Re: ulimit and ssh?
peter360 a écrit : > Thanks for the explanation. So my understanding of the way ssh works is > still incorrect. I am confused about at which point the two parameters, > "-c" and "ulimit -a" were converted into three, "-c", "ulimit", and "-a". I > guess I need to read the source of ssh and bash to really understand this... Enabling tracing options usually helps a lot. Instead of this: sshlocalhost bash-c 'ulimit -a' Try this: ssh -v localhost bash -x -c 'ulimit -a' (And as usual, check out the "Guide to unix shell quoting")
Bash-4.0 Official Patch 29
BASH PATCH REPORT
=
Bash-Release: 4.0
Patch-ID: bash40-029
Bug-Reported-by:Christian Krause
Bug-Reference-ID: Thu, 25 Jun 2009 21:47:59 +0200
Bug-Reference-URL:
http://lists.gnu.org/archive/html/bug-bash/2009-06/msg00078.html
Bug-Description:
Previous versions of bash accepted strings in the initial environment
that were not valid shell variable assignments, usually because the
names were invalid, but still created shell variables from them and
passed them to child processes in the environment.
Bash-4.0 ignores those names and does not pass them to child processes.
Some users and automated processes depend on invalid variables being
ignored and passed to child processes.
This patch makes bash continue to ignore the invalid names, but pass
them to child processes in the export environment.
Patch:
*** ../bash-4.0-patched/variables.c 2009-01-04 14:32:46.0 -0500
--- variables.c 2009-06-29 09:17:20.0 -0400
***
*** 253,256
--- 255,259
static int visible_var __P((SHELL_VAR *));
static int visible_and_exported __P((SHELL_VAR *));
+ static int export_environment_candidate __P((SHELL_VAR *));
static int local_and_exported __P((SHELL_VAR *));
static int variable_in_context __P((SHELL_VAR *));
***
*** 376,383
# endif
#endif
else if (legal_identifier (name))
{
temp_var = bind_variable (name, string, 0);
! VSETATTR (temp_var, (att_exported | att_imported));
array_needs_making = 1;
}
--- 379,393
# endif
#endif
+ #if 0
else if (legal_identifier (name))
+ #else
+ else
+ #endif
{
temp_var = bind_variable (name, string, 0);
! if (legal_identifier (name))
! VSETATTR (temp_var, (att_exported | att_imported));
! else
! VSETATTR (temp_var, (att_exported | att_imported | att_invisible));
array_needs_making = 1;
}
***
*** 3083,3086
--- 3098,3111
}
+ /* Candidate variables for the export environment are either valid variables
+with the export attribute or invalid variables inherited from the initial
+environment and simply passed through. */
+ static int
+ export_environment_candidate (var)
+ SHELL_VAR *var;
+ {
+ return (exported_p (var) && (invisible_p (var) == 0 || imported_p (var)));
+ }
+
/* Return non-zero if VAR is a local variable in the current context and
is exported. */
***
*** 3439,3443
--- 3464,3472
SHELL_VAR **vars;
+ #if 0
vars = map_over (visible_and_exported, vcxt);
+ #else
+ vars = map_over (export_environment_candidate, vcxt);
+ #endif
if (vars == 0)
*** ../bash-4.0/patchlevel.h2009-01-04 14:32:40.0 -0500
--- patchlevel.h2009-02-22 16:11:31.0 -0500
***
*** 26,30
looks for to find the patch level (for the sccs version string). */
! #define PATCHLEVEL 28
#endif /* _PATCHLEVEL_H_ */
--- 26,30
looks for to find the patch level (for the sccs version string). */
! #define PATCHLEVEL 29
#endif /* _PATCHLEVEL_H_ */
--
``The lyf so short, the craft so long to lerne.'' - Chaucer
``Ars longa, vita brevis'' - Hippocrates
Chet Ramey, ITS, CWRUc...@case.eduhttp://cnswww.cns.cwru.edu/~chet/
Bash-4.0 Official Patch 30
BASH PATCH REPORT
=
Bash-Release: 4.0
Patch-ID: bash40-030
Bug-Reported-by:Henning Bekel
Bug-Reference-ID: <7c6eacf262ct...@mid.individual.net>
Bug-Reference-URL:
http://lists.gnu.org/archive/html/bug-bash/2009-07/msg00054.html
Bug-Description:
A shell function invoked with `bind -x' is supposed to be able to move the
cursor by setting READLINE_POINT. The effects of this assignment were
sometimes ignored.
Patch:
*** ../bash-4.0-patched/bashline.c 2009-01-08 09:29:24.0 -0500
--- bashline.c 2009-07-16 14:13:41.0 -0400
***
*** 3389,3393
register int i;
intmax_t mi;
- int save_point;
sh_parser_state_t ps;
char *cmd, *value, *l;
--- 3389,3392
***
*** 3433,3437
VSETATTR (v, att_exported);
l = value_cell (v);
- save_point = rl_point;
value = inttostr (rl_point, ibuf, sizeof (ibuf));
v = bind_int_variable ("READLINE_POINT", value);
--- 3432,3435
***
*** 3451,3455
{
i = mi;
! if (i != save_point)
{
rl_point = i;
--- 3449,3453
{
i = mi;
! if (i != rl_point)
{
rl_point = i;
*** ../bash-4.0/patchlevel.h2009-01-04 14:32:40.0 -0500
--- patchlevel.h2009-02-22 16:11:31.0 -0500
***
*** 26,30
looks for to find the patch level (for the sccs version string). */
! #define PATCHLEVEL 29
#endif /* _PATCHLEVEL_H_ */
--- 26,30
looks for to find the patch level (for the sccs version string). */
! #define PATCHLEVEL 30
#endif /* _PATCHLEVEL_H_ */
--
``The lyf so short, the craft so long to lerne.'' - Chaucer
``Ars longa, vita brevis'' - Hippocrates
Chet Ramey, ITS, CWRUc...@case.eduhttp://cnswww.cns.cwru.edu/~chet/
Bash-4.0 Official Patch 31
BASH PATCH REPORT
=
Bash-Release: 4.0
Patch-ID: bash40-031
Bug-Reported-by:Roman Rakus
Bug-Reference-ID: <4a93f6e9.4050...@redhat.com>
Bug-Reference-URL:
Bug-Description:
An implicit assignment to index "0" of an existing array variable caused
the shell to crash when the variable was unset.
Patch:
*** ../bash-4.0-patched/arrayfunc.c 2009-03-08 21:24:39.0 -0400
--- arrayfunc.c 2009-08-24 09:29:43.0 -0400
***
*** 99,103
hash = assoc_create (0);
if (oldval)
! assoc_insert (hash, "0", oldval);
FREE (value_cell (var));
--- 99,103
hash = assoc_create (0);
if (oldval)
! assoc_insert (hash, savestring ("0"), oldval);
FREE (value_cell (var));
*** ../bash-4.0-patched/variables.c 2009-01-04 14:32:46.0 -0500
--- variables.c 2009-08-24 09:29:58.0 -0400
***
*** 2218,
else if (assoc_p (entry))
{
! assoc_insert (assoc_cell (entry), "0", newval);
free (newval);
}
--- 2218,
else if (assoc_p (entry))
{
! assoc_insert (assoc_cell (entry), savestring ("0"), newval);
free (newval);
}
*** ../bash-4.0/patchlevel.h2009-01-04 14:32:40.0 -0500
--- patchlevel.h2009-02-22 16:11:31.0 -0500
***
*** 26,30
looks for to find the patch level (for the sccs version string). */
! #define PATCHLEVEL 30
#endif /* _PATCHLEVEL_H_ */
--- 26,30
looks for to find the patch level (for the sccs version string). */
! #define PATCHLEVEL 31
#endif /* _PATCHLEVEL_H_ */
--
``The lyf so short, the craft so long to lerne.'' - Chaucer
``Ars longa, vita brevis'' - Hippocrates
Chet Ramey, ITS, CWRUc...@case.eduhttp://cnswww.cns.cwru.edu/~chet/
Bash-4.0 Official Patch 32
BASH PATCH REPORT
=
Bash-Release: 4.0
Patch-ID: bash40-032
Bug-Reported-by:mu...@muszi.kite.hu
Bug-Reference-ID: <20090826113159.18815.qm...@muszi.kite.hu>
Bug-Reference-URL:
http://lists.gnu.org/archive/html/bug-bash/2009-08/msg00090.html
Bug-Description:
Bash-4.0 has a memory leak when processing ${!pre...@}.
Patch:
*** ../bash-4.0-patched/subst.c 2009-07-22 23:18:55.0 -0400
--- subst.c 2009-08-26 23:08:51.0 -0400
***
*** 6607,6611
}
free (x);
! free (xlist);
free (temp1);
*indexp = sindex;
--- 6769,6773
}
free (x);
! dispose_words (xlist);
free (temp1);
*indexp = sindex;
*** ../bash-4.0/patchlevel.h2009-01-04 14:32:40.0 -0500
--- patchlevel.h2009-02-22 16:11:31.0 -0500
***
*** 26,30
looks for to find the patch level (for the sccs version string). */
! #define PATCHLEVEL 31
#endif /* _PATCHLEVEL_H_ */
--- 26,30
looks for to find the patch level (for the sccs version string). */
! #define PATCHLEVEL 32
#endif /* _PATCHLEVEL_H_ */
--
``The lyf so short, the craft so long to lerne.'' - Chaucer
``Ars longa, vita brevis'' - Hippocrates
Chet Ramey, ITS, CWRUc...@case.eduhttp://cnswww.cns.cwru.edu/~chet/
Bash-4.0 Official Patch 33
BASH PATCH REPORT
=
Bash-Release: 4.0
Patch-ID: bash40-033
Bug-Reported-by:Dr. Werner Fink
Bug-Reference-ID: <200907010951.n619p76i013...@boole.suse.de>
Bug-Reference-URL:
http://lists.gnu.org/archive/html/bug-bash/2009-07/msg0.html
Bug-Description:
Bash-4.0 has a memory leak in the `read' builtin when the number of fields
read is not the same as the number of variables passed as arguments.
Patch:
*** ../bash-4.0-patched/builtins/read.def 2009-03-08 21:24:45.0
-0400
--- builtins/read.def 2009-07-01 15:32:42.0 -0400
***
*** 764,768
tofree = input_string = t;
else
! input_string = strip_trailing_ifs_whitespace (t1, ifs_chars,
saw_escape);
}
#endif
--- 764,771
tofree = input_string = t;
else
! {
! input_string = strip_trailing_ifs_whitespace (t1, ifs_chars,
saw_escape);
! tofree = t;
! }
}
#endif
*** ../bash-4.0/patchlevel.h2009-01-04 14:32:40.0 -0500
--- patchlevel.h2009-02-22 16:11:31.0 -0500
***
*** 26,30
looks for to find the patch level (for the sccs version string). */
! #define PATCHLEVEL 32
#endif /* _PATCHLEVEL_H_ */
--- 26,30
looks for to find the patch level (for the sccs version string). */
! #define PATCHLEVEL 33
#endif /* _PATCHLEVEL_H_ */
--
``The lyf so short, the craft so long to lerne.'' - Chaucer
``Ars longa, vita brevis'' - Hippocrates
Chet Ramey, ITS, CWRUc...@case.eduhttp://cnswww.cns.cwru.edu/~chet/
Sourcing shopt setting removes 'history' from SHELLOPTS
Configuration Information [Automatically generated, do not change]: Machine: i486 OS: linux-gnu Compiler: gcc Compilation CFLAGS: -DPROGRAM='bash' -DCONF_HOSTTYPE='i486' -DCONF_OSTYPE='linux-gnu' -DCONF_MACHTYPE='i486-pc-linux-gnu' -DCONF_VENDOR='pc' -DLOCALEDIR='/usr/share/locale' -DPACKAGE='bash' -DSHELL -DHAVE_CONFIG_H -I. -I../bash -I../bash/include -I../bash/lib -g -O2 -Wall uname output: Linux myhost 2.6.26-2-686 #1 SMP Wed Aug 19 06:06:52 UTC 2009 i686 GNU/Linux Machine Type: i486-pc-linux-gnu Bash Version: 3.2 Patch Level: 39 Release Status: release Description: Sourcing or eval'ing code which touches a shopt setting, causes `history' to disappear from the SHELLOPTS environment variable. Repeat-By: $ # Make sure 'history' is on, and 'verbose' is off $ set -o history +o verbose $ echo $SHELLOPTS braceexpand:histexpand:history:interactive-comments:monitor:vi $ eval set +o verbose $ echo $SHELLOPTS # 'history' is now gone braceexpand:histexpand:interactive-comments:monitor:vi The same happens if you source a file (e.g. t.sh) with just one command (e.g. set -o nounset): $ set -o history $ echo $SHELLOPTS braceexpand:histexpand:history:interactive-comments:monitor:vi $ cat t.sh set -o nounset $ source t.sh $ echo $SHELLOPTS # 'history' is now gone braceexpand:histexpand:interactive-comments:monitor:nounset:vi It doesn't occur if the shopt setting is touched directly (without source or eval): $ # Make sure 'history' is on, and 'verbose' is off $ set -o history +o verbose $ echo $SHELLOPTS braceexpand:histexpand:history:interactive-comments:monitor:vi $ set +o verbose $ echo $SHELLOPTS # 'history' is still there braceexpand:histexpand:history:interactive-comments:monitor:vi Regards, Freddy Vulto http://fvue.nl
Re: Sourcing shopt setting removes 'history' from SHELLOPTS
> Machine Type: i486-pc-linux-gnu > > Bash Version: 3.2 > Patch Level: 39 > Release Status: release > > Description: > Sourcing or eval'ing code which touches a shopt setting, causes > `history' to disappear from the SHELLOPTS environment variable. Look at bash-3.2 patch 41. Chet -- ``The lyf so short, the craft so long to lerne.'' - Chaucer ``Ars longa, vita brevis'' - Hippocrates Chet Ramey, ITS, CWRUc...@case.eduhttp://cnswww.cns.cwru.edu/~chet/
