binutils-gdb/gdb/event-top.c

1198 lines
35 KiB
C
Raw Normal View History

1999-05-11 22:29:07 +02:00
/* Top level stuff for GDB, the GNU debugger.
Copyright (C) 1999-2016 Free Software Foundation, Inc.
1999-05-11 22:29:07 +02:00
Written by Elena Zannoni <ezannoni@cygnus.com> of Cygnus Solutions.
This file is part of GDB.
This program is free software; you can redistribute it and/or modify
it under the terms of the GNU General Public License as published by
the Free Software Foundation; either version 3 of the License, or
1999-05-11 22:29:07 +02:00
(at your option) any later version.
This program is distributed in the hope that it will be useful,
but WITHOUT ANY WARRANTY; without even the implied warranty of
MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
GNU General Public License for more details.
You should have received a copy of the GNU General Public License
along with this program. If not, see <http://www.gnu.org/licenses/>. */
1999-05-11 22:29:07 +02:00
#include "defs.h"
1999-06-14 20:08:47 +02:00
#include "top.h"
1999-05-11 22:29:07 +02:00
#include "inferior.h"
Add new infrun.h header. Move infrun.c declarations out of inferior.h to a new infrun.h file. Tested by building on: i686-w64-mingw32, enable-targets=all x86_64-linux, enable-targets=all i586-pc-msdosdjgpp And also grepped the whole tree for each symbol moved to find where infrun.h might be necessary. gdb/ 2014-05-22 Pedro Alves <palves@redhat.com> * inferior.h (debug_infrun, debug_displaced, stop_on_solib_events) (sync_execution, sched_multi, step_stop_if_no_debug, non_stop) (disable_randomization, enum exec_direction_kind) (execution_direction, stop_registers, start_remote) (clear_proceed_status, proceed, resume, user_visible_resume_ptid) (wait_for_inferior, normal_stop, get_last_target_status) (prepare_for_detach, fetch_inferior_event, init_wait_for_inferior) (insert_step_resume_breakpoint_at_sal) (follow_inferior_reset_breakpoints, stepping_past_instruction_at) (set_step_info, print_stop_event, signal_stop_state) (signal_print_state, signal_pass_state, signal_stop_update) (signal_print_update, signal_pass_update) (update_signals_program_target, clear_exit_convenience_vars) (displaced_step_dump_bytes, update_observer_mode) (signal_catch_update, gdb_signal_from_command): Move declarations ... * infrun.h: ... to this new file. * amd64-tdep.c: Include infrun.h. * annotate.c: Include infrun.h. * arch-utils.c: Include infrun.h. * arm-linux-tdep.c: Include infrun.h. * arm-tdep.c: Include infrun.h. * break-catch-sig.c: Include infrun.h. * breakpoint.c: Include infrun.h. * common/agent.c: Include infrun.h instead of inferior.h. * corelow.c: Include infrun.h. * event-top.c: Include infrun.h. * go32-nat.c: Include infrun.h. * i386-tdep.c: Include infrun.h. * inf-loop.c: Include infrun.h. * infcall.c: Include infrun.h. * infcmd.c: Include infrun.h. * infrun.c: Include infrun.h. * linux-fork.c: Include infrun.h. * linux-nat.c: Include infrun.h. * linux-thread-db.c: Include infrun.h. * monitor.c: Include infrun.h. * nto-tdep.c: Include infrun.h. * procfs.c: Include infrun.h. * record-btrace.c: Include infrun.h. * record-full.c: Include infrun.h. * remote-m32r-sdi.c: Include infrun.h. * remote-mips.c: Include infrun.h. * remote-notif.c: Include infrun.h. * remote-sim.c: Include infrun.h. * remote.c: Include infrun.h. * reverse.c: Include infrun.h. * rs6000-tdep.c: Include infrun.h. * s390-linux-tdep.c: Include infrun.h. * solib-irix.c: Include infrun.h. * solib-osf.c: Include infrun.h. * solib-svr4.c: Include infrun.h. * target.c: Include infrun.h. * top.c: Include infrun.h. * windows-nat.c: Include infrun.h. * mi/mi-interp.c: Include infrun.h. * mi/mi-main.c: Include infrun.h. * python/py-threadevent.c: Include infrun.h.
2014-05-22 13:29:11 +02:00
#include "infrun.h"
1999-10-26 05:43:48 +02:00
#include "target.h"
1999-07-07 22:19:36 +02:00
#include "terminal.h" /* for job_control */
1999-06-21 15:27:42 +02:00
#include "event-loop.h"
1999-09-22 05:28:34 +02:00
#include "event-top.h"
#include "interps.h"
#include <signal.h>
2006-07-21 16:46:56 +02:00
#include "cli/cli-script.h" /* for reset_command_nest_depth */
#include "main.h"
Add "executing" property to threads. * inferior.h (target_executing): Delete. * gdbthread.h (struct thread_info): Add executing_ field. (set_executing, is_executing): New. * thread.c (main_thread_executing): New. (init_thread_list): Clear it and also main_thread_running. (is_running): Return false if target has no execution. (any_running, is_executing, set_executing): New. * top.c: Include "gdbthread.h". (target_executing): Delete. (execute_command): Replace target_executing check by any_running. * event-top.c: Include "gdbthread.h". (display_gdb_prompt, command_handler): Replace target_executing by is_running. * inf-loop.c: Include "gdbthread.h". Don't mark as not executing here. Replace target_executing by is_running. * infrun.c (handle_inferior_event): Mark all threads as not-executing. * linux-nat.c (linux_nat_resume): Don't mark thread as executing here. * stack.c (get_selected_block): Return null if inferior is executing. * target.c (target_resume): Mark resumed ptid as executing. * breakpoint.c (until_break_command): Replace target_executing check by is_executing. * remote.c (remote_async_resume): Don't mark inferior as executing here. * mi/mi-interp.c (mi_cmd_interpreter_exec): Replace target_executing by any_running. * mi/mi-main.c (mi_cmd_exec_interrupt, mi_cmd_execute) (mi_execute_async_cli_command): Replace target_executing by is_running. * frame.c (get_current_frame): Error out if the current thread is executing. (has_stack_frames): New. (get_selected_frame, deprecated_safe_get_selected_frame): Check has_stack_frames. * Makefile.in (event-top.o, frame.o, inf-loop.o, top.o): Depend on $(gdbthread_h).
2008-07-10 00:16:15 +02:00
#include "gdbthread.h"
#include "observer.h"
#include "continuations.h"
#include "gdbcmd.h" /* for dont_repeat() */
With some changes to how software single-step (SSS) breakpoints are handled, one of those being to place SSS breakpoints on the breakpoint chain as all other breakpoints, annota1.exp times out with lots and lots of breakpoint-invalid and frame-changed annotations. All those extra annotations are actually unnecessary. For one, SSS breakpoints are internal breakpoints, so the frontend shouldn't care if they were added, removed or changed. Then, there's really no point in emitting "breakpoints-invalid" or "frames-invalid" more than once between times the frontend/user can actually issues GDB commands; the frontend will have to wait for the GDB prompt to refresh its state, so emitting those annotations at most once between prompts is enough. Non-stop or async would complicate this, but no frontend will be using annotations in those modes (one of goes of emacs switching to MI was non-stop mode support, AFAIK). The previous patch reveals there has been an intention in the past to suppress multiple breakpoints-invalid annotations caused by ignore count changes. As the previous patch shows, that's always been broken, but in any case, this patch actually makes it work. The next patch will remove several annotation-specific calls in breakpoint.c in favor of always using the breakpoint modified & friends observers, and that causes yet more of these annotations, because several calls to the corresponding annotate_* functions in breakpoint.c are missing, particularly in newer code. So all in all, here's a simple mechanism that avoids sending the same annotation to the frontend more than once until gdb is ready to accept further commands. Tested on x86_64 Fedora 17. 2013-01-22 Pedro Alves <palves@redhat.com> * annotate.c: Include "inferior.h". (frames_invalid_emitted) (breakpoints_invalid_emitted): New globals. (async_background_execution_p): New function. (annotate_breakpoints_changed, annotate_frames_invalid): Skip emitting the annotation if it has already been emitted. (annotate_display_prompt): New function. * annotate.h (annotate_display_prompt): New declaration. * event-top.c: Include annotate.h. (display_gdb_prompt): Call annotate_display_prompt.
2013-01-22 21:17:10 +01:00
#include "annotate.h"
New commands "mt set per-command {space,time,symtab} {on,off}". * NEWS: Add entry. * event-top.c: #include "maint.h". * main.c: #include "maint.h". * maint.c: #include <sys/time.h>, <time.h>, block.h, top.h, timeval-utils.h, maint.h, cli/cli-setshow.h. (per_command_time, per_command_space): New static globals. (per_command_symtab): New static global. (per_command_setlist, per_command_showlist): New static globals. (struct cmd_stats): Move here from utils.c. (set_per_command_time): Renamed from set_display_time in utils.c and moved here. All callers updated. (set_per_command_space): Renamed from set_display_space in utils.c and moved here. All callers updated. (count_symtabs_and_blocks): New function. (report_command_stats): Moved here from utils.c. Add support for printing symtab stats. Only print data if enabled before command executed. (make_command_stats_cleanup): Ditto. (sert_per_command_cmd, show_per_command_cmd): New functions. (_initialize_maint_cmds): Add new commands mt set per-command {space,time,symtab} {on,off}. * maint.h: New file. * top.c: #include "maint.h". * utils.c (reset_prompt_for_continue_wait_time): New function. (get_prompt_for_continue_wait_time): New function. * utils.h (reset_prompt_for_continue_wait_time): Declare (get_prompt_for_continue_wait_time): Declare. (make_command_stats_cleanup): Moved to maint.h. (set_display_time, set_display_space): Moved to maint.h and renamed to set_per_command_time, set_per_command_space. * cli/cli-setshow.c (parse_cli_boolean_value): Renamed from parse_binary_operation and made non-static. Don't call error, just return an error marker. All callers updated. * cli/cli-setshow.h (parse_cli_boolean_value): Declare. doc/ * gdb.texinfo (Maintenance Commands): Add docs for "mt set per-command {space,time,symtab} {on,off}". testsuite/ * gdb.base/maint.exp: Update tests for per-command stats.
2013-03-21 18:37:30 +01:00
#include "maint.h"
#include "buffer.h"
Introduce interruptible_select We have places where we call a blocking gdb_select expecting that a Ctrl-C will unblock it. However, if the Ctrl-C is pressed just before gdb_select, the SIGINT handler runs before gdb_select, and thus gdb_select won't return. For example gdb_readline_no_editing: QUIT; /* Wait until at least one byte of data is available. Control-C can interrupt gdb_select, but not fgetc. */ FD_ZERO (&readfds); FD_SET (fd, &readfds); if (gdb_select (fd + 1, &readfds, NULL, NULL, NULL) == -1) and stdio_file_read: /* For the benefit of Windows, call gdb_select before reading from the file. Wait until at least one byte of data is available. Control-C can interrupt gdb_select, but not read. */ { fd_set readfds; FD_ZERO (&readfds); FD_SET (stdio->fd, &readfds); if (gdb_select (stdio->fd + 1, &readfds, NULL, NULL, NULL) == -1) return -1; } return read (stdio->fd, buf, length_buf); This is a race classically fixed with either the self-pipe trick, or by blocking SIGINT and then using pselect instead of select. Blocking SIGINT most of the time would mean that check_quit_flag (and thus QUIT) would need to do a syscall every time it is called, which sounds best avoided, since QUIT is called in many loops. Thus we take the self-pipe trick route (wrapped in a serial event). Instead of having all places that need this manually add an extra file descriptor to the set of gdb_select's watched file descriptors, we introduce a wrapper, interruptible_select, that does that. The Windows version of gdb_select actually does not suffer from this, because mingw-hdep.c:gdb_call_async_signal_handler sets a Windows event that gdb_select always waits on. So this patch can be seen as generalization of that technique. We can't remove that extra event from mingw-hdep.c until we get rid of immediate_quit though. gdb/ChangeLog: 2016-04-12 Pedro Alves <palves@redhat.com> * defs.h: Extend QUIT-related comments to mention interruptible_select. (quit_serial_event_set, quit_serial_event_clear): Declare. * event-top.c: Include "ser-event.h" and "gdb_select.h". (quit_serial_event): New global. (async_init_signals): Make quit_serial_event. (quit_serial_event_set, quit_serial_event_clear) (quit_serial_event_fd, interruptible_select): New functions. * extension.c (set_quit_flag): Set the quit serial event. (check_quit_flag): Clear the quit serial event. * gdb_select.h (interruptible_select): New declaration. * guile/scm-ports.c (ioscm_input_waiting): Use interruptible_select instead of gdb_select. * top.c (gdb_readline_no_editing): Likewise. * ui-file.c (stdio_file_read): Likewise.
2016-04-12 17:49:30 +02:00
#include "ser-event.h"
#include "gdb_select.h"
1999-08-31 03:14:27 +02:00
/* readline include files. */
#include "readline/readline.h"
#include "readline/history.h"
1999-05-11 22:29:07 +02:00
/* readline defines this. */
#undef savestring
1999-09-22 05:28:34 +02:00
static void rl_callback_read_char_wrapper (gdb_client_data client_data);
static void command_line_handler (char *rl);
static void change_line_handler (void);
static char *top_level_prompt (void);
1999-05-11 22:29:07 +02:00
/* Signal handlers. */
#ifdef SIGQUIT
1999-09-22 05:28:34 +02:00
static void handle_sigquit (int sig);
#endif
#ifdef SIGHUP
1999-09-22 05:28:34 +02:00
static void handle_sighup (int sig);
#endif
1999-09-22 05:28:34 +02:00
static void handle_sigfpe (int sig);
1999-05-11 22:29:07 +02:00
/* Functions to be invoked by the event loop in response to
signals. */
#if defined (SIGQUIT) || defined (SIGHUP)
1999-09-22 05:28:34 +02:00
static void async_do_nothing (gdb_client_data);
#endif
#ifdef SIGHUP
1999-09-22 05:28:34 +02:00
static void async_disconnect (gdb_client_data);
#endif
1999-09-22 05:28:34 +02:00
static void async_float_handler (gdb_client_data);
#ifdef STOP_SIGNAL
1999-09-22 05:28:34 +02:00
static void async_stop_sig (gdb_client_data);
#endif
static void async_sigterm_handler (gdb_client_data arg);
1999-05-11 22:29:07 +02:00
/* Readline offers an alternate interface, via callback
functions. These are all included in the file callback.c in the
1999-05-11 22:29:07 +02:00
readline distribution. This file provides (mainly) a function, which
the event loop uses as callback (i.e. event handler) whenever an event
is detected on the standard input file descriptor.
readline_callback_read_char is called (by the GDB event loop) whenever
there is a new character ready on the input stream. This function
1999-05-11 22:29:07 +02:00
incrementally builds a buffer internal to readline where it
accumulates the line read up to the point of invocation. In the
special case in which the character read is newline, the function
invokes a GDB supplied callback routine, which does the processing of
a full command line. This latter routine is the asynchronous analog
of the old command_line_input in gdb. Instead of invoking (and waiting
1999-05-11 22:29:07 +02:00
for) readline to read the command line and pass it back to
command_loop for processing, the new command_line_handler function has
the command line already available as its parameter. INPUT_HANDLER is
to be set to the function that readline will invoke when a complete
line of input is ready. CALL_READLINE is to be set to the function
that readline offers as callback to the event_loop. */
1999-05-11 22:29:07 +02:00
1999-09-22 05:28:34 +02:00
void (*input_handler) (char *);
void (*call_readline) (gdb_client_data);
1999-05-11 22:29:07 +02:00
/* Important variables for the event loop. */
1999-05-11 22:29:07 +02:00
/* This is used to determine if GDB is using the readline library or
its own simplified form of readline. It is used by the asynchronous
1999-06-14 20:08:47 +02:00
form of the set editing command.
1999-05-25 20:09:09 +02:00
ezannoni: as of 1999-04-29 I expect that this
1999-05-11 22:29:07 +02:00
variable will not be used after gdb is changed to use the event
loop as default engine, and event-top.c is merged into top.c. */
1999-05-11 22:29:07 +02:00
int async_command_editing_p;
1999-08-31 03:14:27 +02:00
/* This is used to display the notification of the completion of an
asynchronous execution command. */
1999-08-31 03:14:27 +02:00
int exec_done_display_p = 0;
1999-05-11 22:29:07 +02:00
/* This is the file descriptor for the input stream that GDB uses to
read commands from. */
1999-05-11 22:29:07 +02:00
int input_fd;
Fix the processing of Meta-key commands in TUI This patch fixes the annoying bug where key sequences such as Alt_F or Alt_B (go forward or backwards by a word) do not behave promptly in TUI. You have to press a third key in order for the key sequence to register. This is mostly ncurses' fault. Calling wgetch() normally causes ncurses to read only a single key from stdin. However if the key read is the start-sequence key (^[ a.k.a. ESC) then wgetch() reads TWO keys from stdin, storing the 2nd key into an internal FIFO buffer and returning the start-sequence key. The extraneous read of the 2nd key makes us miss its corresponding stdin event, so the event loop blocks until a third key is pressed. This explains why such key sequences do not behave promptly in TUI. To fix this issue, we must somehow compensate for the missed stdin event corresponding to the 2nd byte of a key sequence. This patch achieves this by hacking up the stdin event handler to conditionally execute the readline callback multiple times in a row. This is done via a new global variable, call_stdin_event_handler_again_p, which is set from tui_getc() when we receive a start-sequence key and notice extra pending input in the ncurses buffer. Tested on x86_64-unknown-linux-gnu. gdb/ChangeLog: * event-top.h (call_stdin_event_handler_again_p): Declare. * event-top.c (call_stdin_event_handler_again_p): Define. (stdin_event_handler): Use it. * tui/tui-io.c (tui_getc): Prepare to call the stdin event handler again if there is pending input following a start sequence.
2014-11-23 11:03:39 +01:00
/* Used by the stdin event handler to compensate for missed stdin events.
Setting this to a non-zero value inside an stdin callback makes the callback
run again. */
int call_stdin_event_handler_again_p;
/* Signal handling variables. */
1999-05-11 22:29:07 +02:00
/* Each of these is a pointer to a function that the event loop will
invoke if the corresponding signal has received. The real signal
1999-05-11 22:29:07 +02:00
handlers mark these functions as ready to be executed and the event
loop, in a later iteration, calls them. See the function
invoke_async_signal_handler. */
static struct async_signal_handler *sigint_token;
1999-05-11 22:29:07 +02:00
#ifdef SIGHUP
static struct async_signal_handler *sighup_token;
1999-05-11 22:29:07 +02:00
#endif
#ifdef SIGQUIT
static struct async_signal_handler *sigquit_token;
#endif
static struct async_signal_handler *sigfpe_token;
1999-06-14 20:08:47 +02:00
#ifdef STOP_SIGNAL
static struct async_signal_handler *sigtstp_token;
1999-06-14 20:08:47 +02:00
#endif
static struct async_signal_handler *async_sigterm_token;
1999-06-14 20:08:47 +02:00
/* This hook is called by rl_callback_read_char_wrapper after each
character is processed. */
void (*after_char_processing_hook) (void);
1999-05-11 22:29:07 +02:00
/* Wrapper function for calling into the readline library. The event
loop expects the callback function to have a paramter, while
readline expects none. */
1999-09-22 05:28:34 +02:00
static void
rl_callback_read_char_wrapper (gdb_client_data client_data)
{
rl_callback_read_char ();
if (after_char_processing_hook)
(*after_char_processing_hook) ();
1999-09-22 05:28:34 +02:00
}
1999-05-11 22:29:07 +02:00
/* Initialize all the necessary variables, start the event loop,
register readline, and stdin, start the loop. The DATA is the
interpreter data cookie, ignored for now. */
1999-05-11 22:29:07 +02:00
void
cli_command_loop (void *data)
1999-05-11 22:29:07 +02:00
{
display_gdb_prompt (0);
1999-05-11 22:29:07 +02:00
/* Now it's time to start the event loop. */
1999-06-28 18:06:02 +02:00
start_event_loop ();
1999-05-11 22:29:07 +02:00
}
/* Change the function to be invoked every time there is a character
ready on stdin. This is used when the user sets the editing off,
1999-05-11 22:29:07 +02:00
therefore bypassing readline, and letting gdb handle the input
itself, via gdb_readline_no_editing_callback. Also it is used in
the opposite case in which the user sets editing on again, by
restoring readline handling of the input. */
1999-05-25 20:09:09 +02:00
static void
1999-09-22 05:28:34 +02:00
change_line_handler (void)
1999-05-11 22:29:07 +02:00
{
/* NOTE: this operates on input_fd, not instream. If we are reading
commands from a file, instream will point to the file. However in
1999-09-22 05:28:34 +02:00
async mode, we always read commands from a file with editing
off. This means that the 'set editing on/off' will have effect
only on the interactive session. */
1999-09-22 05:28:34 +02:00
1999-05-11 22:29:07 +02:00
if (async_command_editing_p)
{
/* Turn on editing by using readline. */
1999-09-22 05:28:34 +02:00
call_readline = rl_callback_read_char_wrapper;
1999-06-14 20:08:47 +02:00
input_handler = command_line_handler;
1999-05-11 22:29:07 +02:00
}
else
{
/* Turn off editing by using gdb_readline_no_editing_callback. */
PR python/17372 - Python hangs when displaying help() This is more of a readline/terminal issue than a Python one. PR17372 is a regression in 7.8 caused by the fix for PR17072: commit 0017922d0292d8c374584f6100874580659c9973 Author: Pedro Alves <palves@redhat.com> Date: Mon Jul 14 19:55:32 2014 +0100 Background execution + pagination aborts readline/gdb gdb_readline_wrapper_line removes the handler after a line is processed. Usually, we'll end up re-displaying the prompt, and that reinstalls the handler. But if the output is coming out of handling a stop event, we don't re-display the prompt, and nothing restores the handler. So the next input wakes up the event loop and calls into readline, which aborts. ... gdb/ 2014-07-14 Pedro Alves <palves@redhat.com> PR gdb/17072 * top.c (gdb_readline_wrapper_line): Tweak comment. (gdb_readline_wrapper_cleanup): If readline is enabled, reinstall the input handler callback. The problem is that installing the input handler callback also preps the terminal, putting it in raw mode and with echo disabled, which is bad if we're going to call a command that assumes cooked/canonical mode, and echo enabled, like in the case of the PR, Python's interactive shell. Another example I came up with that doesn't depend on Python is starting a subshell with "(gdb) shell /bin/sh" from a multi-line command. Tests covering both these examples are added. The fix is to revert the original fix for PR gdb/17072, and instead restore the callback handler after processing an asynchronous target event. Furthermore, calling rl_callback_handler_install when we already have some input in readline's line buffer discards that input, which is obviously a bad thing to do while the user is typing. No specific test is added for that, because I first tried calling it even if the callback handler was still installed and that resulted in hundreds of failures in the testsuite. gdb/ 2014-10-29 Pedro Alves <palves@redhat.com> PR python/17372 * event-top.c (change_line_handler): Call gdb_rl_callback_handler_remove instead of rl_callback_handler_remove. (callback_handler_installed): New global. (gdb_rl_callback_handler_remove, gdb_rl_callback_handler_install) (gdb_rl_callback_handler_reinstall): New functions. (display_gdb_prompt): Call gdb_rl_callback_handler_remove and gdb_rl_callback_handler_install instead of rl_callback_handler_remove and rl_callback_handler_install. (gdb_disable_readline): Call gdb_rl_callback_handler_remove instead of rl_callback_handler_remove. * event-top.h (gdb_rl_callback_handler_remove) (gdb_rl_callback_handler_install) (gdb_rl_callback_handler_reinstall): New declarations. * infrun.c (reinstall_readline_callback_handler_cleanup): New cleanup function. (fetch_inferior_event): Install it. * top.c (gdb_readline_wrapper_line) Call gdb_rl_callback_handler_remove instead of rl_callback_handler_remove. (gdb_readline_wrapper_cleanup): Don't call rl_callback_handler_install. gdb/testsuite/ 2014-10-29 Pedro Alves <palves@redhat.com> PR python/17372 * gdb.python/python.exp: Test a multi-line command that spawns interactive Python. * gdb.base/multi-line-starts-subshell.exp: New file.
2014-10-23 18:13:35 +02:00
gdb_rl_callback_handler_remove ();
call_readline = gdb_readline_no_editing_callback;
1999-06-14 20:08:47 +02:00
/* Set up the command handler as well, in case we are called as
first thing from .gdbinit. */
1999-06-14 20:08:47 +02:00
input_handler = command_line_handler;
1999-05-11 22:29:07 +02:00
}
}
PR python/17372 - Python hangs when displaying help() This is more of a readline/terminal issue than a Python one. PR17372 is a regression in 7.8 caused by the fix for PR17072: commit 0017922d0292d8c374584f6100874580659c9973 Author: Pedro Alves <palves@redhat.com> Date: Mon Jul 14 19:55:32 2014 +0100 Background execution + pagination aborts readline/gdb gdb_readline_wrapper_line removes the handler after a line is processed. Usually, we'll end up re-displaying the prompt, and that reinstalls the handler. But if the output is coming out of handling a stop event, we don't re-display the prompt, and nothing restores the handler. So the next input wakes up the event loop and calls into readline, which aborts. ... gdb/ 2014-07-14 Pedro Alves <palves@redhat.com> PR gdb/17072 * top.c (gdb_readline_wrapper_line): Tweak comment. (gdb_readline_wrapper_cleanup): If readline is enabled, reinstall the input handler callback. The problem is that installing the input handler callback also preps the terminal, putting it in raw mode and with echo disabled, which is bad if we're going to call a command that assumes cooked/canonical mode, and echo enabled, like in the case of the PR, Python's interactive shell. Another example I came up with that doesn't depend on Python is starting a subshell with "(gdb) shell /bin/sh" from a multi-line command. Tests covering both these examples are added. The fix is to revert the original fix for PR gdb/17072, and instead restore the callback handler after processing an asynchronous target event. Furthermore, calling rl_callback_handler_install when we already have some input in readline's line buffer discards that input, which is obviously a bad thing to do while the user is typing. No specific test is added for that, because I first tried calling it even if the callback handler was still installed and that resulted in hundreds of failures in the testsuite. gdb/ 2014-10-29 Pedro Alves <palves@redhat.com> PR python/17372 * event-top.c (change_line_handler): Call gdb_rl_callback_handler_remove instead of rl_callback_handler_remove. (callback_handler_installed): New global. (gdb_rl_callback_handler_remove, gdb_rl_callback_handler_install) (gdb_rl_callback_handler_reinstall): New functions. (display_gdb_prompt): Call gdb_rl_callback_handler_remove and gdb_rl_callback_handler_install instead of rl_callback_handler_remove and rl_callback_handler_install. (gdb_disable_readline): Call gdb_rl_callback_handler_remove instead of rl_callback_handler_remove. * event-top.h (gdb_rl_callback_handler_remove) (gdb_rl_callback_handler_install) (gdb_rl_callback_handler_reinstall): New declarations. * infrun.c (reinstall_readline_callback_handler_cleanup): New cleanup function. (fetch_inferior_event): Install it. * top.c (gdb_readline_wrapper_line) Call gdb_rl_callback_handler_remove instead of rl_callback_handler_remove. (gdb_readline_wrapper_cleanup): Don't call rl_callback_handler_install. gdb/testsuite/ 2014-10-29 Pedro Alves <palves@redhat.com> PR python/17372 * gdb.python/python.exp: Test a multi-line command that spawns interactive Python. * gdb.base/multi-line-starts-subshell.exp: New file.
2014-10-23 18:13:35 +02:00
/* The functions below are wrappers for rl_callback_handler_remove and
rl_callback_handler_install that keep track of whether the callback
handler is installed in readline. This is necessary because after
handling a target event of a background execution command, we may
need to reinstall the callback handler if it was removed due to a
secondary prompt. See gdb_readline_wrapper_line. We don't
unconditionally install the handler for every target event because
that also clears the line buffer, thus installing it while the user
is typing would lose input. */
/* Whether we've registered a callback handler with readline. */
static int callback_handler_installed;
/* See event-top.h, and above. */
void
gdb_rl_callback_handler_remove (void)
{
rl_callback_handler_remove ();
callback_handler_installed = 0;
}
/* See event-top.h, and above. Note this wrapper doesn't have an
actual callback parameter because we always install
INPUT_HANDLER. */
void
gdb_rl_callback_handler_install (const char *prompt)
{
/* Calling rl_callback_handler_install resets readline's input
buffer. Calling this when we were already processing input
therefore loses input. */
gdb_assert (!callback_handler_installed);
rl_callback_handler_install (prompt, input_handler);
callback_handler_installed = 1;
}
/* See event-top.h, and above. */
void
gdb_rl_callback_handler_reinstall (void)
{
if (!callback_handler_installed)
{
/* Passing NULL as prompt argument tells readline to not display
a prompt. */
gdb_rl_callback_handler_install (NULL);
}
}
/* Displays the prompt. If the argument NEW_PROMPT is NULL, the
prompt that is displayed is the current top level prompt.
Otherwise, it displays whatever NEW_PROMPT is as a local/secondary
prompt.
This is used after each gdb command has completed, and in the
following cases:
1. When the user enters a command line which is ended by '\'
indicating that the command will continue on the next line. In
that case the prompt that is displayed is the empty string.
1999-06-14 20:08:47 +02:00
2. When the user is entering 'commands' for a breakpoint, or
actions for a tracepoint. In this case the prompt will be '>'
3. On prompting for pagination. */
1999-05-11 22:29:07 +02:00
void
Fix build with Python 3.4 (PR python/16784) The type of the function pointer PyOS_ReadlineFunctionPointer (part of the Python C API), which we use, slightly changed starting with Python 3.4. The signature went from PyAPI_DATA(char) *(*PyOS_ReadlineFunctionPointer)(FILE *, FILE *, char *); to PyAPI_DATA(char) *(*PyOS_ReadlineFunctionPointer)(FILE *, FILE *, const char *); The parameter that changed is the prompt text. This commits adjust gdb accordingly by making the prompt_arg parameter const, as well as the fallouts of that. I needed to rework how annotations are added to the prompt, since the it is now const. If annotations are enabled, it will make a copy of the prompt overwrite the prompt variable that is used throughout the function. Otherwise, no copy is done and the original prompt_arg value is passed. I changed the signature of deprecated_readline_hook. I would've changed any user of it, but it seems like nothing is using it, Built-tested with python 2.7.x, 3.3.y and 3.4.z. gdb/ChangeLog: * defs.h (gdb_readline): Constify argument. (gdb_readline_wrapper): Same. (command_line_input): Same. (deprecated_readline_hook): Same. * top.c (deprecated_readline_hook): Same. (gdb_readline): Same. (gdb_readline_wrapper): Same. (command_line_input): Constify argument. Pass prompt to called functions instead of local_prompt, overwriting prompt if using annotations. * event-top.h (display_gdb_prompt): Constify argument. * event-top.c (display_gdb_prompt): Same. * python/py-gdb-readline.c (gdbpy_readline_wrapper): Constify argument if building with Python 3.4 and up. Signed-off-by: Simon Marchi <simon.marchi@ericsson.com>
2014-12-15 17:38:03 +01:00
display_gdb_prompt (const char *new_prompt)
1999-05-11 22:29:07 +02:00
{
char *actual_gdb_prompt = NULL;
struct cleanup *old_chain;
1999-05-11 22:29:07 +02:00
With some changes to how software single-step (SSS) breakpoints are handled, one of those being to place SSS breakpoints on the breakpoint chain as all other breakpoints, annota1.exp times out with lots and lots of breakpoint-invalid and frame-changed annotations. All those extra annotations are actually unnecessary. For one, SSS breakpoints are internal breakpoints, so the frontend shouldn't care if they were added, removed or changed. Then, there's really no point in emitting "breakpoints-invalid" or "frames-invalid" more than once between times the frontend/user can actually issues GDB commands; the frontend will have to wait for the GDB prompt to refresh its state, so emitting those annotations at most once between prompts is enough. Non-stop or async would complicate this, but no frontend will be using annotations in those modes (one of goes of emacs switching to MI was non-stop mode support, AFAIK). The previous patch reveals there has been an intention in the past to suppress multiple breakpoints-invalid annotations caused by ignore count changes. As the previous patch shows, that's always been broken, but in any case, this patch actually makes it work. The next patch will remove several annotation-specific calls in breakpoint.c in favor of always using the breakpoint modified & friends observers, and that causes yet more of these annotations, because several calls to the corresponding annotate_* functions in breakpoint.c are missing, particularly in newer code. So all in all, here's a simple mechanism that avoids sending the same annotation to the frontend more than once until gdb is ready to accept further commands. Tested on x86_64 Fedora 17. 2013-01-22 Pedro Alves <palves@redhat.com> * annotate.c: Include "inferior.h". (frames_invalid_emitted) (breakpoints_invalid_emitted): New globals. (async_background_execution_p): New function. (annotate_breakpoints_changed, annotate_frames_invalid): Skip emitting the annotation if it has already been emitted. (annotate_display_prompt): New function. * annotate.h (annotate_display_prompt): New declaration. * event-top.c: Include annotate.h. (display_gdb_prompt): Call annotate_display_prompt.
2013-01-22 21:17:10 +01:00
annotate_display_prompt ();
2006-07-21 16:46:56 +02:00
/* Reset the nesting depth used when trace-commands is set. */
reset_command_nest_depth ();
old_chain = make_cleanup (free_current_contents, &actual_gdb_prompt);
/* Do not call the python hook on an explicit prompt change as
passed to this function, as this forms a secondary/local prompt,
IE, displayed but not set. */
if (! new_prompt)
1999-07-20 01:30:11 +02:00
{
if (sync_execution)
{
/* This is to trick readline into not trying to display the
prompt. Even though we display the prompt using this
function, readline still tries to do its own display if
we don't call rl_callback_handler_install and
rl_callback_handler_remove (which readline detects
because a global variable is not set). If readline did
that, it could mess up gdb signal handlers for SIGINT.
Readline assumes that between calls to rl_set_signals and
rl_clear_signals gdb doesn't do anything with the signal
handlers. Well, that's not the case, because when the
target executes we change the SIGINT signal handler. If
we allowed readline to display the prompt, the signal
handler change would happen exactly between the calls to
the above two functions. Calling
rl_callback_handler_remove(), does the job. */
PR python/17372 - Python hangs when displaying help() This is more of a readline/terminal issue than a Python one. PR17372 is a regression in 7.8 caused by the fix for PR17072: commit 0017922d0292d8c374584f6100874580659c9973 Author: Pedro Alves <palves@redhat.com> Date: Mon Jul 14 19:55:32 2014 +0100 Background execution + pagination aborts readline/gdb gdb_readline_wrapper_line removes the handler after a line is processed. Usually, we'll end up re-displaying the prompt, and that reinstalls the handler. But if the output is coming out of handling a stop event, we don't re-display the prompt, and nothing restores the handler. So the next input wakes up the event loop and calls into readline, which aborts. ... gdb/ 2014-07-14 Pedro Alves <palves@redhat.com> PR gdb/17072 * top.c (gdb_readline_wrapper_line): Tweak comment. (gdb_readline_wrapper_cleanup): If readline is enabled, reinstall the input handler callback. The problem is that installing the input handler callback also preps the terminal, putting it in raw mode and with echo disabled, which is bad if we're going to call a command that assumes cooked/canonical mode, and echo enabled, like in the case of the PR, Python's interactive shell. Another example I came up with that doesn't depend on Python is starting a subshell with "(gdb) shell /bin/sh" from a multi-line command. Tests covering both these examples are added. The fix is to revert the original fix for PR gdb/17072, and instead restore the callback handler after processing an asynchronous target event. Furthermore, calling rl_callback_handler_install when we already have some input in readline's line buffer discards that input, which is obviously a bad thing to do while the user is typing. No specific test is added for that, because I first tried calling it even if the callback handler was still installed and that resulted in hundreds of failures in the testsuite. gdb/ 2014-10-29 Pedro Alves <palves@redhat.com> PR python/17372 * event-top.c (change_line_handler): Call gdb_rl_callback_handler_remove instead of rl_callback_handler_remove. (callback_handler_installed): New global. (gdb_rl_callback_handler_remove, gdb_rl_callback_handler_install) (gdb_rl_callback_handler_reinstall): New functions. (display_gdb_prompt): Call gdb_rl_callback_handler_remove and gdb_rl_callback_handler_install instead of rl_callback_handler_remove and rl_callback_handler_install. (gdb_disable_readline): Call gdb_rl_callback_handler_remove instead of rl_callback_handler_remove. * event-top.h (gdb_rl_callback_handler_remove) (gdb_rl_callback_handler_install) (gdb_rl_callback_handler_reinstall): New declarations. * infrun.c (reinstall_readline_callback_handler_cleanup): New cleanup function. (fetch_inferior_event): Install it. * top.c (gdb_readline_wrapper_line) Call gdb_rl_callback_handler_remove instead of rl_callback_handler_remove. (gdb_readline_wrapper_cleanup): Don't call rl_callback_handler_install. gdb/testsuite/ 2014-10-29 Pedro Alves <palves@redhat.com> PR python/17372 * gdb.python/python.exp: Test a multi-line command that spawns interactive Python. * gdb.base/multi-line-starts-subshell.exp: New file.
2014-10-23 18:13:35 +02:00
gdb_rl_callback_handler_remove ();
do_cleanups (old_chain);
return;
}
else
{
/* Display the top level prompt. */
actual_gdb_prompt = top_level_prompt ();
}
1999-05-11 22:29:07 +02:00
}
else
actual_gdb_prompt = xstrdup (new_prompt);
1999-05-11 22:29:07 +02:00
if (async_command_editing_p)
{
PR python/17372 - Python hangs when displaying help() This is more of a readline/terminal issue than a Python one. PR17372 is a regression in 7.8 caused by the fix for PR17072: commit 0017922d0292d8c374584f6100874580659c9973 Author: Pedro Alves <palves@redhat.com> Date: Mon Jul 14 19:55:32 2014 +0100 Background execution + pagination aborts readline/gdb gdb_readline_wrapper_line removes the handler after a line is processed. Usually, we'll end up re-displaying the prompt, and that reinstalls the handler. But if the output is coming out of handling a stop event, we don't re-display the prompt, and nothing restores the handler. So the next input wakes up the event loop and calls into readline, which aborts. ... gdb/ 2014-07-14 Pedro Alves <palves@redhat.com> PR gdb/17072 * top.c (gdb_readline_wrapper_line): Tweak comment. (gdb_readline_wrapper_cleanup): If readline is enabled, reinstall the input handler callback. The problem is that installing the input handler callback also preps the terminal, putting it in raw mode and with echo disabled, which is bad if we're going to call a command that assumes cooked/canonical mode, and echo enabled, like in the case of the PR, Python's interactive shell. Another example I came up with that doesn't depend on Python is starting a subshell with "(gdb) shell /bin/sh" from a multi-line command. Tests covering both these examples are added. The fix is to revert the original fix for PR gdb/17072, and instead restore the callback handler after processing an asynchronous target event. Furthermore, calling rl_callback_handler_install when we already have some input in readline's line buffer discards that input, which is obviously a bad thing to do while the user is typing. No specific test is added for that, because I first tried calling it even if the callback handler was still installed and that resulted in hundreds of failures in the testsuite. gdb/ 2014-10-29 Pedro Alves <palves@redhat.com> PR python/17372 * event-top.c (change_line_handler): Call gdb_rl_callback_handler_remove instead of rl_callback_handler_remove. (callback_handler_installed): New global. (gdb_rl_callback_handler_remove, gdb_rl_callback_handler_install) (gdb_rl_callback_handler_reinstall): New functions. (display_gdb_prompt): Call gdb_rl_callback_handler_remove and gdb_rl_callback_handler_install instead of rl_callback_handler_remove and rl_callback_handler_install. (gdb_disable_readline): Call gdb_rl_callback_handler_remove instead of rl_callback_handler_remove. * event-top.h (gdb_rl_callback_handler_remove) (gdb_rl_callback_handler_install) (gdb_rl_callback_handler_reinstall): New declarations. * infrun.c (reinstall_readline_callback_handler_cleanup): New cleanup function. (fetch_inferior_event): Install it. * top.c (gdb_readline_wrapper_line) Call gdb_rl_callback_handler_remove instead of rl_callback_handler_remove. (gdb_readline_wrapper_cleanup): Don't call rl_callback_handler_install. gdb/testsuite/ 2014-10-29 Pedro Alves <palves@redhat.com> PR python/17372 * gdb.python/python.exp: Test a multi-line command that spawns interactive Python. * gdb.base/multi-line-starts-subshell.exp: New file.
2014-10-23 18:13:35 +02:00
gdb_rl_callback_handler_remove ();
gdb_rl_callback_handler_install (actual_gdb_prompt);
1999-05-11 22:29:07 +02:00
}
/* new_prompt at this point can be the top of the stack or the one
passed in. It can't be NULL. */
else
1999-05-11 22:29:07 +02:00
{
/* Don't use a _filtered function here. It causes the assumed
character position to be off, since the newline we read from
the user is not accounted for. */
fputs_unfiltered (actual_gdb_prompt, gdb_stdout);
1999-05-11 22:29:07 +02:00
gdb_flush (gdb_stdout);
}
do_cleanups (old_chain);
1999-05-11 22:29:07 +02:00
}
/* Return the top level prompt, as specified by "set prompt", possibly
overriden by the python gdb.prompt_hook hook, and then composed
with the prompt prefix and suffix (annotations). The caller is
responsible for freeing the returned string. */
static char *
top_level_prompt (void)
1999-05-11 22:29:07 +02:00
{
char *prompt;
1999-05-11 22:29:07 +02:00
/* Give observers a chance of changing the prompt. E.g., the python
`gdb.prompt_hook' is installed as an observer. */
observer_notify_before_prompt (get_prompt ());
prompt = get_prompt ();
1999-05-11 22:29:07 +02:00
if (annotation_level >= 2)
1999-05-11 22:29:07 +02:00
{
/* Prefix needs to have new line at end. */
const char prefix[] = "\n\032\032pre-prompt\n";
/* Suffix needs to have a new line at end and \032 \032 at
beginning. */
const char suffix[] = "\n\032\032prompt\n";
1999-05-11 22:29:07 +02:00
return concat (prefix, prompt, suffix, NULL);
}
return xstrdup (prompt);
1999-05-11 22:29:07 +02:00
}
1999-09-22 05:28:34 +02:00
/* Get a pointer to the command line buffer. This is used to
construct a whole line of input from partial input. */
static struct buffer *
get_command_line_buffer (void)
{
static struct buffer line_buffer;
static int line_buffer_initialized;
if (!line_buffer_initialized)
{
buffer_init (&line_buffer);
line_buffer_initialized = 1;
}
return &line_buffer;
}
/* When there is an event ready on the stdin file descriptor, instead
1999-09-22 05:28:34 +02:00
of calling readline directly throught the callback function, or
instead of calling gdb_readline_no_editing_callback, give gdb a
chance to detect errors and do something. */
1999-09-22 05:28:34 +02:00
void
1999-10-06 01:13:56 +02:00
stdin_event_handler (int error, gdb_client_data client_data)
1999-09-22 05:28:34 +02:00
{
if (error)
{
printf_unfiltered (_("error detected on stdin\n"));
1999-10-06 01:13:56 +02:00
delete_file_handler (input_fd);
/* If stdin died, we may as well kill gdb. */
2000-01-18 01:55:13 +01:00
quit_command ((char *) 0, stdin == instream);
1999-09-22 05:28:34 +02:00
}
else
Fix the processing of Meta-key commands in TUI This patch fixes the annoying bug where key sequences such as Alt_F or Alt_B (go forward or backwards by a word) do not behave promptly in TUI. You have to press a third key in order for the key sequence to register. This is mostly ncurses' fault. Calling wgetch() normally causes ncurses to read only a single key from stdin. However if the key read is the start-sequence key (^[ a.k.a. ESC) then wgetch() reads TWO keys from stdin, storing the 2nd key into an internal FIFO buffer and returning the start-sequence key. The extraneous read of the 2nd key makes us miss its corresponding stdin event, so the event loop blocks until a third key is pressed. This explains why such key sequences do not behave promptly in TUI. To fix this issue, we must somehow compensate for the missed stdin event corresponding to the 2nd byte of a key sequence. This patch achieves this by hacking up the stdin event handler to conditionally execute the readline callback multiple times in a row. This is done via a new global variable, call_stdin_event_handler_again_p, which is set from tui_getc() when we receive a start-sequence key and notice extra pending input in the ncurses buffer. Tested on x86_64-unknown-linux-gnu. gdb/ChangeLog: * event-top.h (call_stdin_event_handler_again_p): Declare. * event-top.c (call_stdin_event_handler_again_p): Define. (stdin_event_handler): Use it. * tui/tui-io.c (tui_getc): Prepare to call the stdin event handler again if there is pending input following a start sequence.
2014-11-23 11:03:39 +01:00
{
Don't set immediate_quit in prompt_for_continue immediate_quit used to be necessary back when prompt_for_continue used blocking fread, but nowadays it uses gdb_readline_wrapper, which is implemented in terms of a nested event loop, which already knows how to react to SIGINT: #0 throw_it (reason=RETURN_QUIT, error=GDB_NO_ERROR, fmt=0x9d6d7e "Quit", ap=0x7fffffffcb88) at .../src/gdb/common/common-exceptions.c:324 #1 0x00000000007bab5d in throw_vquit (fmt=0x9d6d7e "Quit", ap=0x7fffffffcb88) at .../src/gdb/common/common-exceptions.c:366 #2 0x00000000007bac9f in throw_quit (fmt=0x9d6d7e "Quit") at .../src/gdb/common/common-exceptions.c:385 #3 0x0000000000773a2d in quit () at .../src/gdb/utils.c:1039 #4 0x000000000065d81b in async_request_quit (arg=0x0) at .../src/gdb/event-top.c:893 #5 0x000000000065c27b in invoke_async_signal_handlers () at .../src/gdb/event-loop.c:949 #6 0x000000000065aeef in gdb_do_one_event () at .../src/gdb/event-loop.c:280 #7 0x0000000000770838 in gdb_readline_wrapper (prompt=0x7fffffffcd40 "---Type <return> to continue, or q <return> to quit---") at .../src/gdb/top.c:873 The need for the QUIT in stdin_event_handler is then exposed by the gdb.base/double-prompt-target-event-error.exp test, which has: # We're now stopped in a pagination query while handling a # target event (printing where the program stopped). Quitting # the pagination should result in only one prompt being # output. send_gdb "\003p 1\n" Without that change we'd get: Continuing. ---Type <return> to continue, or q <return> to quit---PASS: gdb.base/double-prompt-target-event-error.exp: ctrlc target event: continue: continue to pagination ^CpQuit (gdb) 1 Undefined command: "1". Try "help". (gdb) PASS: gdb.base/double-prompt-target-event-error.exp: ctrlc target event: continue: first prompt ERROR: Undefined command "". UNRESOLVED: gdb.base/double-prompt-target-event-error.exp: ctrlc target event: continue: no double prompt Vs: Continuing. ---Type <return> to continue, or q <return> to quit---PASS: gdb.base/double-prompt-target-event-error.exp: ctrlc target event: continue: continue to pagination ^CQuit (gdb) p 1 $1 = 1 (gdb) PASS: gdb.base/double-prompt-target-event-error.exp: ctrlc target event: continue: first prompt PASS: gdb.base/double-prompt-target-event-error.exp: ctrlc target event: continue: no double prompt gdb/ChangeLog: 2016-04-12 Pedro Alves <palves@redhat.com> * event-top.c (stdin_event_handler): Call QUIT; (prompt_for_continue): Don't run with immediate_quit set.
2016-04-12 17:49:30 +02:00
/* This makes sure a ^C immediately followed by further input is
always processed in that order. E.g,. with input like
"^Cprint 1\n", the SIGINT handler runs, marks the async signal
handler, and then select/poll may return with stdin ready,
instead of -1/EINTR. The
gdb.base/double-prompt-target-event-error.exp test exercises
this. */
QUIT;
Fix the processing of Meta-key commands in TUI This patch fixes the annoying bug where key sequences such as Alt_F or Alt_B (go forward or backwards by a word) do not behave promptly in TUI. You have to press a third key in order for the key sequence to register. This is mostly ncurses' fault. Calling wgetch() normally causes ncurses to read only a single key from stdin. However if the key read is the start-sequence key (^[ a.k.a. ESC) then wgetch() reads TWO keys from stdin, storing the 2nd key into an internal FIFO buffer and returning the start-sequence key. The extraneous read of the 2nd key makes us miss its corresponding stdin event, so the event loop blocks until a third key is pressed. This explains why such key sequences do not behave promptly in TUI. To fix this issue, we must somehow compensate for the missed stdin event corresponding to the 2nd byte of a key sequence. This patch achieves this by hacking up the stdin event handler to conditionally execute the readline callback multiple times in a row. This is done via a new global variable, call_stdin_event_handler_again_p, which is set from tui_getc() when we receive a start-sequence key and notice extra pending input in the ncurses buffer. Tested on x86_64-unknown-linux-gnu. gdb/ChangeLog: * event-top.h (call_stdin_event_handler_again_p): Declare. * event-top.c (call_stdin_event_handler_again_p): Define. (stdin_event_handler): Use it. * tui/tui-io.c (tui_getc): Prepare to call the stdin event handler again if there is pending input following a start sequence.
2014-11-23 11:03:39 +01:00
do
{
call_stdin_event_handler_again_p = 0;
(*call_readline) (client_data);
} while (call_stdin_event_handler_again_p != 0);
}
1999-09-22 05:28:34 +02:00
}
1999-09-28 23:55:21 +02:00
/* Re-enable stdin after the end of an execution command in
synchronous mode, or after an error from the target, and we aborted
the exec operation. */
1999-09-28 23:55:21 +02:00
void
async_enable_stdin (void)
1999-09-28 23:55:21 +02:00
{
if (sync_execution)
{
/* See NOTE in async_disable_stdin(). */
/* FIXME: cagney/1999-09-27: Call this before clearing
sync_execution. Current target_terminal_ours() implementations
check for sync_execution before switching the terminal. */
target_terminal_ours ();
sync_execution = 0;
}
1999-09-28 23:55:21 +02:00
}
/* Disable reads from stdin (the console) marking the command as
synchronous. */
1999-09-28 23:55:21 +02:00
void
async_disable_stdin (void)
{
sync_execution = 1;
1999-09-28 23:55:21 +02:00
}
1999-05-11 22:29:07 +02:00
1999-09-28 23:55:21 +02:00
/* Handle a gdb command line. This function is called when
handle_line_of_input has concatenated one or more input lines into
a whole command. */
void
1999-09-22 05:28:34 +02:00
command_handler (char *command)
1999-05-11 22:29:07 +02:00
{
struct cleanup *stat_chain;
char *c;
1999-05-11 22:29:07 +02:00
if (instream == stdin)
1999-05-11 22:29:07 +02:00
reinitialize_more_filter ();
stat_chain = make_command_stats_cleanup (1);
1999-05-11 22:29:07 +02:00
/* Do not execute commented lines. */
for (c = command; *c == ' ' || *c == '\t'; c++)
;
if (c[0] != '#')
{
execute_command (command, instream == stdin);
1999-07-07 22:19:36 +02:00
/* Do any commands attached to breakpoint we stopped at. */
bpstat_do_actions ();
}
1999-07-07 22:19:36 +02:00
do_cleanups (stat_chain);
1999-07-05 19:58:44 +02:00
}
/* Append RL, an input line returned by readline or one of its
emulations, to CMD_LINE_BUFFER. Returns the command line if we
have a whole command line ready to be processed by the command
interpreter or NULL if the command line isn't complete yet (input
line ends in a backslash). Takes ownership of RL. */
1999-05-11 22:29:07 +02:00
static char *
command_line_append_input_line (struct buffer *cmd_line_buffer, char *rl)
1999-05-11 22:29:07 +02:00
{
char *cmd;
size_t len;
1999-05-11 22:29:07 +02:00
len = strlen (rl);
1999-05-11 22:29:07 +02:00
if (len > 0 && rl[len - 1] == '\\')
1999-05-11 22:29:07 +02:00
{
/* Don't copy the backslash and wait for more. */
buffer_grow (cmd_line_buffer, rl, len - 1);
cmd = NULL;
1999-05-11 22:29:07 +02:00
}
else
1999-05-11 22:29:07 +02:00
{
/* Copy whole line including terminating null, and we're
done. */
buffer_grow (cmd_line_buffer, rl, len + 1);
cmd = cmd_line_buffer->buffer;
1999-05-11 22:29:07 +02:00
}
/* Allocated in readline. */
xfree (rl);
1999-05-11 22:29:07 +02:00
return cmd;
}
1999-05-11 22:29:07 +02:00
/* Handle a line of input coming from readline.
1999-05-11 22:29:07 +02:00
If the read line ends with a continuation character (backslash),
save the partial input in CMD_LINE_BUFFER (except the backslash),
and return NULL. Otherwise, save the partial input and return a
pointer to CMD_LINE_BUFFER's buffer (null terminated), indicating a
whole command line is ready to be executed.
1999-05-11 22:29:07 +02:00
Returns EOF on end of file.
1999-05-11 22:29:07 +02:00
If REPEAT, handle command repetitions:
1999-05-11 22:29:07 +02:00
- If the input command line is NOT empty, the command returned is
copied into the global 'saved_command_line' var so that it can
be repeated later.
- OTOH, if the input command line IS empty, return the previously
saved command instead of the empty input line.
*/
1999-05-11 22:29:07 +02:00
char *
handle_line_of_input (struct buffer *cmd_line_buffer,
char *rl, int repeat, char *annotation_suffix)
{
char *p1;
char *cmd;
if (rl == NULL)
return (char *) EOF;
cmd = command_line_append_input_line (cmd_line_buffer, rl);
if (cmd == NULL)
return NULL;
1999-05-11 22:29:07 +02:00
/* We have a complete command line now. Prepare for the next
command, but leave ownership of memory to the buffer . */
cmd_line_buffer->used_size = 0;
if (annotation_level > 1 && instream == stdin)
1999-05-11 22:29:07 +02:00
{
printf_unfiltered (("\n\032\032post-"));
puts_unfiltered (annotation_suffix);
printf_unfiltered (("\n"));
}
#define SERVER_COMMAND_PREFIX "server "
if (startswith (cmd, SERVER_COMMAND_PREFIX))
{
/* Note that we don't set `saved_command_line'. Between this
and the check in dont_repeat, this insures that repeating
will still do the right thing. */
return cmd + strlen (SERVER_COMMAND_PREFIX);
1999-05-11 22:29:07 +02:00
}
/* Do history expansion if that is wished. */
if (history_expansion_p && instream == stdin
&& ISATTY (instream))
{
char *history_value;
int expanded;
expanded = history_expand (cmd, &history_value);
1999-05-11 22:29:07 +02:00
if (expanded)
{
size_t len;
1999-05-11 22:29:07 +02:00
/* Print the changes. */
printf_unfiltered ("%s\n", history_value);
/* If there was an error, call this function again. */
if (expanded < 0)
{
2000-12-15 02:01:51 +01:00
xfree (history_value);
return cmd;
1999-05-11 22:29:07 +02:00
}
/* history_expand returns an allocated string. Just replace
our buffer with it. */
len = strlen (history_value);
xfree (buffer_finish (cmd_line_buffer));
cmd_line_buffer->buffer = history_value;
cmd_line_buffer->buffer_size = len + 1;
cmd = history_value;
1999-05-11 22:29:07 +02:00
}
}
/* If we just got an empty line, and that is supposed to repeat the
previous command, return the previously saved command. */
for (p1 = cmd; *p1 == ' ' || *p1 == '\t'; p1++)
;
if (repeat && *p1 == '\0')
return saved_command_line;
/* Add command to history if appropriate. Note: lines consisting
solely of comments are also added to the command history. This
is useful when you type a command, and then realize you don't
want to execute it quite yet. You can comment out the command
and then later fetch it from the value history and remove the
'#'. The kill ring is probably better, but some people are in
the habit of commenting things out. */
if (*cmd != '\0' && input_from_terminal_p ())
gdb_add_history (cmd);
1999-05-11 22:29:07 +02:00
/* Save into global buffer if appropriate. */
if (repeat)
1999-05-11 22:29:07 +02:00
{
xfree (saved_command_line);
saved_command_line = xstrdup (cmd);
return saved_command_line;
1999-05-11 22:29:07 +02:00
}
else
return cmd;
}
1999-05-11 22:29:07 +02:00
/* Handle a complete line of input. This is called by the callback
mechanism within the readline library. Deal with incomplete
commands as well, by saving the partial input in a global
buffer.
1999-05-11 22:29:07 +02:00
NOTE: This is the asynchronous version of the command_line_input
function. */
1999-05-11 22:29:07 +02:00
void
command_line_handler (char *rl)
{
struct buffer *line_buffer = get_command_line_buffer ();
char *cmd;
1999-05-11 22:29:07 +02:00
cmd = handle_line_of_input (line_buffer, rl, instream == stdin, "prompt");
if (cmd == (char *) EOF)
1999-05-11 22:29:07 +02:00
{
/* stdin closed. The connection with the terminal is gone.
This happens at the end of a testsuite run, after Expect has
hung up but GDB is still alive. In such a case, we just quit
gdb killing the inferior program too. */
printf_unfiltered ("quit\n");
execute_command ("quit", stdin == instream);
}
else if (cmd == NULL)
{
/* We don't have a full line yet. Print an empty prompt. */
display_gdb_prompt ("");
}
else
{
command_handler (cmd);
display_gdb_prompt (0);
1999-05-11 22:29:07 +02:00
}
}
/* Does reading of input from terminal w/o the editing features
provided by the readline library. Calls the line input handler
once we have a whole input line. */
1999-05-11 22:29:07 +02:00
1999-06-28 18:06:02 +02:00
void
gdb_readline_no_editing_callback (gdb_client_data client_data)
1999-05-11 22:29:07 +02:00
{
int c;
char *result;
struct buffer line_buffer;
1999-08-16 21:57:19 +02:00
static int done_once = 0;
buffer_init (&line_buffer);
1999-08-16 21:57:19 +02:00
/* Unbuffer the input stream, so that, later on, the calls to fgetc
fetch only one char at the time from the stream. The fgetc's will
1999-08-16 21:57:19 +02:00
get up to the first newline, but there may be more chars in the
stream after '\n'. If we buffer the input and fgetc drains the
1999-08-16 21:57:19 +02:00
stream, getting stuff beyond the newline as well, a select, done
afterwards will not trigger. */
1999-08-16 21:57:19 +02:00
if (!done_once && !ISATTY (instream))
{
setbuf (instream, NULL);
done_once = 1;
}
1999-05-11 22:29:07 +02:00
/* We still need the while loop here, even though it would seem
obvious to invoke gdb_readline_no_editing_callback at every
character entered. If not using the readline library, the
terminal is in cooked mode, which sends the characters all at
once. Poll will notice that the input fd has changed state only
after enter is pressed. At this point we still need to fetch all
the chars entered. */
1999-05-11 22:29:07 +02:00
while (1)
{
/* Read from stdin if we are executing a user defined command.
This is the right thing for prompt_for_continue, at least. */
c = fgetc (instream ? instream : stdin);
if (c == EOF)
{
if (line_buffer.used_size > 0)
{
/* The last line does not end with a newline. Return it, and
if we are called again fgetc will still return EOF and
we'll return NULL then. */
break;
}
xfree (buffer_finish (&line_buffer));
1999-06-14 20:08:47 +02:00
(*input_handler) (0);
return;
1999-05-11 22:29:07 +02:00
}
if (c == '\n')
{
if (line_buffer.used_size > 0
&& line_buffer.buffer[line_buffer.used_size - 1] == '\r')
line_buffer.used_size--;
1999-05-11 22:29:07 +02:00
break;
}
buffer_grow_char (&line_buffer, c);
1999-05-11 22:29:07 +02:00
}
buffer_grow_char (&line_buffer, '\0');
result = buffer_finish (&line_buffer);
1999-06-14 20:08:47 +02:00
(*input_handler) (result);
1999-05-11 22:29:07 +02:00
}
Introduce interruptible_select We have places where we call a blocking gdb_select expecting that a Ctrl-C will unblock it. However, if the Ctrl-C is pressed just before gdb_select, the SIGINT handler runs before gdb_select, and thus gdb_select won't return. For example gdb_readline_no_editing: QUIT; /* Wait until at least one byte of data is available. Control-C can interrupt gdb_select, but not fgetc. */ FD_ZERO (&readfds); FD_SET (fd, &readfds); if (gdb_select (fd + 1, &readfds, NULL, NULL, NULL) == -1) and stdio_file_read: /* For the benefit of Windows, call gdb_select before reading from the file. Wait until at least one byte of data is available. Control-C can interrupt gdb_select, but not read. */ { fd_set readfds; FD_ZERO (&readfds); FD_SET (stdio->fd, &readfds); if (gdb_select (stdio->fd + 1, &readfds, NULL, NULL, NULL) == -1) return -1; } return read (stdio->fd, buf, length_buf); This is a race classically fixed with either the self-pipe trick, or by blocking SIGINT and then using pselect instead of select. Blocking SIGINT most of the time would mean that check_quit_flag (and thus QUIT) would need to do a syscall every time it is called, which sounds best avoided, since QUIT is called in many loops. Thus we take the self-pipe trick route (wrapped in a serial event). Instead of having all places that need this manually add an extra file descriptor to the set of gdb_select's watched file descriptors, we introduce a wrapper, interruptible_select, that does that. The Windows version of gdb_select actually does not suffer from this, because mingw-hdep.c:gdb_call_async_signal_handler sets a Windows event that gdb_select always waits on. So this patch can be seen as generalization of that technique. We can't remove that extra event from mingw-hdep.c until we get rid of immediate_quit though. gdb/ChangeLog: 2016-04-12 Pedro Alves <palves@redhat.com> * defs.h: Extend QUIT-related comments to mention interruptible_select. (quit_serial_event_set, quit_serial_event_clear): Declare. * event-top.c: Include "ser-event.h" and "gdb_select.h". (quit_serial_event): New global. (async_init_signals): Make quit_serial_event. (quit_serial_event_set, quit_serial_event_clear) (quit_serial_event_fd, interruptible_select): New functions. * extension.c (set_quit_flag): Set the quit serial event. (check_quit_flag): Clear the quit serial event. * gdb_select.h (interruptible_select): New declaration. * guile/scm-ports.c (ioscm_input_waiting): Use interruptible_select instead of gdb_select. * top.c (gdb_readline_no_editing): Likewise. * ui-file.c (stdio_file_read): Likewise.
2016-04-12 17:49:30 +02:00
/* The serial event associated with the QUIT flag. set_quit_flag sets
this, and check_quit_flag clears it. Used by interruptible_select
to be able to do interruptible I/O with no race with the SIGINT
handler. */
static struct serial_event *quit_serial_event;
1999-05-11 22:29:07 +02:00
/* Initialization of signal handlers and tokens. There is a function
handle_sig* for each of the signals GDB cares about. Specifically:
1999-05-11 22:29:07 +02:00
SIGINT, SIGFPE, SIGQUIT, SIGTSTP, SIGHUP, SIGWINCH. These
functions are the actual signal handlers associated to the signals
via calls to signal(). The only job for these functions is to
enqueue the appropriate event/procedure with the event loop. Such
procedures are the old signal handlers. The event loop will take
1999-05-11 22:29:07 +02:00
care of invoking the queued procedures to perform the usual tasks
associated with the reception of the signal. */
1999-05-25 20:09:09 +02:00
/* NOTE: 1999-04-30 This is the asynchronous version of init_signals.
1999-05-11 22:29:07 +02:00
init_signals will become obsolete as we move to have to event loop
as the default for gdb. */
1999-05-11 22:29:07 +02:00
void
1999-09-22 05:28:34 +02:00
async_init_signals (void)
1999-07-07 22:19:36 +02:00
{
Fix signal handler/event-loop races GDB's core signal handling suffers from a classical signal handler / mainline code race: int gdb_do_one_event (void) { ... /* First let's see if there are any asynchronous signal handlers that are ready. These would be the result of invoking any of the signal handlers. */ if (invoke_async_signal_handlers ()) return 1; ... /* Block waiting for a new event. (...). */ if (gdb_wait_for_event (1) < 0) return -1; ... } If a signal is delivered while gdb is blocked in the poll/select inside gdb_wait_for_event, then the select/poll breaks with EINTR, we'll loop back around and call invoke_async_signal_handlers. However, if the signal handler runs between invoke_async_signal_handlers and gdb_wait_for_event, gdb_wait_for_event will block, until the next unrelated event... The fix is to a struct serial_event, and register it in the set of files that select/poll in gdb_wait_for_event waits on. The signal handlers that defer work to invoke_async_signal_handlers call mark_async_signal_handler, which is adjusted to also set the new serial event in addition to setting a flag, and is thus now is garanteed to immediately unblock the next gdb_select/poll call, up until invoke_async_signal_handlers is called and the event is cleared. gdb/ChangeLog: 2016-04-12 Pedro Alves <palves@redhat.com> * event-loop.c: Include "ser-event.h". (async_signal_handlers_serial_event): New global. (async_signals_handler, initialize_async_signal_handlers): New functions. (mark_async_signal_handler): Set async_signal_handlers_serial_event. (invoke_async_signal_handlers): Clear async_signal_handlers_serial_event. * event-top.c (async_init_signals): Call initialize_async_signal_handlers.
2016-04-12 17:49:30 +02:00
initialize_async_signal_handlers ();
Introduce interruptible_select We have places where we call a blocking gdb_select expecting that a Ctrl-C will unblock it. However, if the Ctrl-C is pressed just before gdb_select, the SIGINT handler runs before gdb_select, and thus gdb_select won't return. For example gdb_readline_no_editing: QUIT; /* Wait until at least one byte of data is available. Control-C can interrupt gdb_select, but not fgetc. */ FD_ZERO (&readfds); FD_SET (fd, &readfds); if (gdb_select (fd + 1, &readfds, NULL, NULL, NULL) == -1) and stdio_file_read: /* For the benefit of Windows, call gdb_select before reading from the file. Wait until at least one byte of data is available. Control-C can interrupt gdb_select, but not read. */ { fd_set readfds; FD_ZERO (&readfds); FD_SET (stdio->fd, &readfds); if (gdb_select (stdio->fd + 1, &readfds, NULL, NULL, NULL) == -1) return -1; } return read (stdio->fd, buf, length_buf); This is a race classically fixed with either the self-pipe trick, or by blocking SIGINT and then using pselect instead of select. Blocking SIGINT most of the time would mean that check_quit_flag (and thus QUIT) would need to do a syscall every time it is called, which sounds best avoided, since QUIT is called in many loops. Thus we take the self-pipe trick route (wrapped in a serial event). Instead of having all places that need this manually add an extra file descriptor to the set of gdb_select's watched file descriptors, we introduce a wrapper, interruptible_select, that does that. The Windows version of gdb_select actually does not suffer from this, because mingw-hdep.c:gdb_call_async_signal_handler sets a Windows event that gdb_select always waits on. So this patch can be seen as generalization of that technique. We can't remove that extra event from mingw-hdep.c until we get rid of immediate_quit though. gdb/ChangeLog: 2016-04-12 Pedro Alves <palves@redhat.com> * defs.h: Extend QUIT-related comments to mention interruptible_select. (quit_serial_event_set, quit_serial_event_clear): Declare. * event-top.c: Include "ser-event.h" and "gdb_select.h". (quit_serial_event): New global. (async_init_signals): Make quit_serial_event. (quit_serial_event_set, quit_serial_event_clear) (quit_serial_event_fd, interruptible_select): New functions. * extension.c (set_quit_flag): Set the quit serial event. (check_quit_flag): Clear the quit serial event. * gdb_select.h (interruptible_select): New declaration. * guile/scm-ports.c (ioscm_input_waiting): Use interruptible_select instead of gdb_select. * top.c (gdb_readline_no_editing): Likewise. * ui-file.c (stdio_file_read): Likewise.
2016-04-12 17:49:30 +02:00
quit_serial_event = make_serial_event ();
1999-05-11 22:29:07 +02:00
signal (SIGINT, handle_sigint);
sigint_token =
1999-06-14 20:08:47 +02:00
create_async_signal_handler (async_request_quit, NULL);
signal (SIGTERM, handle_sigterm);
async_sigterm_token
= create_async_signal_handler (async_sigterm_handler, NULL);
1999-05-11 22:29:07 +02:00
/* If SIGTRAP was set to SIG_IGN, then the SIG_IGN will get passed
to the inferior and breakpoints will be ignored. */
#ifdef SIGTRAP
signal (SIGTRAP, SIG_DFL);
#endif
#ifdef SIGQUIT
1999-05-11 22:29:07 +02:00
/* If we initialize SIGQUIT to SIG_IGN, then the SIG_IGN will get
passed to the inferior, which we don't want. It would be
possible to do a "signal (SIGQUIT, SIG_DFL)" after we fork, but
on BSD4.3 systems using vfork, that can affect the
GDB process as well as the inferior (the signal handling tables
might be in memory, shared between the two). Since we establish
a handler for SIGQUIT, when we call exec it will set the signal
to SIG_DFL for us. */
signal (SIGQUIT, handle_sigquit);
sigquit_token =
1999-06-14 20:08:47 +02:00
create_async_signal_handler (async_do_nothing, NULL);
#endif
1999-05-11 22:29:07 +02:00
#ifdef SIGHUP
if (signal (SIGHUP, handle_sighup) != SIG_IGN)
sighup_token =
1999-06-14 20:08:47 +02:00
create_async_signal_handler (async_disconnect, NULL);
1999-05-11 22:29:07 +02:00
else
sighup_token =
1999-06-14 20:08:47 +02:00
create_async_signal_handler (async_do_nothing, NULL);
1999-05-11 22:29:07 +02:00
#endif
signal (SIGFPE, handle_sigfpe);
sigfpe_token =
1999-06-14 20:08:47 +02:00
create_async_signal_handler (async_float_handler, NULL);
1999-05-11 22:29:07 +02:00
1999-06-14 20:08:47 +02:00
#ifdef STOP_SIGNAL
sigtstp_token =
create_async_signal_handler (async_stop_sig, NULL);
#endif
}
Introduce interruptible_select We have places where we call a blocking gdb_select expecting that a Ctrl-C will unblock it. However, if the Ctrl-C is pressed just before gdb_select, the SIGINT handler runs before gdb_select, and thus gdb_select won't return. For example gdb_readline_no_editing: QUIT; /* Wait until at least one byte of data is available. Control-C can interrupt gdb_select, but not fgetc. */ FD_ZERO (&readfds); FD_SET (fd, &readfds); if (gdb_select (fd + 1, &readfds, NULL, NULL, NULL) == -1) and stdio_file_read: /* For the benefit of Windows, call gdb_select before reading from the file. Wait until at least one byte of data is available. Control-C can interrupt gdb_select, but not read. */ { fd_set readfds; FD_ZERO (&readfds); FD_SET (stdio->fd, &readfds); if (gdb_select (stdio->fd + 1, &readfds, NULL, NULL, NULL) == -1) return -1; } return read (stdio->fd, buf, length_buf); This is a race classically fixed with either the self-pipe trick, or by blocking SIGINT and then using pselect instead of select. Blocking SIGINT most of the time would mean that check_quit_flag (and thus QUIT) would need to do a syscall every time it is called, which sounds best avoided, since QUIT is called in many loops. Thus we take the self-pipe trick route (wrapped in a serial event). Instead of having all places that need this manually add an extra file descriptor to the set of gdb_select's watched file descriptors, we introduce a wrapper, interruptible_select, that does that. The Windows version of gdb_select actually does not suffer from this, because mingw-hdep.c:gdb_call_async_signal_handler sets a Windows event that gdb_select always waits on. So this patch can be seen as generalization of that technique. We can't remove that extra event from mingw-hdep.c until we get rid of immediate_quit though. gdb/ChangeLog: 2016-04-12 Pedro Alves <palves@redhat.com> * defs.h: Extend QUIT-related comments to mention interruptible_select. (quit_serial_event_set, quit_serial_event_clear): Declare. * event-top.c: Include "ser-event.h" and "gdb_select.h". (quit_serial_event): New global. (async_init_signals): Make quit_serial_event. (quit_serial_event_set, quit_serial_event_clear) (quit_serial_event_fd, interruptible_select): New functions. * extension.c (set_quit_flag): Set the quit serial event. (check_quit_flag): Clear the quit serial event. * gdb_select.h (interruptible_select): New declaration. * guile/scm-ports.c (ioscm_input_waiting): Use interruptible_select instead of gdb_select. * top.c (gdb_readline_no_editing): Likewise. * ui-file.c (stdio_file_read): Likewise.
2016-04-12 17:49:30 +02:00
/* See defs.h. */
void
quit_serial_event_set (void)
{
serial_event_set (quit_serial_event);
}
/* See defs.h. */
void
quit_serial_event_clear (void)
{
serial_event_clear (quit_serial_event);
}
/* Return the selectable file descriptor of the serial event
associated with the quit flag. */
static int
quit_serial_event_fd (void)
{
return serial_event_fd (quit_serial_event);
}
target remote: Don't rely on immediate_quit (introduce quit handlers) remote.c is the last user of immediate_quit. It's relied on to immediately break the initial remote connection sync up, if the user does Ctrl-C, assuming that was because the target isn't responding. At that stage, since the connection isn't synced yet, disconnecting is the only safe thing to do. This commit reworks that, to not rely on throwing from the SIGINT signal handler. So, this commit: - Introduces the concept of a "quit handler". This is used to override what does the QUIT macro do when the quit flag is set. - Makes the "struct serial" reachar / write code call QUIT in the partial read/write loops, so the current quit handler is invoked whenever a serial->read_prim / serial->write_prim returns EINTR. - Makes the "struct serial" reachar / write code call interruptible_select instead of gdb_select, so that QUITs are detected in a race-free manner. - Stops remote.c from setting immediate_quit during the initial connection. - Instead, we install a custom quit handler whenever we're calling into the serial code. This custom quit handler knows to immediately throw a quit when we're in the initial connection setup, and otherwise defer handling the quit/Ctrl-C request to later, when we're safely out of a packet command/response sequence. This also is what is now responsible for handling "double Ctrl-C because target connection is stuck/wedged." - remote.c no longer installs a specialized SIGINT handlers, and instead re-uses the quit flag. Since we want to rely on the QUIT macro, the SIGINT handler must also set the quit. And the easiest is just to not install custom SIGINT handler in remote.c. Let the standard SIGINT handler do its job of setting the quit flag. Centralizing SIGINT handlers seems like a good thing to me, anyway. gdb/ChangeLog: 2016-04-12 Pedro Alves <palves@redhat.com> * defs.h (quit_handler_ftype, quit_handler) (make_cleanup_override_quit_handler, default_quit_handler): New. (QUIT): Adjust comments. * event-top.c (default_quit_handler): New function. (quit_handler): New global. (struct quit_handler_cleanup_data): New. (restore_quit_handler, restore_quit_handler_dtor) (make_cleanup_override_quit_handler): New. (async_request_quit): Call QUIT. * remote.c (struct remote_state) <got_ctrlc_during_io>: New field. (async_sigint_remote_twice_token, async_sigint_remote_token): Delete. (remote_close): Update comments. (remote_start_remote): Don't set immediate_quit. Set starting_up earlier. (remote_serial_quit_handler, remote_unpush_and_throw): New functions. (remote_open_1): Clear got_ctrlc_during_io. Set remote_async_terminal_ours_p unconditionally. (async_initialize_sigint_signal_handler) (async_handle_remote_sigint, async_handle_remote_sigint_twice) (remote_check_pending_interrupt, async_remote_interrupt) (async_remote_interrupt_twice) (async_cleanup_sigint_signal_handler, ofunc) (sync_remote_interrupt, sync_remote_interrupt_twice): Delete. (remote_terminal_inferior, remote_terminal_ours): Remove async checks. (remote_wait_as): Don't install a SIGINT handler in sync mode. (readchar, remote_serial_write): Override the quit handler with remote_serial_quit_handler. (getpkt_or_notif_sane_1): Don't call QUIT. (initialize_remote_ops): Don't install remote_check_pending_interrupt. (_initialize_remote): Don't create async_sigint_remote_token and async_sigint_remote_twice_token. * ser-base.c (ser_base_wait_for): Call QUIT and use interruptible_select. (ser_base_write): Call QUIT. * ser-go32.c (dos_readchar, dos_write): Call QUIT. * ser-unix.c (wait_for): Don't use VTIME. Always take the gdb_select path, but call QUIT and interruptible_select. * utils.c (maybe_quit): Call the current quit handler. Don't call target_check_pending_interrupt. (defaulted_query, prompt_for_continue): Override the quit handler with the default quit handler.
2016-04-12 17:49:32 +02:00
/* See defs.h. */
void
default_quit_handler (void)
{
if (check_quit_flag ())
{
if (target_terminal_is_ours ())
quit ();
else
target_pass_ctrlc ();
}
}
/* See defs.h. */
quit_handler_ftype *quit_handler = default_quit_handler;
/* Data for make_cleanup_override_quit_handler. Wrap the previous
handler pointer in a data struct because it's not portable to cast
a function pointer to a data pointer, which is what make_cleanup
expects. */
struct quit_handler_cleanup_data
{
/* The previous quit handler. */
quit_handler_ftype *prev_handler;
};
/* Cleanup call that restores the previous quit handler. */
static void
restore_quit_handler (void *arg)
{
struct quit_handler_cleanup_data *data
= (struct quit_handler_cleanup_data *) arg;
quit_handler = data->prev_handler;
}
/* Destructor for the quit handler cleanup. */
static void
restore_quit_handler_dtor (void *arg)
{
xfree (arg);
}
/* See defs.h. */
struct cleanup *
make_cleanup_override_quit_handler (quit_handler_ftype *new_quit_handler)
{
struct cleanup *old_chain;
struct quit_handler_cleanup_data *data;
data = XNEW (struct quit_handler_cleanup_data);
data->prev_handler = quit_handler;
old_chain = make_cleanup_dtor (restore_quit_handler, data,
restore_quit_handler_dtor);
quit_handler = new_quit_handler;
return old_chain;
}
Introduce interruptible_select We have places where we call a blocking gdb_select expecting that a Ctrl-C will unblock it. However, if the Ctrl-C is pressed just before gdb_select, the SIGINT handler runs before gdb_select, and thus gdb_select won't return. For example gdb_readline_no_editing: QUIT; /* Wait until at least one byte of data is available. Control-C can interrupt gdb_select, but not fgetc. */ FD_ZERO (&readfds); FD_SET (fd, &readfds); if (gdb_select (fd + 1, &readfds, NULL, NULL, NULL) == -1) and stdio_file_read: /* For the benefit of Windows, call gdb_select before reading from the file. Wait until at least one byte of data is available. Control-C can interrupt gdb_select, but not read. */ { fd_set readfds; FD_ZERO (&readfds); FD_SET (stdio->fd, &readfds); if (gdb_select (stdio->fd + 1, &readfds, NULL, NULL, NULL) == -1) return -1; } return read (stdio->fd, buf, length_buf); This is a race classically fixed with either the self-pipe trick, or by blocking SIGINT and then using pselect instead of select. Blocking SIGINT most of the time would mean that check_quit_flag (and thus QUIT) would need to do a syscall every time it is called, which sounds best avoided, since QUIT is called in many loops. Thus we take the self-pipe trick route (wrapped in a serial event). Instead of having all places that need this manually add an extra file descriptor to the set of gdb_select's watched file descriptors, we introduce a wrapper, interruptible_select, that does that. The Windows version of gdb_select actually does not suffer from this, because mingw-hdep.c:gdb_call_async_signal_handler sets a Windows event that gdb_select always waits on. So this patch can be seen as generalization of that technique. We can't remove that extra event from mingw-hdep.c until we get rid of immediate_quit though. gdb/ChangeLog: 2016-04-12 Pedro Alves <palves@redhat.com> * defs.h: Extend QUIT-related comments to mention interruptible_select. (quit_serial_event_set, quit_serial_event_clear): Declare. * event-top.c: Include "ser-event.h" and "gdb_select.h". (quit_serial_event): New global. (async_init_signals): Make quit_serial_event. (quit_serial_event_set, quit_serial_event_clear) (quit_serial_event_fd, interruptible_select): New functions. * extension.c (set_quit_flag): Set the quit serial event. (check_quit_flag): Clear the quit serial event. * gdb_select.h (interruptible_select): New declaration. * guile/scm-ports.c (ioscm_input_waiting): Use interruptible_select instead of gdb_select. * top.c (gdb_readline_no_editing): Likewise. * ui-file.c (stdio_file_read): Likewise.
2016-04-12 17:49:30 +02:00
/* Handle a SIGINT. */
1999-07-07 22:19:36 +02:00
void
1999-09-22 05:28:34 +02:00
handle_sigint (int sig)
1999-05-11 22:29:07 +02:00
{
signal (sig, handle_sigint);
/* We could be running in a loop reading in symfiles or something so
it may be quite a while before we get back to the event loop. So
set quit_flag to 1 here. Then if QUIT is called before we get to
the event loop, we will unwind as expected. */
set_quit_flag ();
1999-05-11 22:29:07 +02:00
/* If immediate_quit is set, we go ahead and process the SIGINT right
away, even if we usually would defer this to the event loop. The
1999-05-11 22:29:07 +02:00
assumption here is that it is safe to process ^C immediately if
immediate_quit is set. If we didn't, SIGINT would be really
1999-05-11 22:29:07 +02:00
processed only the next time through the event loop. To get to
that point, though, the command that we want to interrupt needs to
finish first, which is unacceptable. If immediate quit is not set,
we process SIGINT the next time through the loop, which is fine. */
gdb_call_async_signal_handler (sigint_token, immediate_quit);
1999-05-11 22:29:07 +02:00
}
Introduce interruptible_select We have places where we call a blocking gdb_select expecting that a Ctrl-C will unblock it. However, if the Ctrl-C is pressed just before gdb_select, the SIGINT handler runs before gdb_select, and thus gdb_select won't return. For example gdb_readline_no_editing: QUIT; /* Wait until at least one byte of data is available. Control-C can interrupt gdb_select, but not fgetc. */ FD_ZERO (&readfds); FD_SET (fd, &readfds); if (gdb_select (fd + 1, &readfds, NULL, NULL, NULL) == -1) and stdio_file_read: /* For the benefit of Windows, call gdb_select before reading from the file. Wait until at least one byte of data is available. Control-C can interrupt gdb_select, but not read. */ { fd_set readfds; FD_ZERO (&readfds); FD_SET (stdio->fd, &readfds); if (gdb_select (stdio->fd + 1, &readfds, NULL, NULL, NULL) == -1) return -1; } return read (stdio->fd, buf, length_buf); This is a race classically fixed with either the self-pipe trick, or by blocking SIGINT and then using pselect instead of select. Blocking SIGINT most of the time would mean that check_quit_flag (and thus QUIT) would need to do a syscall every time it is called, which sounds best avoided, since QUIT is called in many loops. Thus we take the self-pipe trick route (wrapped in a serial event). Instead of having all places that need this manually add an extra file descriptor to the set of gdb_select's watched file descriptors, we introduce a wrapper, interruptible_select, that does that. The Windows version of gdb_select actually does not suffer from this, because mingw-hdep.c:gdb_call_async_signal_handler sets a Windows event that gdb_select always waits on. So this patch can be seen as generalization of that technique. We can't remove that extra event from mingw-hdep.c until we get rid of immediate_quit though. gdb/ChangeLog: 2016-04-12 Pedro Alves <palves@redhat.com> * defs.h: Extend QUIT-related comments to mention interruptible_select. (quit_serial_event_set, quit_serial_event_clear): Declare. * event-top.c: Include "ser-event.h" and "gdb_select.h". (quit_serial_event): New global. (async_init_signals): Make quit_serial_event. (quit_serial_event_set, quit_serial_event_clear) (quit_serial_event_fd, interruptible_select): New functions. * extension.c (set_quit_flag): Set the quit serial event. (check_quit_flag): Clear the quit serial event. * gdb_select.h (interruptible_select): New declaration. * guile/scm-ports.c (ioscm_input_waiting): Use interruptible_select instead of gdb_select. * top.c (gdb_readline_no_editing): Likewise. * ui-file.c (stdio_file_read): Likewise.
2016-04-12 17:49:30 +02:00
/* See gdb_select.h. */
int
interruptible_select (int n,
fd_set *readfds, fd_set *writefds, fd_set *exceptfds,
struct timeval *timeout)
{
fd_set my_readfds;
int fd;
int res;
if (readfds == NULL)
{
readfds = &my_readfds;
FD_ZERO (&my_readfds);
}
fd = quit_serial_event_fd ();
FD_SET (fd, readfds);
if (n <= fd)
n = fd + 1;
do
{
res = gdb_select (n, readfds, writefds, exceptfds, timeout);
}
while (res == -1 && errno == EINTR);
if (res == 1 && FD_ISSET (fd, readfds))
{
errno = EINTR;
return -1;
}
return res;
}
/* Handle GDB exit upon receiving SIGTERM if target_can_async_p (). */
static void
async_sigterm_handler (gdb_client_data arg)
{
quit_force (NULL, stdin == instream);
}
/* See defs.h. */
volatile int sync_quit_force_run;
/* Quit GDB if SIGTERM is received.
GDB would quit anyway, but this way it will clean up properly. */
void
handle_sigterm (int sig)
{
signal (sig, handle_sigterm);
sync_quit_force_run = 1;
set_quit_flag ();
mark_async_signal_handler (async_sigterm_token);
}
/* Do the quit. All the checks have been done by the caller. */
1999-07-07 22:19:36 +02:00
void
1999-09-22 05:28:34 +02:00
async_request_quit (gdb_client_data arg)
1999-05-11 22:29:07 +02:00
{
/* If the quit_flag has gotten reset back to 0 by the time we get
back here, that means that an exception was thrown to unwind the
current command before we got back to the event loop. So there
is no reason to call quit again here. */
target remote: Don't rely on immediate_quit (introduce quit handlers) remote.c is the last user of immediate_quit. It's relied on to immediately break the initial remote connection sync up, if the user does Ctrl-C, assuming that was because the target isn't responding. At that stage, since the connection isn't synced yet, disconnecting is the only safe thing to do. This commit reworks that, to not rely on throwing from the SIGINT signal handler. So, this commit: - Introduces the concept of a "quit handler". This is used to override what does the QUIT macro do when the quit flag is set. - Makes the "struct serial" reachar / write code call QUIT in the partial read/write loops, so the current quit handler is invoked whenever a serial->read_prim / serial->write_prim returns EINTR. - Makes the "struct serial" reachar / write code call interruptible_select instead of gdb_select, so that QUITs are detected in a race-free manner. - Stops remote.c from setting immediate_quit during the initial connection. - Instead, we install a custom quit handler whenever we're calling into the serial code. This custom quit handler knows to immediately throw a quit when we're in the initial connection setup, and otherwise defer handling the quit/Ctrl-C request to later, when we're safely out of a packet command/response sequence. This also is what is now responsible for handling "double Ctrl-C because target connection is stuck/wedged." - remote.c no longer installs a specialized SIGINT handlers, and instead re-uses the quit flag. Since we want to rely on the QUIT macro, the SIGINT handler must also set the quit. And the easiest is just to not install custom SIGINT handler in remote.c. Let the standard SIGINT handler do its job of setting the quit flag. Centralizing SIGINT handlers seems like a good thing to me, anyway. gdb/ChangeLog: 2016-04-12 Pedro Alves <palves@redhat.com> * defs.h (quit_handler_ftype, quit_handler) (make_cleanup_override_quit_handler, default_quit_handler): New. (QUIT): Adjust comments. * event-top.c (default_quit_handler): New function. (quit_handler): New global. (struct quit_handler_cleanup_data): New. (restore_quit_handler, restore_quit_handler_dtor) (make_cleanup_override_quit_handler): New. (async_request_quit): Call QUIT. * remote.c (struct remote_state) <got_ctrlc_during_io>: New field. (async_sigint_remote_twice_token, async_sigint_remote_token): Delete. (remote_close): Update comments. (remote_start_remote): Don't set immediate_quit. Set starting_up earlier. (remote_serial_quit_handler, remote_unpush_and_throw): New functions. (remote_open_1): Clear got_ctrlc_during_io. Set remote_async_terminal_ours_p unconditionally. (async_initialize_sigint_signal_handler) (async_handle_remote_sigint, async_handle_remote_sigint_twice) (remote_check_pending_interrupt, async_remote_interrupt) (async_remote_interrupt_twice) (async_cleanup_sigint_signal_handler, ofunc) (sync_remote_interrupt, sync_remote_interrupt_twice): Delete. (remote_terminal_inferior, remote_terminal_ours): Remove async checks. (remote_wait_as): Don't install a SIGINT handler in sync mode. (readchar, remote_serial_write): Override the quit handler with remote_serial_quit_handler. (getpkt_or_notif_sane_1): Don't call QUIT. (initialize_remote_ops): Don't install remote_check_pending_interrupt. (_initialize_remote): Don't create async_sigint_remote_token and async_sigint_remote_twice_token. * ser-base.c (ser_base_wait_for): Call QUIT and use interruptible_select. (ser_base_write): Call QUIT. * ser-go32.c (dos_readchar, dos_write): Call QUIT. * ser-unix.c (wait_for): Don't use VTIME. Always take the gdb_select path, but call QUIT and interruptible_select. * utils.c (maybe_quit): Call the current quit handler. Don't call target_check_pending_interrupt. (defaulted_query, prompt_for_continue): Override the quit handler with the default quit handler.
2016-04-12 17:49:32 +02:00
QUIT;
1999-05-11 22:29:07 +02:00
}
#ifdef SIGQUIT
/* Tell the event loop what to do if SIGQUIT is received.
See event-signal.c. */
1999-07-07 22:19:36 +02:00
static void
1999-09-22 05:28:34 +02:00
handle_sigquit (int sig)
1999-05-11 22:29:07 +02:00
{
mark_async_signal_handler (sigquit_token);
1999-05-11 22:29:07 +02:00
signal (sig, handle_sigquit);
}
#endif
1999-05-11 22:29:07 +02:00
#if defined (SIGQUIT) || defined (SIGHUP)
/* Called by the event loop in response to a SIGQUIT or an
ignored SIGHUP. */
1999-07-07 22:19:36 +02:00
static void
1999-09-22 05:28:34 +02:00
async_do_nothing (gdb_client_data arg)
1999-05-11 22:29:07 +02:00
{
/* Empty function body. */
1999-05-11 22:29:07 +02:00
}
#endif
1999-05-11 22:29:07 +02:00
#ifdef SIGHUP
/* Tell the event loop what to do if SIGHUP is received.
See event-signal.c. */
1999-07-07 22:19:36 +02:00
static void
2000-07-30 03:48:28 +02:00
handle_sighup (int sig)
1999-05-11 22:29:07 +02:00
{
mark_async_signal_handler (sighup_token);
1999-05-11 22:29:07 +02:00
signal (sig, handle_sighup);
}
/* Called by the event loop to process a SIGHUP. */
1999-07-07 22:19:36 +02:00
static void
1999-09-22 05:28:34 +02:00
async_disconnect (gdb_client_data arg)
1999-05-11 22:29:07 +02:00
{
Split TRY_CATCH into TRY + CATCH This patch splits the TRY_CATCH macro into three, so that we go from this: ~~~ volatile gdb_exception ex; TRY_CATCH (ex, RETURN_MASK_ERROR) { } if (ex.reason < 0) { } ~~~ to this: ~~~ TRY { } CATCH (ex, RETURN_MASK_ERROR) { } END_CATCH ~~~ Thus, we'll be getting rid of the local volatile exception object, and declaring the caught exception in the catch block. This allows reimplementing TRY/CATCH in terms of C++ exceptions when building in C++ mode, while still allowing to build GDB in C mode (using setjmp/longjmp), as a transition step. TBC, after this patch, is it _not_ valid to have code between the TRY and the CATCH blocks, like: TRY { } // some code here. CATCH (ex, RETURN_MASK_ERROR) { } END_CATCH Just like it isn't valid to do that with C++'s native try/catch. By switching to creating the exception object inside the CATCH block scope, we can get rid of all the explicitly allocated volatile exception objects all over the tree, and map the CATCH block more directly to C++'s catch blocks. The majority of the TRY_CATCH -> TRY+CATCH+END_CATCH conversion was done with a script, rerun from scratch at every rebase, no manual editing involved. After the mechanical conversion, a few places needed manual intervention, to fix preexisting cases where we were using the exception object outside of the TRY_CATCH block, and cases where we were using "else" after a 'if (ex.reason) < 0)' [a CATCH after this patch]. The result was folded into this patch so that GDB still builds at each incremental step. END_CATCH is necessary for two reasons: First, because we name the exception object in the CATCH block, which requires creating a scope, which in turn must be closed somewhere. Declaring the exception variable in the initializer field of a for block, like: #define CATCH(EXCEPTION, mask) \ for (struct gdb_exception EXCEPTION; \ exceptions_state_mc_catch (&EXCEPTION, MASK); \ EXCEPTION = exception_none) would avoid needing END_CATCH, but alas, in C mode, we build with C90, which doesn't allow mixed declarations and code. Second, because when TRY/CATCH are wired to real C++ try/catch, as long as we need to handle cleanup chains, even if there's no CATCH block that wants to catch the exception, we need for stop at every frame in the unwind chain and run cleanups, then rethrow. That will be done in END_CATCH. After we require C++, we'll still need TRY/CATCH/END_CATCH until cleanups are completely phased out -- TRY/CATCH in C++ mode will save/restore the current cleanup chain, like in C mode, and END_CATCH catches otherwise uncaugh exceptions, runs cleanups and rethrows, so that C++ cleanups and exceptions can coexist. IMO, this still makes the TRY/CATCH code look a bit more like a newcomer would expect, so IMO worth it even if we weren't considering C++. gdb/ChangeLog. 2015-03-07 Pedro Alves <palves@redhat.com> * common/common-exceptions.c (struct catcher) <exception>: No longer a pointer to volatile exception. Now an exception value. <mask>: Delete field. (exceptions_state_mc_init): Remove all parameters. Adjust. (exceptions_state_mc): No longer pop the catcher here. (exceptions_state_mc_catch): New function. (throw_exception): Adjust. * common/common-exceptions.h (exceptions_state_mc_init): Remove all parameters. (exceptions_state_mc_catch): Declare. (TRY_CATCH): Rename to ... (TRY): ... this. Remove EXCEPTION and MASK parameters. (CATCH, END_CATCH): New. All callers adjusted. gdb/gdbserver/ChangeLog: 2015-03-07 Pedro Alves <palves@redhat.com> Adjust all callers of TRY_CATCH to use TRY/CATCH/END_CATCH instead.
2015-03-07 16:14:14 +01:00
TRY
{
quit_cover ();
}
Split TRY_CATCH into TRY + CATCH This patch splits the TRY_CATCH macro into three, so that we go from this: ~~~ volatile gdb_exception ex; TRY_CATCH (ex, RETURN_MASK_ERROR) { } if (ex.reason < 0) { } ~~~ to this: ~~~ TRY { } CATCH (ex, RETURN_MASK_ERROR) { } END_CATCH ~~~ Thus, we'll be getting rid of the local volatile exception object, and declaring the caught exception in the catch block. This allows reimplementing TRY/CATCH in terms of C++ exceptions when building in C++ mode, while still allowing to build GDB in C mode (using setjmp/longjmp), as a transition step. TBC, after this patch, is it _not_ valid to have code between the TRY and the CATCH blocks, like: TRY { } // some code here. CATCH (ex, RETURN_MASK_ERROR) { } END_CATCH Just like it isn't valid to do that with C++'s native try/catch. By switching to creating the exception object inside the CATCH block scope, we can get rid of all the explicitly allocated volatile exception objects all over the tree, and map the CATCH block more directly to C++'s catch blocks. The majority of the TRY_CATCH -> TRY+CATCH+END_CATCH conversion was done with a script, rerun from scratch at every rebase, no manual editing involved. After the mechanical conversion, a few places needed manual intervention, to fix preexisting cases where we were using the exception object outside of the TRY_CATCH block, and cases where we were using "else" after a 'if (ex.reason) < 0)' [a CATCH after this patch]. The result was folded into this patch so that GDB still builds at each incremental step. END_CATCH is necessary for two reasons: First, because we name the exception object in the CATCH block, which requires creating a scope, which in turn must be closed somewhere. Declaring the exception variable in the initializer field of a for block, like: #define CATCH(EXCEPTION, mask) \ for (struct gdb_exception EXCEPTION; \ exceptions_state_mc_catch (&EXCEPTION, MASK); \ EXCEPTION = exception_none) would avoid needing END_CATCH, but alas, in C mode, we build with C90, which doesn't allow mixed declarations and code. Second, because when TRY/CATCH are wired to real C++ try/catch, as long as we need to handle cleanup chains, even if there's no CATCH block that wants to catch the exception, we need for stop at every frame in the unwind chain and run cleanups, then rethrow. That will be done in END_CATCH. After we require C++, we'll still need TRY/CATCH/END_CATCH until cleanups are completely phased out -- TRY/CATCH in C++ mode will save/restore the current cleanup chain, like in C mode, and END_CATCH catches otherwise uncaugh exceptions, runs cleanups and rethrows, so that C++ cleanups and exceptions can coexist. IMO, this still makes the TRY/CATCH code look a bit more like a newcomer would expect, so IMO worth it even if we weren't considering C++. gdb/ChangeLog. 2015-03-07 Pedro Alves <palves@redhat.com> * common/common-exceptions.c (struct catcher) <exception>: No longer a pointer to volatile exception. Now an exception value. <mask>: Delete field. (exceptions_state_mc_init): Remove all parameters. Adjust. (exceptions_state_mc): No longer pop the catcher here. (exceptions_state_mc_catch): New function. (throw_exception): Adjust. * common/common-exceptions.h (exceptions_state_mc_init): Remove all parameters. (exceptions_state_mc_catch): Declare. (TRY_CATCH): Rename to ... (TRY): ... this. Remove EXCEPTION and MASK parameters. (CATCH, END_CATCH): New. All callers adjusted. gdb/gdbserver/ChangeLog: 2015-03-07 Pedro Alves <palves@redhat.com> Adjust all callers of TRY_CATCH to use TRY/CATCH/END_CATCH instead.
2015-03-07 16:14:14 +01:00
CATCH (exception, RETURN_MASK_ALL)
{
fputs_filtered ("Could not kill the program being debugged",
gdb_stderr);
exception_print (gdb_stderr, exception);
}
Split TRY_CATCH into TRY + CATCH This patch splits the TRY_CATCH macro into three, so that we go from this: ~~~ volatile gdb_exception ex; TRY_CATCH (ex, RETURN_MASK_ERROR) { } if (ex.reason < 0) { } ~~~ to this: ~~~ TRY { } CATCH (ex, RETURN_MASK_ERROR) { } END_CATCH ~~~ Thus, we'll be getting rid of the local volatile exception object, and declaring the caught exception in the catch block. This allows reimplementing TRY/CATCH in terms of C++ exceptions when building in C++ mode, while still allowing to build GDB in C mode (using setjmp/longjmp), as a transition step. TBC, after this patch, is it _not_ valid to have code between the TRY and the CATCH blocks, like: TRY { } // some code here. CATCH (ex, RETURN_MASK_ERROR) { } END_CATCH Just like it isn't valid to do that with C++'s native try/catch. By switching to creating the exception object inside the CATCH block scope, we can get rid of all the explicitly allocated volatile exception objects all over the tree, and map the CATCH block more directly to C++'s catch blocks. The majority of the TRY_CATCH -> TRY+CATCH+END_CATCH conversion was done with a script, rerun from scratch at every rebase, no manual editing involved. After the mechanical conversion, a few places needed manual intervention, to fix preexisting cases where we were using the exception object outside of the TRY_CATCH block, and cases where we were using "else" after a 'if (ex.reason) < 0)' [a CATCH after this patch]. The result was folded into this patch so that GDB still builds at each incremental step. END_CATCH is necessary for two reasons: First, because we name the exception object in the CATCH block, which requires creating a scope, which in turn must be closed somewhere. Declaring the exception variable in the initializer field of a for block, like: #define CATCH(EXCEPTION, mask) \ for (struct gdb_exception EXCEPTION; \ exceptions_state_mc_catch (&EXCEPTION, MASK); \ EXCEPTION = exception_none) would avoid needing END_CATCH, but alas, in C mode, we build with C90, which doesn't allow mixed declarations and code. Second, because when TRY/CATCH are wired to real C++ try/catch, as long as we need to handle cleanup chains, even if there's no CATCH block that wants to catch the exception, we need for stop at every frame in the unwind chain and run cleanups, then rethrow. That will be done in END_CATCH. After we require C++, we'll still need TRY/CATCH/END_CATCH until cleanups are completely phased out -- TRY/CATCH in C++ mode will save/restore the current cleanup chain, like in C mode, and END_CATCH catches otherwise uncaugh exceptions, runs cleanups and rethrows, so that C++ cleanups and exceptions can coexist. IMO, this still makes the TRY/CATCH code look a bit more like a newcomer would expect, so IMO worth it even if we weren't considering C++. gdb/ChangeLog. 2015-03-07 Pedro Alves <palves@redhat.com> * common/common-exceptions.c (struct catcher) <exception>: No longer a pointer to volatile exception. Now an exception value. <mask>: Delete field. (exceptions_state_mc_init): Remove all parameters. Adjust. (exceptions_state_mc): No longer pop the catcher here. (exceptions_state_mc_catch): New function. (throw_exception): Adjust. * common/common-exceptions.h (exceptions_state_mc_init): Remove all parameters. (exceptions_state_mc_catch): Declare. (TRY_CATCH): Rename to ... (TRY): ... this. Remove EXCEPTION and MASK parameters. (CATCH, END_CATCH): New. All callers adjusted. gdb/gdbserver/ChangeLog: 2015-03-07 Pedro Alves <palves@redhat.com> Adjust all callers of TRY_CATCH to use TRY/CATCH/END_CATCH instead.
2015-03-07 16:14:14 +01:00
END_CATCH
Split TRY_CATCH into TRY + CATCH This patch splits the TRY_CATCH macro into three, so that we go from this: ~~~ volatile gdb_exception ex; TRY_CATCH (ex, RETURN_MASK_ERROR) { } if (ex.reason < 0) { } ~~~ to this: ~~~ TRY { } CATCH (ex, RETURN_MASK_ERROR) { } END_CATCH ~~~ Thus, we'll be getting rid of the local volatile exception object, and declaring the caught exception in the catch block. This allows reimplementing TRY/CATCH in terms of C++ exceptions when building in C++ mode, while still allowing to build GDB in C mode (using setjmp/longjmp), as a transition step. TBC, after this patch, is it _not_ valid to have code between the TRY and the CATCH blocks, like: TRY { } // some code here. CATCH (ex, RETURN_MASK_ERROR) { } END_CATCH Just like it isn't valid to do that with C++'s native try/catch. By switching to creating the exception object inside the CATCH block scope, we can get rid of all the explicitly allocated volatile exception objects all over the tree, and map the CATCH block more directly to C++'s catch blocks. The majority of the TRY_CATCH -> TRY+CATCH+END_CATCH conversion was done with a script, rerun from scratch at every rebase, no manual editing involved. After the mechanical conversion, a few places needed manual intervention, to fix preexisting cases where we were using the exception object outside of the TRY_CATCH block, and cases where we were using "else" after a 'if (ex.reason) < 0)' [a CATCH after this patch]. The result was folded into this patch so that GDB still builds at each incremental step. END_CATCH is necessary for two reasons: First, because we name the exception object in the CATCH block, which requires creating a scope, which in turn must be closed somewhere. Declaring the exception variable in the initializer field of a for block, like: #define CATCH(EXCEPTION, mask) \ for (struct gdb_exception EXCEPTION; \ exceptions_state_mc_catch (&EXCEPTION, MASK); \ EXCEPTION = exception_none) would avoid needing END_CATCH, but alas, in C mode, we build with C90, which doesn't allow mixed declarations and code. Second, because when TRY/CATCH are wired to real C++ try/catch, as long as we need to handle cleanup chains, even if there's no CATCH block that wants to catch the exception, we need for stop at every frame in the unwind chain and run cleanups, then rethrow. That will be done in END_CATCH. After we require C++, we'll still need TRY/CATCH/END_CATCH until cleanups are completely phased out -- TRY/CATCH in C++ mode will save/restore the current cleanup chain, like in C mode, and END_CATCH catches otherwise uncaugh exceptions, runs cleanups and rethrows, so that C++ cleanups and exceptions can coexist. IMO, this still makes the TRY/CATCH code look a bit more like a newcomer would expect, so IMO worth it even if we weren't considering C++. gdb/ChangeLog. 2015-03-07 Pedro Alves <palves@redhat.com> * common/common-exceptions.c (struct catcher) <exception>: No longer a pointer to volatile exception. Now an exception value. <mask>: Delete field. (exceptions_state_mc_init): Remove all parameters. Adjust. (exceptions_state_mc): No longer pop the catcher here. (exceptions_state_mc_catch): New function. (throw_exception): Adjust. * common/common-exceptions.h (exceptions_state_mc_init): Remove all parameters. (exceptions_state_mc_catch): Declare. (TRY_CATCH): Rename to ... (TRY): ... this. Remove EXCEPTION and MASK parameters. (CATCH, END_CATCH): New. All callers adjusted. gdb/gdbserver/ChangeLog: 2015-03-07 Pedro Alves <palves@redhat.com> Adjust all callers of TRY_CATCH to use TRY/CATCH/END_CATCH instead.
2015-03-07 16:14:14 +01:00
TRY
{
gdb/ Code cleanup. * bfd-target.c (target_bfd_xclose): Remove parameter quitting. * bsd-kvm.c (bsd_kvm_close): Likewise. * bsd-uthread.c (bsd_uthread_close): Likewise. * corelow.c (core_close): Likewise. (core_close_cleanup): Remove parameter quitting from a caller. * event-top.c (async_disconnect): Likewise. * exec.c (exec_close_1): Remove parameter quitting. * go32-nat.c (go32_close): Likewise. * linux-nat.c (linux_nat_close): Remove parameter quitting. Remove parameter quitting from a caller. * mips-linux-nat.c (super_close): Remove parameter quitting from the variable. (mips_linux_close): Remove parameter quitting. Remove parameter quitting from a caller. * monitor.c (monitor_close): Remove parameter quitting. * monitor.h (monitor_close): Likewise. * record-btrace.c (record_btrace_close): Likewise. * record-full.c (record_full_close): Likewise. * remote-m32r-sdi.c (m32r_close): Remove parameter quitting and remove it also from fprintf_unfiltered. * remote-mips.c (mips_close): Remove parameter quitting. (mips_detach): Remove parameter quitting from a caller. * remote-sim.c (gdbsim_close): Remove parameter quitting. (gdbsim_close): Remove duplicate function comment. Remove parameter quitting and remove it also from printf_filtered. * remote.c (remote_close): Remove parameter quitting. * solib-svr4.c (enable_break): Remove parameter quitting from a caller. * target.c (update_current_target): Remove parameter int from to_close de_fault. (push_target, unpush_target, pop_target): Remove parameter quitting from a caller. (pop_all_targets_above, pop_all_targets): Remove parameter quitting. Remove parameter quitting from a caller. (target_preopen): Remove parameter quitting from a caller. (target_close): Remove parameter quitting. Remove parameter quitting from a caller two times. Remove parameter quitting also from fprintf_unfiltered. * target.h (struct target_ops): Remove parameter quitting and as int from fields to_xclose and to_close. (extern struct target_ops current_target): (target_close, pop_all_targets): Remove parameter quitting. Update the comment. (pop_all_targets_above): Remove parameter quitting. * top.c (quit_target): Remove parameter quitting from a caller. * tracepoint.c (tfile_close): Remove parameter quitting. * windows-nat.c (windows_close): Remove parameter quitting.
2013-03-20 16:46:24 +01:00
pop_all_targets ();
}
Split TRY_CATCH into TRY + CATCH This patch splits the TRY_CATCH macro into three, so that we go from this: ~~~ volatile gdb_exception ex; TRY_CATCH (ex, RETURN_MASK_ERROR) { } if (ex.reason < 0) { } ~~~ to this: ~~~ TRY { } CATCH (ex, RETURN_MASK_ERROR) { } END_CATCH ~~~ Thus, we'll be getting rid of the local volatile exception object, and declaring the caught exception in the catch block. This allows reimplementing TRY/CATCH in terms of C++ exceptions when building in C++ mode, while still allowing to build GDB in C mode (using setjmp/longjmp), as a transition step. TBC, after this patch, is it _not_ valid to have code between the TRY and the CATCH blocks, like: TRY { } // some code here. CATCH (ex, RETURN_MASK_ERROR) { } END_CATCH Just like it isn't valid to do that with C++'s native try/catch. By switching to creating the exception object inside the CATCH block scope, we can get rid of all the explicitly allocated volatile exception objects all over the tree, and map the CATCH block more directly to C++'s catch blocks. The majority of the TRY_CATCH -> TRY+CATCH+END_CATCH conversion was done with a script, rerun from scratch at every rebase, no manual editing involved. After the mechanical conversion, a few places needed manual intervention, to fix preexisting cases where we were using the exception object outside of the TRY_CATCH block, and cases where we were using "else" after a 'if (ex.reason) < 0)' [a CATCH after this patch]. The result was folded into this patch so that GDB still builds at each incremental step. END_CATCH is necessary for two reasons: First, because we name the exception object in the CATCH block, which requires creating a scope, which in turn must be closed somewhere. Declaring the exception variable in the initializer field of a for block, like: #define CATCH(EXCEPTION, mask) \ for (struct gdb_exception EXCEPTION; \ exceptions_state_mc_catch (&EXCEPTION, MASK); \ EXCEPTION = exception_none) would avoid needing END_CATCH, but alas, in C mode, we build with C90, which doesn't allow mixed declarations and code. Second, because when TRY/CATCH are wired to real C++ try/catch, as long as we need to handle cleanup chains, even if there's no CATCH block that wants to catch the exception, we need for stop at every frame in the unwind chain and run cleanups, then rethrow. That will be done in END_CATCH. After we require C++, we'll still need TRY/CATCH/END_CATCH until cleanups are completely phased out -- TRY/CATCH in C++ mode will save/restore the current cleanup chain, like in C mode, and END_CATCH catches otherwise uncaugh exceptions, runs cleanups and rethrows, so that C++ cleanups and exceptions can coexist. IMO, this still makes the TRY/CATCH code look a bit more like a newcomer would expect, so IMO worth it even if we weren't considering C++. gdb/ChangeLog. 2015-03-07 Pedro Alves <palves@redhat.com> * common/common-exceptions.c (struct catcher) <exception>: No longer a pointer to volatile exception. Now an exception value. <mask>: Delete field. (exceptions_state_mc_init): Remove all parameters. Adjust. (exceptions_state_mc): No longer pop the catcher here. (exceptions_state_mc_catch): New function. (throw_exception): Adjust. * common/common-exceptions.h (exceptions_state_mc_init): Remove all parameters. (exceptions_state_mc_catch): Declare. (TRY_CATCH): Rename to ... (TRY): ... this. Remove EXCEPTION and MASK parameters. (CATCH, END_CATCH): New. All callers adjusted. gdb/gdbserver/ChangeLog: 2015-03-07 Pedro Alves <palves@redhat.com> Adjust all callers of TRY_CATCH to use TRY/CATCH/END_CATCH instead.
2015-03-07 16:14:14 +01:00
CATCH (exception, RETURN_MASK_ALL)
{
}
END_CATCH
signal (SIGHUP, SIG_DFL); /*FIXME: ??????????? */
raise (SIGHUP);
1999-05-11 22:29:07 +02:00
}
#endif
1999-06-14 20:08:47 +02:00
#ifdef STOP_SIGNAL
1999-07-07 22:19:36 +02:00
void
1999-09-22 05:28:34 +02:00
handle_stop_sig (int sig)
1999-06-14 20:08:47 +02:00
{
mark_async_signal_handler (sigtstp_token);
1999-07-07 22:19:36 +02:00
signal (sig, handle_stop_sig);
1999-06-14 20:08:47 +02:00
}
static void
1999-09-22 05:28:34 +02:00
async_stop_sig (gdb_client_data arg)
1999-06-14 20:08:47 +02:00
{
char *prompt = get_prompt ();
1999-06-14 20:08:47 +02:00
#if STOP_SIGNAL == SIGTSTP
signal (SIGTSTP, SIG_DFL);
1999-10-06 01:13:56 +02:00
#if HAVE_SIGPROCMASK
{
sigset_t zero;
1999-10-06 01:13:56 +02:00
sigemptyset (&zero);
sigprocmask (SIG_SETMASK, &zero, 0);
}
#elif HAVE_SIGSETMASK
1999-06-14 20:08:47 +02:00
sigsetmask (0);
1999-10-06 01:13:56 +02:00
#endif
raise (SIGTSTP);
1999-06-14 20:08:47 +02:00
signal (SIGTSTP, handle_stop_sig);
#else
signal (STOP_SIGNAL, handle_stop_sig);
#endif
printf_unfiltered ("%s", prompt);
gdb_flush (gdb_stdout);
/* Forget about any previous command -- null line now will do
nothing. */
1999-06-14 20:08:47 +02:00
dont_repeat ();
}
#endif /* STOP_SIGNAL */
/* Tell the event loop what to do if SIGFPE is received.
See event-signal.c. */
1999-07-07 22:19:36 +02:00
static void
1999-09-22 05:28:34 +02:00
handle_sigfpe (int sig)
1999-05-11 22:29:07 +02:00
{
mark_async_signal_handler (sigfpe_token);
1999-05-11 22:29:07 +02:00
signal (sig, handle_sigfpe);
}
/* Event loop will call this functin to process a SIGFPE. */
1999-07-07 22:19:36 +02:00
static void
1999-09-22 05:28:34 +02:00
async_float_handler (gdb_client_data arg)
1999-05-11 22:29:07 +02:00
{
/* This message is based on ANSI C, section 4.7. Note that integer
divide by zero causes this, so "float" is a misnomer. */
2005-02-10 Andrew Cagney <cagney@gnu.org> Mark up all error and warning messages. * ada-lang.c, amd64-tdep.c, arch-utils.c, breakpoint.c: Update. * bsd-kvm.c, bsd-uthread.c, coff-solib.h, coffread.c: Update. * core-aout.c, core-regset.c, corefile.c, corelow.c: Update. * cp-abi.c, cp-support.c, cp-valprint.c, cris-tdep.c: Update. * dbxread.c, demangle.c, doublest.c, dsrec.c: Update. * dve3900-rom.c, dwarf2expr.c, dwarf2loc.c: Update. * dwarf2read.c, dwarfread.c, elfread.c, eval.c: Update. * event-top.c, exec.c, expprint.c, f-lang.c: Update. * f-typeprint.c, f-valprint.c, fbsd-nat.c, findvar.c: Update. * frame.c, frv-linux-tdep.c, gcore.c, gdbtypes.c: Update. * gnu-nat.c, gnu-v2-abi.c, gnu-v3-abi.c, go32-nat.c: Update. * hpacc-abi.c, hppa-hpux-nat.c, hppa-hpux-tdep.c: Update. * hppa-linux-nat.c, hppa-linux-tdep.c, hppa-tdep.c: Update. * hpread.c, hpux-thread.c, i386-linux-nat.c: Update. * i386-linux-tdep.c, i386-tdep.c, i386bsd-nat.c: Update. * i386gnu-nat.c, i387-tdep.c, ia64-linux-nat.c: Update. * ia64-tdep.c, inf-child.c, inf-ptrace.c, inf-ttrace.c: Update. * infcall.c, infcmd.c, inflow.c, infptrace.c, infrun.c: Update. * inftarg.c, interps.c, irix5-nat.c, jv-lang.c: Update. * kod-cisco.c, kod.c, language.c, libunwind-frame.c: Update. * linespec.c, linux-nat.c, linux-thread-db.c, m2-lang.c: Update. * m32r-rom.c, m68hc11-tdep.c, m68k-tdep.c: Update. * m68klinux-nat.c, macrocmd.c, macroexp.c, main.c: Update. * maint.c, mdebugread.c, mem-break.c, memattr.c: Update. * mips-linux-tdep.c, mips-tdep.c, mipsread.c, monitor.c: Update. * nlmread.c, nto-procfs.c, objc-lang.c, objfiles.c: Update. * observer.c, ocd.c, p-lang.c, p-typeprint.c: Update. * p-valprint.c, pa64solib.c, parse.c, ppc-linux-tdep.c: Update. * ppcnbsd-tdep.c, printcmd.c, procfs.c, remote-e7000.c: Update. * remote-fileio.c, remote-m32r-sdi.c, remote-rdi.c: Update. * remote-rdp.c, remote-sim.c, remote-st.c: Update. * remote-utils.c, remote-utils.h, remote.c: Update. * rom68k-rom.c, rs6000-nat.c, s390-tdep.c, scm-lang.c: Update. * ser-e7kpc.c, ser-tcp.c, ser-unix.c, sh-tdep.c: Update. * sh3-rom.c, shnbsd-tdep.c, sol-thread.c, solib-aix5.c: Update. * solib-frv.c, solib-irix.c, solib-osf.c, solib-pa64.c: Update. * solib-som.c, solib-sunos.c, solib-svr4.c, solib.c: Update. * somread.c, somsolib.c, source.c, stabsread.c: Update. * stack.c, std-regs.c, symfile-mem.c, symfile.c: Update. * symmisc.c, symtab.c, target.c, thread.c, top.c: Update. * tracepoint.c, trad-frame.c, typeprint.c, utils.c: Update. * uw-thread.c, valarith.c, valops.c, valprint.c: Update. * value.c, varobj.c, version.in, win32-nat.c, wince.c: Update. * xcoffread.c, xcoffsolib.c, cli/cli-cmds.c: Update. * cli/cli-decode.c, cli/cli-dump.c, cli/cli-logging.c: Update. * cli/cli-script.c, cli/cli-setshow.c, mi/mi-cmd-break.c: Update. * mi/mi-cmd-disas.c, mi/mi-cmd-env.c, mi/mi-cmd-file.c: Update. * mi/mi-cmd-stack.c, mi/mi-cmd-var.c, mi/mi-getopt.c: Update. * mi/mi-symbol-cmds.c, tui/tui-layout.c, tui/tui-stack.c: Update. * tui/tui-win.c: Update.
2005-02-11 05:06:14 +01:00
error (_("Erroneous arithmetic operation."));
1999-05-11 22:29:07 +02:00
}
/* Called by do_setshow_command. */
void
set_async_editing_command (char *args, int from_tty,
struct cmd_list_element *c)
1999-05-11 22:29:07 +02:00
{
change_line_handler ();
}
1999-06-14 20:08:47 +02:00
/* Set things up for readline to be invoked via the alternate
interface, i.e. via a callback function (rl_callback_read_char),
and hook up instream to the event loop. */
1999-06-14 20:08:47 +02:00
void
gdb_setup_readline (void)
1999-06-14 20:08:47 +02:00
{
/* This function is a noop for the sync case. The assumption is
that the sync setup is ALL done in gdb_init, and we would only
mess it up here. The sync stuff should really go away over
time. */
if (!batch_silent)
gdb_stdout = stdio_fileopen (stdout);
gdb_stderr = stderr_fileopen ();
gdb_stdlog = gdb_stderr; /* for moment */
gdb_stdtarg = gdb_stderr; /* for moment */
gdb_stdtargerr = gdb_stderr; /* for moment */
/* If the input stream is connected to a terminal, turn on
editing. */
if (ISATTY (instream))
1999-06-21 15:27:42 +02:00
{
/* Tell gdb that we will be using the readline library. This
could be overwritten by a command in .gdbinit like 'set
editing on' or 'off'. */
async_command_editing_p = 1;
/* When a character is detected on instream by select or poll,
readline will be invoked via this callback function. */
call_readline = rl_callback_read_char_wrapper;
1999-06-21 15:27:42 +02:00
}
else
{
async_command_editing_p = 0;
call_readline = gdb_readline_no_editing_callback;
}
/* When readline has read an end-of-line character, it passes the
complete line to gdb for processing; command_line_handler is the
function that does this. */
input_handler = command_line_handler;
/* Tell readline to use the same input stream that gdb uses. */
rl_instream = instream;
/* Get a file descriptor for the input stream, so that we can
register it with the event loop. */
input_fd = fileno (instream);
/* Now we need to create the event sources for the input file
descriptor. */
/* At this point in time, this is the only event source that we
register with the even loop. Another source is going to be the
target program (inferior), but that must be registered only when
it actually exists (I.e. after we say 'run' or after we connect
to a remote target. */
add_file_handler (input_fd, stdin_event_handler, 0);
1999-06-14 20:08:47 +02:00
}
/* Disable command input through the standard CLI channels. Used in
the suspend proc for interpreters that use the standard gdb readline
interface, like the cli & the mi. */
void
gdb_disable_readline (void)
{
/* FIXME - It is too heavyweight to delete and remake these every
time you run an interpreter that needs readline. It is probably
better to have the interpreters cache these, which in turn means
that this needs to be moved into interpreter specific code. */
#if 0
ui_file_delete (gdb_stdout);
ui_file_delete (gdb_stderr);
gdb_stdlog = NULL;
gdb_stdtarg = NULL;
gdb_stdtargerr = NULL;
#endif
PR python/17372 - Python hangs when displaying help() This is more of a readline/terminal issue than a Python one. PR17372 is a regression in 7.8 caused by the fix for PR17072: commit 0017922d0292d8c374584f6100874580659c9973 Author: Pedro Alves <palves@redhat.com> Date: Mon Jul 14 19:55:32 2014 +0100 Background execution + pagination aborts readline/gdb gdb_readline_wrapper_line removes the handler after a line is processed. Usually, we'll end up re-displaying the prompt, and that reinstalls the handler. But if the output is coming out of handling a stop event, we don't re-display the prompt, and nothing restores the handler. So the next input wakes up the event loop and calls into readline, which aborts. ... gdb/ 2014-07-14 Pedro Alves <palves@redhat.com> PR gdb/17072 * top.c (gdb_readline_wrapper_line): Tweak comment. (gdb_readline_wrapper_cleanup): If readline is enabled, reinstall the input handler callback. The problem is that installing the input handler callback also preps the terminal, putting it in raw mode and with echo disabled, which is bad if we're going to call a command that assumes cooked/canonical mode, and echo enabled, like in the case of the PR, Python's interactive shell. Another example I came up with that doesn't depend on Python is starting a subshell with "(gdb) shell /bin/sh" from a multi-line command. Tests covering both these examples are added. The fix is to revert the original fix for PR gdb/17072, and instead restore the callback handler after processing an asynchronous target event. Furthermore, calling rl_callback_handler_install when we already have some input in readline's line buffer discards that input, which is obviously a bad thing to do while the user is typing. No specific test is added for that, because I first tried calling it even if the callback handler was still installed and that resulted in hundreds of failures in the testsuite. gdb/ 2014-10-29 Pedro Alves <palves@redhat.com> PR python/17372 * event-top.c (change_line_handler): Call gdb_rl_callback_handler_remove instead of rl_callback_handler_remove. (callback_handler_installed): New global. (gdb_rl_callback_handler_remove, gdb_rl_callback_handler_install) (gdb_rl_callback_handler_reinstall): New functions. (display_gdb_prompt): Call gdb_rl_callback_handler_remove and gdb_rl_callback_handler_install instead of rl_callback_handler_remove and rl_callback_handler_install. (gdb_disable_readline): Call gdb_rl_callback_handler_remove instead of rl_callback_handler_remove. * event-top.h (gdb_rl_callback_handler_remove) (gdb_rl_callback_handler_install) (gdb_rl_callback_handler_reinstall): New declarations. * infrun.c (reinstall_readline_callback_handler_cleanup): New cleanup function. (fetch_inferior_event): Install it. * top.c (gdb_readline_wrapper_line) Call gdb_rl_callback_handler_remove instead of rl_callback_handler_remove. (gdb_readline_wrapper_cleanup): Don't call rl_callback_handler_install. gdb/testsuite/ 2014-10-29 Pedro Alves <palves@redhat.com> PR python/17372 * gdb.python/python.exp: Test a multi-line command that spawns interactive Python. * gdb.base/multi-line-starts-subshell.exp: New file.
2014-10-23 18:13:35 +02:00
gdb_rl_callback_handler_remove ();
delete_file_handler (input_fd);
}