 16:10:41 +08:00
|
|
|
return -1;
|
|
|
|
|
2021-10-05 15:46:47 +08:00
|
|
|
if (!SetConsoleMode(hconin, cmode_in & ~bits)) {
|
2012-12-04 16:10:41 +08:00
|
|
|
CloseHandle(hconin);
|
|
|
|
hconin = INVALID_HANDLE_VALUE;
|
2022-02-23 02:53:33 +08:00
|
|
|
sigchain_pop_common();
|
2012-12-04 16:10:41 +08:00
|
|
|
return -1;
|
|
|
|
}
|
|
|
|
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
2022-03-17 02:54:02 +08:00
|
|
|
static int disable_echo(enum save_term_flags flags)
|
2020-01-15 02:43:47 +08:00
|
|
|
{
|
2022-03-17 02:54:02 +08:00
|
|
|
return disable_bits(flags, ENABLE_ECHO_INPUT);
|
2020-01-15 02:43:47 +08:00
|
|
|
}
|
|
|
|
|
2022-03-17 02:54:02 +08:00
|
|
|
static int enable_non_canonical(enum save_term_flags flags)
|
2020-01-15 02:43:49 +08:00
|
|
|
{
|
2022-03-17 02:54:02 +08:00
|
|
|
return disable_bits(flags,
|
|
|
|
ENABLE_ECHO_INPUT | ENABLE_LINE_INPUT | ENABLE_PROCESSED_INPUT);
|
2020-01-15 02:43:49 +08:00
|
|
|
}
|
2020-01-15 02:43:47 +08:00
|
|
|
|
built-in add -p: handle Escape sequences in interactive.singlekey mode
This recapitulates part of b5cc003253c8 (add -i: ignore terminal escape
sequences, 2011-05-17):
add -i: ignore terminal escape sequences
On the author's terminal, the up-arrow input sequence is ^[[A, and
thus fat-fingering an up-arrow into 'git checkout -p' is quite
dangerous: git-add--interactive.perl will ignore the ^[ and [
characters and happily treat A as "discard everything".
As a band-aid fix, use Term::Cap to get all terminal capabilities.
Then use the heuristic that any capability value that starts with ^[
(i.e., \e in perl) must be a key input sequence. Finally, given an
input that starts with ^[, read more characters until we have read a
full escape sequence, then return that to the caller. We use a
timeout of 0.5 seconds on the subsequent reads to avoid getting stuck
if the user actually input a lone ^[.
Since none of the currently recognized keys start with ^[, the net
result is that the sequence as a whole will be ignored and the help
displayed.
Note that we leave part for later which uses "Term::Cap to get all
terminal capabilities", for several reasons:
1. it is actually not really necessary, as the timeout of 0.5 seconds
should be plenty sufficient to catch Escape sequences,
2. it is cleaner to keep the change to special-case Escape sequences
separate from the change that reads all terminal capabilities to
speed things up, and
3. in practice, relying on the terminal capabilities is a bit overrated,
as the information could be incomplete, or plain wrong. For example,
in this developer's tmux sessions, the terminal capabilities claim
that the "cursor up" sequence is ^[M, but the actual sequence
produced by the "cursor up" key is ^[[A.
Signed-off-by: Johannes Schindelin <johannes.schindelin@gmx.de>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2020-01-15 02:43:51 +08:00
|
|
|
/*
|
|
|
|
* Override `getchar()`, as the default implementation does not use
|
|
|
|
* `ReadFile()`.
|
|
|
|
*
|
|
|
|
* This poses a problem when we want to see whether the standard
|
|
|
|
* input has more characters, as the default of Git for Windows is to start the
|
|
|
|
* Bash in a MinTTY, which uses a named pipe to emulate a pty, in which case
|
|
|
|
* our `poll()` emulation calls `PeekNamedPipe()`, which seems to require
|
|
|
|
* `ReadFile()` to be called first to work properly (it only reports 0
|
|
|
|
* available bytes, otherwise).
|
|
|
|
*
|
|
|
|
* So let's just override `getchar()` with a version backed by `ReadFile()` and
|
|
|
|
* go our merry ways from here.
|
|
|
|
*/
|
|
|
|
static int mingw_getchar(void)
|
|
|
|
{
|
|
|
|
DWORD read = 0;
|
|
|
|
unsigned char ch;
|
|
|
|
|
|
|
|
if (!ReadFile(GetStdHandle(STD_INPUT_HANDLE), &ch, 1, &read, NULL))
|
|
|
|
return EOF;
|
|
|
|
|
|
|
|
if (!read) {
|
|
|
|
error("Unexpected 0 read");
|
|
|
|
return EOF;
|
|
|
|
}
|
|
|
|
|
|
|
|
return ch;
|
|
|
|
}
|
|
|
|
#define getchar mingw_getchar
|
|
|
|
|
2022-03-17 02:54:04 +08:00
|
|
|
static int getchar_with_timeout(int timeout)
|
|
|
|
{
|
|
|
|
struct pollfd pfd = { .fd = 0, .events = POLLIN };
|
|
|
|
|
|
|
|
if (poll(&pfd, 1, timeout) < 1)
|
|
|
|
return EOF;
|
|
|
|
|
|
|
|
return getchar();
|
|
|
|
}
|
|
|
|
|
2012-12-04 16:10:41 +08:00
|
|
|
#endif
|
|
|
|
|
|
|
|
#ifndef FORCE_TEXT
|
|
|
|
#define FORCE_TEXT
|
|
|
|
#endif
|
|
|
|
|
add generic terminal prompt function
When we need to prompt the user for input interactively, we
want to access their terminal directly. We can't rely on
stdio because it may be connected to pipes or files, rather
than the terminal. Instead, we use "getpass()", because it
abstracts the idea of prompting and reading from the
terminal. However, it has some problems:
1. It never echoes the typed characters, which makes it OK
for passwords but annoying for other input (like usernames).
2. Some implementations of getpass() have an extremely
small input buffer (e.g., Solaris 8 is reported to
support only 8 characters).
3. Some implementations of getpass() will fall back to
reading from stdin (e.g., glibc). We explicitly don't
want this, because our stdin may be connected to a pipe
speaking a particular protocol, and reading will
disrupt the protocol flow (e.g., the remote-curl
helper).
4. Some implementations of getpass() turn off signals, so
that hitting "^C" on the terminal does not break out of
the password prompt. This can be a mild annoyance.
Instead, let's provide an abstract "git_terminal_prompt"
function that addresses these concerns. This patch includes
an implementation based on /dev/tty, enabled by setting
HAVE_DEV_TTY. The fallback is to use getpass() as before.
Signed-off-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2011-12-10 18:41:01 +08:00
|
|
|
char *git_terminal_prompt(const char *prompt, int echo)
|
|
|
|
{
|
|
|
|
static struct strbuf buf = STRBUF_INIT;
|
|
|
|
int r;
|
2012-12-04 16:10:40 +08:00
|
|
|
FILE *input_fh, *output_fh;
|
add generic terminal prompt function
When we need to prompt the user for input interactively, we
want to access their terminal directly. We can't rely on
stdio because it may be connected to pipes or files, rather
than the terminal. Instead, we use "getpass()", because it
abstracts the idea of prompting and reading from the
terminal. However, it has some problems:
1. It never echoes the typed characters, which makes it OK
for passwords but annoying for other input (like usernames).
2. Some implementations of getpass() have an extremely
small input buffer (e.g., Solaris 8 is reported to
support only 8 characters).
3. Some implementations of getpass() will fall back to
reading from stdin (e.g., glibc). We explicitly don't
want this, because our stdin may be connected to a pipe
speaking a particular protocol, and reading will
disrupt the protocol flow (e.g., the remote-curl
helper).
4. Some implementations of getpass() turn off signals, so
that hitting "^C" on the terminal does not break out of
the password prompt. This can be a mild annoyance.
Instead, let's provide an abstract "git_terminal_prompt"
function that addresses these concerns. This patch includes
an implementation based on /dev/tty, enabled by setting
HAVE_DEV_TTY. The fallback is to use getpass() as before.
Signed-off-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2011-12-10 18:41:01 +08:00
|
|
|
|
2012-12-04 16:10:41 +08:00
|
|
|
input_fh = fopen(INPUT_PATH, "r" FORCE_TEXT);
|
2012-12-04 16:10:40 +08:00
|
|
|
if (!input_fh)
|
add generic terminal prompt function
When we need to prompt the user for input interactively, we
want to access their terminal directly. We can't rely on
stdio because it may be connected to pipes or files, rather
than the terminal. Instead, we use "getpass()", because it
abstracts the idea of prompting and reading from the
terminal. However, it has some problems:
1. It never echoes the typed characters, which makes it OK
for passwords but annoying for other input (like usernames).
2. Some implementations of getpass() have an extremely
small input buffer (e.g., Solaris 8 is reported to
support only 8 characters).
3. Some implementations of getpass() will fall back to
reading from stdin (e.g., glibc). We explicitly don't
want this, because our stdin may be connected to a pipe
speaking a particular protocol, and reading will
disrupt the protocol flow (e.g., the remote-curl
helper).
4. Some implementations of getpass() turn off signals, so
that hitting "^C" on the terminal does not break out of
the password prompt. This can be a mild annoyance.
Instead, let's provide an abstract "git_terminal_prompt"
function that addresses these concerns. This patch includes
an implementation based on /dev/tty, enabled by setting
HAVE_DEV_TTY. The fallback is to use getpass() as before.
Signed-off-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2011-12-10 18:41:01 +08:00
|
|
|
return NULL;
|
|
|
|
|
2012-12-04 16:10:41 +08:00
|
|
|
output_fh = fopen(OUTPUT_PATH, "w" FORCE_TEXT);
|
2012-12-04 16:10:40 +08:00
|
|
|
if (!output_fh) {
|
|
|
|
fclose(input_fh);
|
|
|
|
return NULL;
|
|
|
|
}
|
|
|
|
|
2022-03-17 02:54:02 +08:00
|
|
|
if (!echo && disable_echo(0)) {
|
2012-12-04 16:10:40 +08:00
|
|
|
fclose(input_fh);
|
|
|
|
fclose(output_fh);
|
2012-12-04 16:10:39 +08:00
|
|
|
return NULL;
|
add generic terminal prompt function
When we need to prompt the user for input interactively, we
want to access their terminal directly. We can't rely on
stdio because it may be connected to pipes or files, rather
than the terminal. Instead, we use "getpass()", because it
abstracts the idea of prompting and reading from the
terminal. However, it has some problems:
1. It never echoes the typed characters, which makes it OK
for passwords but annoying for other input (like usernames).
2. Some implementations of getpass() have an extremely
small input buffer (e.g., Solaris 8 is reported to
support only 8 characters).
3. Some implementations of getpass() will fall back to
reading from stdin (e.g., glibc). We explicitly don't
want this, because our stdin may be connected to a pipe
speaking a particular protocol, and reading will
disrupt the protocol flow (e.g., the remote-curl
helper).
4. Some implementations of getpass() turn off signals, so
that hitting "^C" on the terminal does not break out of
the password prompt. This can be a mild annoyance.
Instead, let's provide an abstract "git_terminal_prompt"
function that addresses these concerns. This patch includes
an implementation based on /dev/tty, enabled by setting
HAVE_DEV_TTY. The fallback is to use getpass() as before.
Signed-off-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2011-12-10 18:41:01 +08:00
|
|
|
}
|
|
|
|
|
2012-12-04 16:10:40 +08:00
|
|
|
fputs(prompt, output_fh);
|
|
|
|
fflush(output_fh);
|
add generic terminal prompt function
When we need to prompt the user for input interactively, we
want to access their terminal directly. We can't rely on
stdio because it may be connected to pipes or files, rather
than the terminal. Instead, we use "getpass()", because it
abstracts the idea of prompting and reading from the
terminal. However, it has some problems:
1. It never echoes the typed characters, which makes it OK
for passwords but annoying for other input (like usernames).
2. Some implementations of getpass() have an extremely
small input buffer (e.g., Solaris 8 is reported to
support only 8 characters).
3. Some implementations of getpass() will fall back to
reading from stdin (e.g., glibc). We explicitly don't
want this, because our stdin may be connected to a pipe
speaking a particular protocol, and reading will
disrupt the protocol flow (e.g., the remote-curl
helper).
4. Some implementations of getpass() turn off signals, so
that hitting "^C" on the terminal does not break out of
the password prompt. This can be a mild annoyance.
Instead, let's provide an abstract "git_terminal_prompt"
function that addresses these concerns. This patch includes
an implementation based on /dev/tty, enabled by setting
HAVE_DEV_TTY. The fallback is to use getpass() as before.
Signed-off-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2011-12-10 18:41:01 +08:00
|
|
|
|
2016-01-14 07:31:17 +08:00
|
|
|
r = strbuf_getline_lf(&buf, input_fh);
|
add generic terminal prompt function
When we need to prompt the user for input interactively, we
want to access their terminal directly. We can't rely on
stdio because it may be connected to pipes or files, rather
than the terminal. Instead, we use "getpass()", because it
abstracts the idea of prompting and reading from the
terminal. However, it has some problems:
1. It never echoes the typed characters, which makes it OK
for passwords but annoying for other input (like usernames).
2. Some implementations of getpass() have an extremely
small input buffer (e.g., Solaris 8 is reported to
support only 8 characters).
3. Some implementations of getpass() will fall back to
reading from stdin (e.g., glibc). We explicitly don't
want this, because our stdin may be connected to a pipe
speaking a particular protocol, and reading will
disrupt the protocol flow (e.g., the remote-curl
helper).
4. Some implementations of getpass() turn off signals, so
that hitting "^C" on the terminal does not break out of
the password prompt. This can be a mild annoyance.
Instead, let's provide an abstract "git_terminal_prompt"
function that addresses these concerns. This patch includes
an implementation based on /dev/tty, enabled by setting
HAVE_DEV_TTY. The fallback is to use getpass() as before.
Signed-off-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2011-12-10 18:41:01 +08:00
|
|
|
if (!echo) {
|
2012-12-04 16:10:40 +08:00
|
|
|
putc('\n', output_fh);
|
|
|
|
fflush(output_fh);
|
add generic terminal prompt function
When we need to prompt the user for input interactively, we
want to access their terminal directly. We can't rely on
stdio because it may be connected to pipes or files, rather
than the terminal. Instead, we use "getpass()", because it
abstracts the idea of prompting and reading from the
terminal. However, it has some problems:
1. It never echoes the typed characters, which makes it OK
for passwords but annoying for other input (like usernames).
2. Some implementations of getpass() have an extremely
small input buffer (e.g., Solaris 8 is reported to
support only 8 characters).
3. Some implementations of getpass() will fall back to
reading from stdin (e.g., glibc). We explicitly don't
want this, because our stdin may be connected to a pipe
speaking a particular protocol, and reading will
disrupt the protocol flow (e.g., the remote-curl
helper).
4. Some implementations of getpass() turn off signals, so
that hitting "^C" on the terminal does not break out of
the password prompt. This can be a mild annoyance.
Instead, let's provide an abstract "git_terminal_prompt"
function that addresses these concerns. This patch includes
an implementation based on /dev/tty, enabled by setting
HAVE_DEV_TTY. The fallback is to use getpass() as before.
Signed-off-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2011-12-10 18:41:01 +08:00
|
|
|
}
|
|
|
|
|
|
|
|
restore_term();
|
2012-12-04 16:10:40 +08:00
|
|
|
fclose(input_fh);
|
|
|
|
fclose(output_fh);
|
add generic terminal prompt function
When we need to prompt the user for input interactively, we
want to access their terminal directly. We can't rely on
stdio because it may be connected to pipes or files, rather
than the terminal. Instead, we use "getpass()", because it
abstracts the idea of prompting and reading from the
terminal. However, it has some problems:
1. It never echoes the typed characters, which makes it OK
for passwords but annoying for other input (like usernames).
2. Some implementations of getpass() have an extremely
small input buffer (e.g., Solaris 8 is reported to
support only 8 characters).
3. Some implementations of getpass() will fall back to
reading from stdin (e.g., glibc). We explicitly don't
want this, because our stdin may be connected to a pipe
speaking a particular protocol, and reading will
disrupt the protocol flow (e.g., the remote-curl
helper).
4. Some implementations of getpass() turn off signals, so
that hitting "^C" on the terminal does not break out of
the password prompt. This can be a mild annoyance.
Instead, let's provide an abstract "git_terminal_prompt"
function that addresses these concerns. This patch includes
an implementation based on /dev/tty, enabled by setting
HAVE_DEV_TTY. The fallback is to use getpass() as before.
Signed-off-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2011-12-10 18:41:01 +08:00
|
|
|
|
|
|
|
if (r == EOF)
|
|
|
|
return NULL;
|
|
|
|
return buf.buf;
|
|
|
|
}
|
|
|
|
|
built-in add -p: handle Escape sequences more efficiently
When `interactive.singlekey = true`, we react immediately to keystrokes,
even to Escape sequences (e.g. when pressing a cursor key).
The problem with Escape sequences is that we do not really know when
they are done, and as a heuristic we poll standard input for half a
second to make sure that we got all of it.
While waiting half a second is not asking for a whole lot, it can become
quite annoying over time, therefore with this patch, we read the
terminal capabilities (if available) and extract known Escape sequences
from there, then stop polling immediately when we detected that the user
pressed a key that generated such a known sequence.
This recapitulates the remaining part of b5cc003253c8 (add -i: ignore
terminal escape sequences, 2011-05-17).
Note: We do *not* query the terminal capabilities directly. That would
either require a lot of platform-specific code, or it would require
linking to a library such as ncurses.
Linking to a library in the built-ins is something we try very hard to
avoid (we even kicked the libcurl dependency to a non-built-in remote
helper, just to shave off a tiny fraction of a second from Git's startup
time). And the platform-specific code would be a maintenance nightmare.
Even worse: in Git for Windows' case, we would need to query MSYS2
pseudo terminals, which `git.exe` simply cannot do (because it is
intentionally *not* an MSYS2 program).
To address this, we simply spawn `infocmp -L -1` and parse its output
(which works even in Git for Windows, because that helper is included in
the end-user facing installations).
This is done only once, as in the Perl version, but it is done only when
the first Escape sequence is encountered, not upon startup of `git add
-i`; This saves on startup time, yet makes reacting to the first Escape
sequence slightly more sluggish. But it allows us to keep the
terminal-related code encapsulated in the `compat/terminal.c` file.
Signed-off-by: Johannes Schindelin <johannes.schindelin@gmx.de>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2020-01-15 02:43:52 +08:00
|
|
|
/*
|
|
|
|
* The `is_known_escape_sequence()` function returns 1 if the passed string
|
|
|
|
* corresponds to an Escape sequence that the terminal capabilities contains.
|
|
|
|
*
|
|
|
|
* To avoid depending on ncurses or other platform-specific libraries, we rely
|
|
|
|
* on the presence of the `infocmp` executable to do the job for us (failing
|
|
|
|
* silently if the program is not available or refused to run).
|
|
|
|
*/
|
|
|
|
struct escape_sequence_entry {
|
|
|
|
struct hashmap_entry entry;
|
|
|
|
char sequence[FLEX_ARRAY];
|
|
|
|
};
|
|
|
|
|
2022-08-26 01:09:48 +08:00
|
|
|
static int sequence_entry_cmp(const void *hashmap_cmp_fn_data UNUSED,
|
hashmap: use expected signatures for comparison functions
We prefer for callback functions to match the signature with which
they'll be called, rather than casting them to the correct type when
assigning function pointers. Even though casting often works in the real
world, it is a violation of the standard.
We did a mass conversion in 939af16eac (hashmap_cmp_fn takes
hashmap_entry params, 2019-10-06), but have grown a few new cases since
then. Because of the cast, the compiler does not complain. However, as
of clang-18, UBSan will catch these at run-time, and the case in
range-diff.c triggers when running t3206.
After seeing that one, I scanned the results of:
git grep '_fn)[^(]' '*.c' | grep -v typedef
and found a similar case in compat/terminal.c (which presumably isn't
called in the test suite, since it doesn't trigger UBSan). There might
be other cases lurking if the cast is done using a typedef that doesn't
end in "_fn", but loosening it finds too many false positives. I also
looked for:
git grep ' = ([a-z_]*) *[a-z]' '*.c'
to find assignments that cast, but nothing looked like a function.
The resulting code is unfortunately a little longer, but the bonus of
using container_of() is that we are no longer restricted to the
hashmap_entry being at the start of the struct.
Signed-off-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2023-08-20 07:55:30 +08:00
|
|
|
const struct hashmap_entry *he1,
|
|
|
|
const struct hashmap_entry *he2,
|
built-in add -p: handle Escape sequences more efficiently
When `interactive.singlekey = true`, we react immediately to keystrokes,
even to Escape sequences (e.g. when pressing a cursor key).
The problem with Escape sequences is that we do not really know when
they are done, and as a heuristic we poll standard input for half a
second to make sure that we got all of it.
While waiting half a second is not asking for a whole lot, it can become
quite annoying over time, therefore with this patch, we read the
terminal capabilities (if available) and extract known Escape sequences
from there, then stop polling immediately when we detected that the user
pressed a key that generated such a known sequence.
This recapitulates the remaining part of b5cc003253c8 (add -i: ignore
terminal escape sequences, 2011-05-17).
Note: We do *not* query the terminal capabilities directly. That would
either require a lot of platform-specific code, or it would require
linking to a library such as ncurses.
Linking to a library in the built-ins is something we try very hard to
avoid (we even kicked the libcurl dependency to a non-built-in remote
helper, just to shave off a tiny fraction of a second from Git's startup
time). And the platform-specific code would be a maintenance nightmare.
Even worse: in Git for Windows' case, we would need to query MSYS2
pseudo terminals, which `git.exe` simply cannot do (because it is
intentionally *not* an MSYS2 program).
To address this, we simply spawn `infocmp -L -1` and parse its output
(which works even in Git for Windows, because that helper is included in
the end-user facing installations).
This is done only once, as in the Perl version, but it is done only when
the first Escape sequence is encountered, not upon startup of `git add
-i`; This saves on startup time, yet makes reacting to the first Escape
sequence slightly more sluggish. But it allows us to keep the
terminal-related code encapsulated in the `compat/terminal.c` file.
Signed-off-by: Johannes Schindelin <johannes.schindelin@gmx.de>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2020-01-15 02:43:52 +08:00
|
|
|
const void *keydata)
|
|
|
|
{
|
hashmap: use expected signatures for comparison functions
We prefer for callback functions to match the signature with which
they'll be called, rather than casting them to the correct type when
assigning function pointers. Even though casting often works in the real
world, it is a violation of the standard.
We did a mass conversion in 939af16eac (hashmap_cmp_fn takes
hashmap_entry params, 2019-10-06), but have grown a few new cases since
then. Because of the cast, the compiler does not complain. However, as
of clang-18, UBSan will catch these at run-time, and the case in
range-diff.c triggers when running t3206.
After seeing that one, I scanned the results of:
git grep '_fn)[^(]' '*.c' | grep -v typedef
and found a similar case in compat/terminal.c (which presumably isn't
called in the test suite, since it doesn't trigger UBSan). There might
be other cases lurking if the cast is done using a typedef that doesn't
end in "_fn", but loosening it finds too many false positives. I also
looked for:
git grep ' = ([a-z_]*) *[a-z]' '*.c'
to find assignments that cast, but nothing looked like a function.
The resulting code is unfortunately a little longer, but the bonus of
using container_of() is that we are no longer restricted to the
hashmap_entry being at the start of the struct.
Signed-off-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2023-08-20 07:55:30 +08:00
|
|
|
const struct escape_sequence_entry
|
|
|
|
*e1 = container_of(he1, const struct escape_sequence_entry, entry),
|
|
|
|
*e2 = container_of(he2, const struct escape_sequence_entry, entry);
|
built-in add -p: handle Escape sequences more efficiently
When `interactive.singlekey = true`, we react immediately to keystrokes,
even to Escape sequences (e.g. when pressing a cursor key).
The problem with Escape sequences is that we do not really know when
they are done, and as a heuristic we poll standard input for half a
second to make sure that we got all of it.
While waiting half a second is not asking for a whole lot, it can become
quite annoying over time, therefore with this patch, we read the
terminal capabilities (if available) and extract known Escape sequences
from there, then stop polling immediately when we detected that the user
pressed a key that generated such a known sequence.
This recapitulates the remaining part of b5cc003253c8 (add -i: ignore
terminal escape sequences, 2011-05-17).
Note: We do *not* query the terminal capabilities directly. That would
either require a lot of platform-specific code, or it would require
linking to a library such as ncurses.
Linking to a library in the built-ins is something we try very hard to
avoid (we even kicked the libcurl dependency to a non-built-in remote
helper, just to shave off a tiny fraction of a second from Git's startup
time). And the platform-specific code would be a maintenance nightmare.
Even worse: in Git for Windows' case, we would need to query MSYS2
pseudo terminals, which `git.exe` simply cannot do (because it is
intentionally *not* an MSYS2 program).
To address this, we simply spawn `infocmp -L -1` and parse its output
(which works even in Git for Windows, because that helper is included in
the end-user facing installations).
This is done only once, as in the Perl version, but it is done only when
the first Escape sequence is encountered, not upon startup of `git add
-i`; This saves on startup time, yet makes reacting to the first Escape
sequence slightly more sluggish. But it allows us to keep the
terminal-related code encapsulated in the `compat/terminal.c` file.
Signed-off-by: Johannes Schindelin <johannes.schindelin@gmx.de>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2020-01-15 02:43:52 +08:00
|
|
|
return strcmp(e1->sequence, keydata ? keydata : e2->sequence);
|
|
|
|
}
|
|
|
|
|
|
|
|
static int is_known_escape_sequence(const char *sequence)
|
|
|
|
{
|
|
|
|
static struct hashmap sequences;
|
|
|
|
static int initialized;
|
|
|
|
|
|
|
|
if (!initialized) {
|
|
|
|
struct child_process cp = CHILD_PROCESS_INIT;
|
|
|
|
struct strbuf buf = STRBUF_INIT;
|
|
|
|
char *p, *eol;
|
|
|
|
|
hashmap: use expected signatures for comparison functions
We prefer for callback functions to match the signature with which
they'll be called, rather than casting them to the correct type when
assigning function pointers. Even though casting often works in the real
world, it is a violation of the standard.
We did a mass conversion in 939af16eac (hashmap_cmp_fn takes
hashmap_entry params, 2019-10-06), but have grown a few new cases since
then. Because of the cast, the compiler does not complain. However, as
of clang-18, UBSan will catch these at run-time, and the case in
range-diff.c triggers when running t3206.
After seeing that one, I scanned the results of:
git grep '_fn)[^(]' '*.c' | grep -v typedef
and found a similar case in compat/terminal.c (which presumably isn't
called in the test suite, since it doesn't trigger UBSan). There might
be other cases lurking if the cast is done using a typedef that doesn't
end in "_fn", but loosening it finds too many false positives. I also
looked for:
git grep ' = ([a-z_]*) *[a-z]' '*.c'
to find assignments that cast, but nothing looked like a function.
The resulting code is unfortunately a little longer, but the bonus of
using container_of() is that we are no longer restricted to the
hashmap_entry being at the start of the struct.
Signed-off-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2023-08-20 07:55:30 +08:00
|
|
|
hashmap_init(&sequences, sequence_entry_cmp, NULL, 0);
|
built-in add -p: handle Escape sequences more efficiently
When `interactive.singlekey = true`, we react immediately to keystrokes,
even to Escape sequences (e.g. when pressing a cursor key).
The problem with Escape sequences is that we do not really know when
they are done, and as a heuristic we poll standard input for half a
second to make sure that we got all of it.
While waiting half a second is not asking for a whole lot, it can become
quite annoying over time, therefore with this patch, we read the
terminal capabilities (if available) and extract known Escape sequences
from there, then stop polling immediately when we detected that the user
pressed a key that generated such a known sequence.
This recapitulates the remaining part of b5cc003253c8 (add -i: ignore
terminal escape sequences, 2011-05-17).
Note: We do *not* query the terminal capabilities directly. That would
either require a lot of platform-specific code, or it would require
linking to a library such as ncurses.
Linking to a library in the built-ins is something we try very hard to
avoid (we even kicked the libcurl dependency to a non-built-in remote
helper, just to shave off a tiny fraction of a second from Git's startup
time). And the platform-specific code would be a maintenance nightmare.
Even worse: in Git for Windows' case, we would need to query MSYS2
pseudo terminals, which `git.exe` simply cannot do (because it is
intentionally *not* an MSYS2 program).
To address this, we simply spawn `infocmp -L -1` and parse its output
(which works even in Git for Windows, because that helper is included in
the end-user facing installations).
This is done only once, as in the Perl version, but it is done only when
the first Escape sequence is encountered, not upon startup of `git add
-i`; This saves on startup time, yet makes reacting to the first Escape
sequence slightly more sluggish. But it allows us to keep the
terminal-related code encapsulated in the `compat/terminal.c` file.
Signed-off-by: Johannes Schindelin <johannes.schindelin@gmx.de>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2020-01-15 02:43:52 +08:00
|
|
|
|
2020-07-29 04:24:53 +08:00
|
|
|
strvec_pushl(&cp.args, "infocmp", "-L", "-1", NULL);
|
built-in add -p: handle Escape sequences more efficiently
When `interactive.singlekey = true`, we react immediately to keystrokes,
even to Escape sequences (e.g. when pressing a cursor key).
The problem with Escape sequences is that we do not really know when
they are done, and as a heuristic we poll standard input for half a
second to make sure that we got all of it.
While waiting half a second is not asking for a whole lot, it can become
quite annoying over time, therefore with this patch, we read the
terminal capabilities (if available) and extract known Escape sequences
from there, then stop polling immediately when we detected that the user
pressed a key that generated such a known sequence.
This recapitulates the remaining part of b5cc003253c8 (add -i: ignore
terminal escape sequences, 2011-05-17).
Note: We do *not* query the terminal capabilities directly. That would
either require a lot of platform-specific code, or it would require
linking to a library such as ncurses.
Linking to a library in the built-ins is something we try very hard to
avoid (we even kicked the libcurl dependency to a non-built-in remote
helper, just to shave off a tiny fraction of a second from Git's startup
time). And the platform-specific code would be a maintenance nightmare.
Even worse: in Git for Windows' case, we would need to query MSYS2
pseudo terminals, which `git.exe` simply cannot do (because it is
intentionally *not* an MSYS2 program).
To address this, we simply spawn `infocmp -L -1` and parse its output
(which works even in Git for Windows, because that helper is included in
the end-user facing installations).
This is done only once, as in the Perl version, but it is done only when
the first Escape sequence is encountered, not upon startup of `git add
-i`; This saves on startup time, yet makes reacting to the first Escape
sequence slightly more sluggish. But it allows us to keep the
terminal-related code encapsulated in the `compat/terminal.c` file.
Signed-off-by: Johannes Schindelin <johannes.schindelin@gmx.de>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2020-01-15 02:43:52 +08:00
|
|
|
if (pipe_command(&cp, NULL, 0, &buf, 0, NULL, 0))
|
|
|
|
strbuf_setlen(&buf, 0);
|
|
|
|
|
|
|
|
for (eol = p = buf.buf; *p; p = eol + 1) {
|
|
|
|
p = strchr(p, '=');
|
|
|
|
if (!p)
|
|
|
|
break;
|
|
|
|
p++;
|
|
|
|
eol = strchrnul(p, '\n');
|
|
|
|
|
|
|
|
if (starts_with(p, "\\E")) {
|
|
|
|
char *comma = memchr(p, ',', eol - p);
|
|
|
|
struct escape_sequence_entry *e;
|
|
|
|
|
|
|
|
p[0] = '^';
|
|
|
|
p[1] = '[';
|
|
|
|
FLEX_ALLOC_MEM(e, sequence, p, comma - p);
|
|
|
|
hashmap_entry_init(&e->entry,
|
|
|
|
strhash(e->sequence));
|
|
|
|
hashmap_add(&sequences, &e->entry);
|
|
|
|
}
|
|
|
|
if (!*eol)
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
initialized = 1;
|
|
|
|
}
|
|
|
|
|
|
|
|
return !!hashmap_get_from_hash(&sequences, strhash(sequence), sequence);
|
|
|
|
}
|
|
|
|
|
2020-01-15 02:43:49 +08:00
|
|
|
int read_key_without_echo(struct strbuf *buf)
|
|
|
|
{
|
|
|
|
static int warning_displayed;
|
|
|
|
int ch;
|
|
|
|
|
2022-03-17 02:54:03 +08:00
|
|
|
if (warning_displayed || enable_non_canonical(SAVE_TERM_STDIN) < 0) {
|
2020-01-15 02:43:49 +08:00
|
|
|
if (!warning_displayed) {
|
|
|
|
warning("reading single keystrokes not supported on "
|
|
|
|
"this platform; reading line instead");
|
|
|
|
warning_displayed = 1;
|
|
|
|
}
|
|
|
|
|
|
|
|
return strbuf_getline(buf, stdin);
|
|
|
|
}
|
|
|
|
|
|
|
|
strbuf_reset(buf);
|
|
|
|
ch = getchar();
|
|
|
|
if (ch == EOF) {
|
|
|
|
restore_term();
|
|
|
|
return EOF;
|
|
|
|
}
|
|
|
|
strbuf_addch(buf, ch);
|
built-in add -p: handle Escape sequences in interactive.singlekey mode
This recapitulates part of b5cc003253c8 (add -i: ignore terminal escape
sequences, 2011-05-17):
add -i: ignore terminal escape sequences
On the author's terminal, the up-arrow input sequence is ^[[A, and
thus fat-fingering an up-arrow into 'git checkout -p' is quite
dangerous: git-add--interactive.perl will ignore the ^[ and [
characters and happily treat A as "discard everything".
As a band-aid fix, use Term::Cap to get all terminal capabilities.
Then use the heuristic that any capability value that starts with ^[
(i.e., \e in perl) must be a key input sequence. Finally, given an
input that starts with ^[, read more characters until we have read a
full escape sequence, then return that to the caller. We use a
timeout of 0.5 seconds on the subsequent reads to avoid getting stuck
if the user actually input a lone ^[.
Since none of the currently recognized keys start with ^[, the net
result is that the sequence as a whole will be ignored and the help
displayed.
Note that we leave part for later which uses "Term::Cap to get all
terminal capabilities", for several reasons:
1. it is actually not really necessary, as the timeout of 0.5 seconds
should be plenty sufficient to catch Escape sequences,
2. it is cleaner to keep the change to special-case Escape sequences
separate from the change that reads all terminal capabilities to
speed things up, and
3. in practice, relying on the terminal capabilities is a bit overrated,
as the information could be incomplete, or plain wrong. For example,
in this developer's tmux sessions, the terminal capabilities claim
that the "cursor up" sequence is ^[M, but the actual sequence
produced by the "cursor up" key is ^[[A.
Signed-off-by: Johannes Schindelin <johannes.schindelin@gmx.de>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2020-01-15 02:43:51 +08:00
|
|
|
|
|
|
|
if (ch == '\033' /* ESC */) {
|
|
|
|
/*
|
|
|
|
* We are most likely looking at an Escape sequence. Let's try
|
|
|
|
* to read more bytes, waiting at most half a second, assuming
|
|
|
|
* that the sequence is complete if we did not receive any byte
|
|
|
|
* within that time.
|
|
|
|
*
|
|
|
|
* Start by replacing the Escape byte with ^[ */
|
|
|
|
strbuf_splice(buf, buf->len - 1, 1, "^[", 2);
|
|
|
|
|
built-in add -p: handle Escape sequences more efficiently
When `interactive.singlekey = true`, we react immediately to keystrokes,
even to Escape sequences (e.g. when pressing a cursor key).
The problem with Escape sequences is that we do not really know when
they are done, and as a heuristic we poll standard input for half a
second to make sure that we got all of it.
While waiting half a second is not asking for a whole lot, it can become
quite annoying over time, therefore with this patch, we read the
terminal capabilities (if available) and extract known Escape sequences
from there, then stop polling immediately when we detected that the user
pressed a key that generated such a known sequence.
This recapitulates the remaining part of b5cc003253c8 (add -i: ignore
terminal escape sequences, 2011-05-17).
Note: We do *not* query the terminal capabilities directly. That would
either require a lot of platform-specific code, or it would require
linking to a library such as ncurses.
Linking to a library in the built-ins is something we try very hard to
avoid (we even kicked the libcurl dependency to a non-built-in remote
helper, just to shave off a tiny fraction of a second from Git's startup
time). And the platform-specific code would be a maintenance nightmare.
Even worse: in Git for Windows' case, we would need to query MSYS2
pseudo terminals, which `git.exe` simply cannot do (because it is
intentionally *not* an MSYS2 program).
To address this, we simply spawn `infocmp -L -1` and parse its output
(which works even in Git for Windows, because that helper is included in
the end-user facing installations).
This is done only once, as in the Perl version, but it is done only when
the first Escape sequence is encountered, not upon startup of `git add
-i`; This saves on startup time, yet makes reacting to the first Escape
sequence slightly more sluggish. But it allows us to keep the
terminal-related code encapsulated in the `compat/terminal.c` file.
Signed-off-by: Johannes Schindelin <johannes.schindelin@gmx.de>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2020-01-15 02:43:52 +08:00
|
|
|
/*
|
|
|
|
* Query the terminal capabilities once about all the Escape
|
|
|
|
* sequences it knows about, so that we can avoid waiting for
|
|
|
|
* half a second when we know that the sequence is complete.
|
|
|
|
*/
|
|
|
|
while (!is_known_escape_sequence(buf->buf)) {
|
2022-03-17 02:54:04 +08:00
|
|
|
ch = getchar_with_timeout(500);
|
built-in add -p: handle Escape sequences in interactive.singlekey mode
This recapitulates part of b5cc003253c8 (add -i: ignore terminal escape
sequences, 2011-05-17):
add -i: ignore terminal escape sequences
On the author's terminal, the up-arrow input sequence is ^[[A, and
thus fat-fingering an up-arrow into 'git checkout -p' is quite
dangerous: git-add--interactive.perl will ignore the ^[ and [
characters and happily treat A as "discard everything".
As a band-aid fix, use Term::Cap to get all terminal capabilities.
Then use the heuristic that any capability value that starts with ^[
(i.e., \e in perl) must be a key input sequence. Finally, given an
input that starts with ^[, read more characters until we have read a
full escape sequence, then return that to the caller. We use a
timeout of 0.5 seconds on the subsequent reads to avoid getting stuck
if the user actually input a lone ^[.
Since none of the currently recognized keys start with ^[, the net
result is that the sequence as a whole will be ignored and the help
displayed.
Note that we leave part for later which uses "Term::Cap to get all
terminal capabilities", for several reasons:
1. it is actually not really necessary, as the timeout of 0.5 seconds
should be plenty sufficient to catch Escape sequences,
2. it is cleaner to keep the change to special-case Escape sequences
separate from the change that reads all terminal capabilities to
speed things up, and
3. in practice, relying on the terminal capabilities is a bit overrated,
as the information could be incomplete, or plain wrong. For example,
in this developer's tmux sessions, the terminal capabilities claim
that the "cursor up" sequence is ^[M, but the actual sequence
produced by the "cursor up" key is ^[[A.
Signed-off-by: Johannes Schindelin <johannes.schindelin@gmx.de>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2020-01-15 02:43:51 +08:00
|
|
|
if (ch == EOF)
|
2022-02-23 02:53:32 +08:00
|
|
|
break;
|
built-in add -p: handle Escape sequences in interactive.singlekey mode
This recapitulates part of b5cc003253c8 (add -i: ignore terminal escape
sequences, 2011-05-17):
add -i: ignore terminal escape sequences
On the author's terminal, the up-arrow input sequence is ^[[A, and
thus fat-fingering an up-arrow into 'git checkout -p' is quite
dangerous: git-add--interactive.perl will ignore the ^[ and [
characters and happily treat A as "discard everything".
As a band-aid fix, use Term::Cap to get all terminal capabilities.
Then use the heuristic that any capability value that starts with ^[
(i.e., \e in perl) must be a key input sequence. Finally, given an
input that starts with ^[, read more characters until we have read a
full escape sequence, then return that to the caller. We use a
timeout of 0.5 seconds on the subsequent reads to avoid getting stuck
if the user actually input a lone ^[.
Since none of the currently recognized keys start with ^[, the net
result is that the sequence as a whole will be ignored and the help
displayed.
Note that we leave part for later which uses "Term::Cap to get all
terminal capabilities", for several reasons:
1. it is actually not really necessary, as the timeout of 0.5 seconds
should be plenty sufficient to catch Escape sequences,
2. it is cleaner to keep the change to special-case Escape sequences
separate from the change that reads all terminal capabilities to
speed things up, and
3. in practice, relying on the terminal capabilities is a bit overrated,
as the information could be incomplete, or plain wrong. For example,
in this developer's tmux sessions, the terminal capabilities claim
that the "cursor up" sequence is ^[M, but the actual sequence
produced by the "cursor up" key is ^[[A.
Signed-off-by: Johannes Schindelin <johannes.schindelin@gmx.de>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2020-01-15 02:43:51 +08:00
|
|
|
strbuf_addch(buf, ch);
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2020-01-15 02:43:49 +08:00
|
|
|
restore_term();
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
add generic terminal prompt function
When we need to prompt the user for input interactively, we
want to access their terminal directly. We can't rely on
stdio because it may be connected to pipes or files, rather
than the terminal. Instead, we use "getpass()", because it
abstracts the idea of prompting and reading from the
terminal. However, it has some problems:
1. It never echoes the typed characters, which makes it OK
for passwords but annoying for other input (like usernames).
2. Some implementations of getpass() have an extremely
small input buffer (e.g., Solaris 8 is reported to
support only 8 characters).
3. Some implementations of getpass() will fall back to
reading from stdin (e.g., glibc). We explicitly don't
want this, because our stdin may be connected to a pipe
speaking a particular protocol, and reading will
disrupt the protocol flow (e.g., the remote-curl
helper).
4. Some implementations of getpass() turn off signals, so
that hitting "^C" on the terminal does not break out of
the password prompt. This can be a mild annoyance.
Instead, let's provide an abstract "git_terminal_prompt"
function that addresses these concerns. This patch includes
an implementation based on /dev/tty, enabled by setting
HAVE_DEV_TTY. The fallback is to use getpass() as before.
Signed-off-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2011-12-10 18:41:01 +08:00
|
|
|
#else
|
|
|
|
|
2022-03-17 02:54:02 +08:00
|
|
|
int save_term(enum save_term_flags flags)
|
2021-10-05 15:46:47 +08:00
|
|
|
{
|
2022-03-17 02:54:02 +08:00
|
|
|
/* no duplex support available */
|
|
|
|
return -!!(flags & SAVE_TERM_DUPLEX);
|
2021-10-05 15:46:47 +08:00
|
|
|
}
|
|
|
|
|
|
|
|
void restore_term(void)
|
|
|
|
{
|
|
|
|
}
|
|
|
|
|
2024-08-31 22:58:56 +08:00
|
|
|
char *git_terminal_prompt(const char *prompt, int echo UNUSED)
|
add generic terminal prompt function
When we need to prompt the user for input interactively, we
want to access their terminal directly. We can't rely on
stdio because it may be connected to pipes or files, rather
than the terminal. Instead, we use "getpass()", because it
abstracts the idea of prompting and reading from the
terminal. However, it has some problems:
1. It never echoes the typed characters, which makes it OK
for passwords but annoying for other input (like usernames).
2. Some implementations of getpass() have an extremely
small input buffer (e.g., Solaris 8 is reported to
support only 8 characters).
3. Some implementations of getpass() will fall back to
reading from stdin (e.g., glibc). We explicitly don't
want this, because our stdin may be connected to a pipe
speaking a particular protocol, and reading will
disrupt the protocol flow (e.g., the remote-curl
helper).
4. Some implementations of getpass() turn off signals, so
that hitting "^C" on the terminal does not break out of
the password prompt. This can be a mild annoyance.
Instead, let's provide an abstract "git_terminal_prompt"
function that addresses these concerns. This patch includes
an implementation based on /dev/tty, enabled by setting
HAVE_DEV_TTY. The fallback is to use getpass() as before.
Signed-off-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2011-12-10 18:41:01 +08:00
|
|
|
{
|
|
|
|
return getpass(prompt);
|
|
|
|
}
|
|
|
|
|
2020-01-15 02:43:49 +08:00
|
|
|
int read_key_without_echo(struct strbuf *buf)
|
|
|
|
{
|
|
|
|
static int warning_displayed;
|
|
|
|
const char *res;
|
|
|
|
|
|
|
|
if (!warning_displayed) {
|
|
|
|
warning("reading single keystrokes not supported on this "
|
|
|
|
"platform; reading line instead");
|
|
|
|
warning_displayed = 1;
|
|
|
|
}
|
|
|
|
|
|
|
|
res = getpass("");
|
|
|
|
strbuf_reset(buf);
|
|
|
|
if (!res)
|
|
|
|
return EOF;
|
|
|
|
strbuf_addstr(buf, res);
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
add generic terminal prompt function
When we need to prompt the user for input interactively, we
want to access their terminal directly. We can't rely on
stdio because it may be connected to pipes or files, rather
than the terminal. Instead, we use "getpass()", because it
abstracts the idea of prompting and reading from the
terminal. However, it has some problems:
1. It never echoes the typed characters, which makes it OK
for passwords but annoying for other input (like usernames).
2. Some implementations of getpass() have an extremely
small input buffer (e.g., Solaris 8 is reported to
support only 8 characters).
3. Some implementations of getpass() will fall back to
reading from stdin (e.g., glibc). We explicitly don't
want this, because our stdin may be connected to a pipe
speaking a particular protocol, and reading will
disrupt the protocol flow (e.g., the remote-curl
helper).
4. Some implementations of getpass() turn off signals, so
that hitting "^C" on the terminal does not break out of
the password prompt. This can be a mild annoyance.
Instead, let's provide an abstract "git_terminal_prompt"
function that addresses these concerns. This patch includes
an implementation based on /dev/tty, enabled by setting
HAVE_DEV_TTY. The fallback is to use getpass() as before.
Signed-off-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2011-12-10 18:41:01 +08:00
|
|
|
#endif
|