[RFC,02/17] Move gdb-specific code out of start_event_loop
Commit Message
This moves some gdb-specific code out of start_event_loop, into a new
function that must be supplied by the event-loop client.
gdb/ChangeLog
2019-02-24 Tom Tromey <tom@tromey.com>
* main.c (handle_event_loop_exception): New function.
* event-loop.c: Don't include event-top.h or observable.h.
(start_event_loop): Remove gdb-specific code. Call
handle_event_loop_exception.
* event-loop.h (handle_event_loop_exception): Declare.
---
gdb/ChangeLog | 8 ++++++++
gdb/event-loop.c | 26 ++------------------------
gdb/event-loop.h | 6 +++++-
gdb/main.c | 28 ++++++++++++++++++++++++++++
4 files changed, 43 insertions(+), 25 deletions(-)
Comments
On 2/24/19 4:51 PM, Tom Tromey wrote:
> This moves some gdb-specific code out of start_event_loop, into a new
> function that must be supplied by the event-loop client.
>
> gdb/ChangeLog
> 2019-02-24 Tom Tromey <tom@tromey.com>
>
> * main.c (handle_event_loop_exception): New function.
> * event-loop.c: Don't include event-top.h or observable.h.
> (start_event_loop): Remove gdb-specific code. Call
> handle_event_loop_exception.
> * event-loop.h (handle_event_loop_exception): Declare.
I think I'd rather move the whole start_event_loop to gdb.
IIUC, gdbserver version wants to let the exception propagate, so
it'd end up being a while (1) loop calling gdb_do_one_event.
I think the result will be clearer with fewer callsbacks and
no need for the shared ability to stop the event loop that is
only used by gdbserver.
Thanks,
Pedro Alves
@@ -19,7 +19,6 @@
#include "defs.h"
#include "event-loop.h"
-#include "event-top.h"
#include "ser-event.h"
#ifdef HAVE_POLL
@@ -33,7 +32,6 @@
#include <sys/types.h>
#include "common/gdb_sys_time.h"
#include "gdb_select.h"
-#include "observable.h"
#include "top.h"
/* Tell create_file_handler what events we are interested in.
@@ -355,7 +353,7 @@ gdb_do_one_event (void)
from the command loop. */
void
-start_event_loop (void)
+start_event_loop ()
{
/* Loop until there is nothing to do. This is the entry point to
the event loop engine. gdb_do_one_event will process one event
@@ -371,27 +369,7 @@ start_event_loop (void)
}
CATCH (ex, RETURN_MASK_ALL)
{
- exception_print (gdb_stderr, ex);
-
- /* If any exception escaped to here, we better enable
- stdin. Otherwise, any command that calls async_disable_stdin,
- and then throws, will leave stdin inoperable. */
- async_enable_stdin ();
- /* If we long-jumped out of do_one_event, we probably didn't
- get around to resetting the prompt, which leaves readline
- in a messed-up state. Reset it here. */
- current_ui->prompt_state = PROMPT_NEEDED;
- gdb::observers::command_error.notify ();
- /* This call looks bizarre, but it is required. If the user
- entered a command that caused an error,
- after_char_processing_hook won't be called from
- rl_callback_read_char_wrapper. Using a cleanup there
- won't work, since we want this function to be called
- after a new prompt is printed. */
- if (after_char_processing_hook)
- (*after_char_processing_hook) ();
- /* Maybe better to set a flag to be checked somewhere as to
- whether display the prompt or not. */
+ handle_event_loop_exception (ex);
}
END_CATCH
@@ -80,7 +80,7 @@ typedef void (timer_handler_func) (gdb_client_data);
/* Exported functions from event-loop.c */
-extern void start_event_loop (void);
+extern void start_event_loop ();
extern int gdb_do_one_event (void);
extern void delete_file_handler (int fd);
extern void add_file_handler (int fd, handler_func *proc,
@@ -131,4 +131,8 @@ extern void clear_async_event_handler (struct async_event_handler *handler);
extern void initialize_async_signal_handlers (void);
+/* Must be defined by client. */
+
+extern void handle_event_loop_exception (const gdb_exception &);
+
#endif /* EVENT_LOOP_H */
@@ -48,6 +48,7 @@
#include <vector>
#include "common/pathstuff.h"
#include "cli/cli-style.h"
+#include "observable.h"
/* The selected interpreter. This will be used as a set command
variable, so it should always be malloc'ed - since
@@ -314,6 +315,33 @@ setup_alternate_signal_stack (void)
here. */
static void captured_command_loop () __attribute__((noinline));
+/* Called to handle an exception when processing an event. */
+void
+handle_event_loop_exception (const gdb_exception &ex)
+{
+ exception_print (gdb_stderr, ex);
+
+ /* If any exception escaped to here, we better enable
+ stdin. Otherwise, any command that calls async_disable_stdin,
+ and then throws, will leave stdin inoperable. */
+ async_enable_stdin ();
+ /* If we long-jumped out of do_one_event, we probably didn't
+ get around to resetting the prompt, which leaves readline
+ in a messed-up state. Reset it here. */
+ current_ui->prompt_state = PROMPT_NEEDED;
+ gdb::observers::command_error.notify ();
+ /* This call looks bizarre, but it is required. If the user
+ entered a command that caused an error,
+ after_char_processing_hook won't be called from
+ rl_callback_read_char_wrapper. Using a cleanup there
+ won't work, since we want this function to be called
+ after a new prompt is printed. */
+ if (after_char_processing_hook)
+ (*after_char_processing_hook) ();
+ /* Maybe better to set a flag to be checked somewhere as to
+ whether display the prompt or not. */
+}
+
static void
captured_command_loop ()
{