From: Tristan Gingold <gingold@adacore.com>
To: "gdb-patches@sourceware.org ml" <gdb-patches@sourceware.org>
Cc: Pedro Alves <pedro@codesourcery.com>
Subject: [RFA] Preliminary work in fork_inferior
Date: Fri, 16 Sep 2011 13:20:00 -0000 [thread overview]
Message-ID: <8B49B563-BE0B-452C-8956-31E00A880D96@adacore.com> (raw)
In-Reply-To: <201109071409.06452.pedro@codesourcery.com>
Hi,
this patch is both a cleanup and a preliminary work for Lion.
It fixes a few weirdness:
- shell_command was always allocated even if not used
- argv was xmalloc'ed but never free
- gdb_flush/_exit sequence for exec failure was duplicated
The preliminary work consists in calling execvp wether or not a shell is executed.
No regressions on GNU/Linux i386
Ok for trunk ?
Tristan.
2011-09-07 Tristan Gingold <gingold@adacore.com>
* fork-child.c (fork_inferior): Update comment.
Use alloca instead of xmalloc for argv. Move
len and shell_command declarations in the block where they are used.
Only call execvp. Factorize some failure code.
diff --git a/gdb/fork-child.c b/gdb/fork-child.c
index bb173e7..e937aec 100644
--- a/gdb/fork-child.c
+++ b/gdb/fork-child.c
@@ -126,9 +126,7 @@ fork_inferior (char *exec_file_arg, char *allargs, char **env,
void (*pre_trace_fun) (void), char *shell_file_arg)
{
int pid;
- char *shell_command;
static char default_shell_file[] = SHELL_FILE;
- int len;
/* Set debug_fork then attach to the child while it sleeps, to debug. */
static int debug_fork = 0;
/* This is set to the result of setpgrp, which if vforked, will be visible
@@ -162,16 +160,6 @@ fork_inferior (char *exec_file_arg, char *allargs, char **env,
shell = 1;
}
- /* Multiplying the length of exec_file by 4 is to account for the
- fact that it may expand when quoted; it is a worst-case number
- based on every character being '. */
- len = 5 + 4 * strlen (exec_file) + 1 + strlen (allargs) + 1 + /*slop */ 12;
- if (exec_wrapper)
- len += strlen (exec_wrapper) + 1;
-
- shell_command = (char *) alloca (len);
- shell_command[0] = '\0';
-
if (!shell)
{
/* We're going to call execvp. Create argument vector.
@@ -180,18 +168,29 @@ fork_inferior (char *exec_file_arg, char *allargs, char **env,
argument. */
int argc = (strlen (allargs) + 1) / 2 + 2;
- argv = (char **) xmalloc (argc * sizeof (*argv));
+ argv = (char **) alloca (argc * sizeof (*argv));
argv[0] = exec_file;
breakup_args (allargs, &argv[1]);
}
else
{
/* We're going to call a shell. */
-
+ char *shell_command;
+ int len;
char *p;
int need_to_quote;
const int escape_bang = escape_bang_in_quoted_argument (shell_file);
+ /* Multiplying the length of exec_file by 4 is to account for the
+ fact that it may expand when quoted; it is a worst-case number
+ based on every character being '. */
+ len = 5 + 4 * strlen (exec_file) + 1 + strlen (allargs) + 1 + /*slop */ 12;
+ if (exec_wrapper)
+ len += strlen (exec_wrapper) + 1;
+
+ shell_command = (char *) alloca (len);
+ shell_command[0] = '\0';
+
strcat (shell_command, "exec ");
/* Add any exec wrapper. That may be a program name with arguments, so
@@ -257,6 +256,16 @@ fork_inferior (char *exec_file_arg, char *allargs, char **env,
strcat (shell_command, " ");
strcat (shell_command, allargs);
+
+ /* If we decided above to start up with a shell, we exec the
+ shell, "-c" says to interpret the next arg as a shell command
+ to execute, and this command is "exec <target-program>
+ <args>". */
+ argv = (char **) alloca (4 * sizeof (char *));
+ argv[0] = shell_file;
+ argv[1] = "-c";
+ argv[2] = shell_command;
+ argv[3] = (char *) 0;
}
/* On some systems an exec will fail if the executable is open. */
@@ -348,29 +357,18 @@ fork_inferior (char *exec_file_arg, char *allargs, char **env,
path to find $SHELL. Rich Pixley says so, and I agree. */
environ = env;
- /* If we decided above to start up with a shell, we exec the
- shell, "-c" says to interpret the next arg as a shell command
- to execute, and this command is "exec <target-program>
- <args>". */
+ execvp (argv[0], argv);
+
+ /* If we get here, it's an error. */
if (shell)
{
- execlp (shell_file, shell_file, "-c", shell_command, (char *) 0);
-
- /* If we get here, it's an error. */
fprintf_unfiltered (gdb_stderr, "Cannot exec %s: %s.\n", shell_file,
safe_strerror (errno));
- gdb_flush (gdb_stderr);
- _exit (0177);
}
else
{
- /* Otherwise, we directly exec the target program with
- execvp. */
int i;
- execvp (exec_file, argv);
-
- /* If we get here, it's an error. */
safe_strerror (errno);
fprintf_unfiltered (gdb_stderr, "Cannot exec %s ", exec_file);
@@ -383,9 +381,9 @@ fork_inferior (char *exec_file_arg, char *allargs, char **env,
i++;
}
fprintf_unfiltered (gdb_stderr, ".\n");
- gdb_flush (gdb_stderr);
- _exit (0177);
}
+ gdb_flush (gdb_stderr);
+ _exit (0177);
}
/* Restore our environment in case a vforked child clob'd it. */
next prev parent reply other threads:[~2011-09-16 12:29 UTC|newest]
Thread overview: 11+ messages / expand[flat|nested] mbox.gz Atom feed top
2011-09-07 13:09 [RFC] Support of Lion (darwin 11) Tristan Gingold
2011-09-07 13:14 ` Pedro Alves
2011-09-07 13:28 ` Pedro Alves
2011-09-07 13:47 ` Tristan Gingold
2011-09-16 13:20 ` Tristan Gingold [this message]
2011-09-16 14:38 ` [RFA] Preliminary work in fork_inferior Pedro Alves
2011-09-16 14:52 ` Tristan Gingold
2011-09-16 15:28 ` Pedro Alves
2011-09-16 18:25 ` Tristan Gingold
2011-09-16 14:47 ` Abhijit Halder
2011-09-16 14:55 ` Tristan Gingold
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=8B49B563-BE0B-452C-8956-31E00A880D96@adacore.com \
--to=gingold@adacore.com \
--cc=gdb-patches@sourceware.org \
--cc=pedro@codesourcery.com \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox