Strange behaviour from jobs -p in a subshell
Consider the following script. While the 3 sleeps are running, both jobs -p and $(jobs -p) will print 3 PIDs. Once the 3 children are finished, jobs -p will continue to print the 3 PIDs of the done Children, but $(jobs -p) will only print 1 PID. $(jobs -p) always seems to print at most 1 PID of a done child. #!/usr/bin/bash (sleep 2 ) & (sleep 2 ) & (sleep 2 ) & while /bin/true do echo A echo $(jobs -p) echo B jobs -p echo C sleep 1 done
Re: Strange behaviour from jobs -p in a subshell
On 11/13/18 4:28 AM, Christopher Jefferson wrote: > Consider the following script. While the 3 sleeps are running, both jobs > -p and $(jobs -p) will print 3 PIDs. Once the 3 children are finished, > jobs -p will continue to print the 3 PIDs of the done Children, but > $(jobs -p) will only print 1 PID. $(jobs -p) always seems to print at > most 1 PID of a done child. Since the $(jobs -p) is run in a subshell, its knowledge of its parent's jobs is transient. In this case, the subshell deletes knowledge of the jobs it inherits from its parent, but hangs onto the last asynchronous job in case the subshell references $!. Chet -- ``The lyf so short, the craft so long to lerne.'' - Chaucer ``Ars longa, vita brevis'' - Hippocrates Chet Ramey, UTech, CWRUc...@case.eduhttp://tiswww.cwru.edu/~chet/
Re: Strange behaviour from jobs -p in a subshell
On Tue, Nov 13, 2018 at 09:59:51AM -0500, Chet Ramey wrote: > On 11/13/18 4:28 AM, Christopher Jefferson wrote: > > Consider the following script. While the 3 sleeps are running, both jobs > > -p and $(jobs -p) will print 3 PIDs. Once the 3 children are finished, > > jobs -p will continue to print the 3 PIDs of the done Children, but > > $(jobs -p) will only print 1 PID. $(jobs -p) always seems to print at > > most 1 PID of a done child. > > Since the $(jobs -p) is run in a subshell, its knowledge of its parent's > jobs is transient. In this case, the subshell deletes knowledge of the > jobs it inherits from its parent, but hangs onto the last asynchronous job > in case the subshell references $!. > > Chet If the goal is to obtain the result of "jobs -p" and use it in a script, I would suggest redirecting the output of jobs -p to a temp file, then reading it. That skips the subshell.
built-in '[' and '/usr/bin/[' yield different results
Configuration Information [Automatically generated, do not change]: Machine: x86_64 OS: linux-gnu Compiler: gcc Compilation CFLAGS: -DPROGRAM='bash' -DCONF_HOSTTYPE='x86_64' -DCONF_OSTYPE='linux-gnu' -DCONF_MACHTYPE='x86_64-pc-linux-gnu' -DCONF_VENDOR='pc' -DLOCALEDIR='/usr/share/locale' -DPACKAGE='bash' -DSHELL -DHAVE_CONFIG_H -I. -I../. -I.././include -I.././lib -Wdate-time -D_FORTIFY_SOURCE=2 -g -O2 -fdebug-prefix-map=/build/bash-vEMnMR/bash-4.4.18=. -fstack-protector-strong -Wformat -Werror=format-security -Wall -Wno-parentheses -Wno-format-security uname output: Linux COOL4 4.4.0-17134-Microsoft #345-Microsoft Wed Sep 19 17:47:00 PST 2018 x86_64 x86_64 x86_64 GNU/Linux Machine Type: x86_64-pc-linux-gnu Bash Version: 4.4 Patch Level: 19 Release Status: release Description: It seems that the built-in 'test' does not work properly when run in a sub-shell. Repeat-By: Under Windows 10, WSL. Start "bash", terminal with shell pops up. 1. This works when typed directly: # make everything explicit export PATH="" # ensure that file1 exists and that file2 does not exist /bin/touch file1 /bin/rm -f file2 # built-in if [ file1 -nt file2 ]; then echo nt; else echo not_nt; fi # external if /usr/bin/[ file1 -nt file2 ]; then echo nt; else echo not_nt; fi # Output is as expected: nt nt 2. This does not work: # Put the above commands into a script, say check.sh # Run with: /bin/sh < check.sh # Or : /bin/sh ./check.sh # Or : /usr/bin/env ./check.sh # Output is always not ok: not_nt nt
Re: built-in '[' and '/usr/bin/[' yield different results
On Tue, Nov 13, 2018 at 05:29:42PM +0100, Service wrote: > Repeat-By: > Under Windows 10, WSL. Then why did you send this to a debian.org address? > Start "bash", terminal with shell pops up. > 2. This does not work: > > # Put the above commands into a script, say check.sh > # Run with: /bin/sh < check.sh > # Or : /bin/sh ./check.sh In both of these cases, you're explicitly calling /bin/sh rather than bash. So you are not using bash's builtin [. You are using WSL's sh's builtin [ command. You might start by identifying which shell WSL uses as its sh. > # Or : /usr/bin/env ./check.sh In this case, it would use the shebang line of check.sh to decide which shell to run. However, I suspect you did not include a shebang line at all. In that case, /usr/bin/env will decide how to handle the exec format error. It might decide to spawn sh for you, in which case this would explain why the result matches the first two results.
Re: built-in '[' and '/usr/bin/[' yield different results
On 13.11. 18:29, Service wrote: # Put the above commands into a script, say check.sh # Run with: /bin/sh < check.sh # Or : /bin/sh ./check.sh # Or : /usr/bin/env ./check.sh # Output is always not ok: not_nt nt $ cat check.sh export PATH="" /bin/touch file1 /bin/rm -f file2 if [ file1 -nt file2 ]; then echo nt; else echo not_nt; fi if /usr/bin/[ file1 -nt file2 ]; then echo nt; else echo not_nt; fi $ bash ./check.sh nt nt $ /bin/sh ./check.sh not_nt nt Isn't that Windows Linux thingy based on Ubuntu? /bin/sh isn't Bash by default on Debian and Ubuntu, so it might be you're just not running the script with Bash. -- Ilkka Virta / itvi...@iki.fi
Re: built-in '[' and '/usr/bin/[' yield different results
On 11/13/18 10:29 AM, Service wrote: # ensure that file1 exists and that file2 does not exist There's your problem. It is inherently ambiguous what timestamp to use when a file is missing (infinitely new or infinitely old, or always an error for not existing); bash's -nt picked one way, while other shells have picked the other. POSIX is silent on the matter (-nt is an extension outside of POSIX), so there is nothing portable you can rely on. /bin/touch file1 /bin/rm -f file2 # built-in if [ file1 -nt file2 ]; then echo nt; else echo not_nt; fi # external if /usr/bin/[ file1 -nt file2 ]; then echo nt; else echo not_nt; fi # Output is as expected: nt nt That is, bash's builtin '[' and coreutil's external '[' happened to pick the same thing: a missing file is treated as infinitely old. 2. This does not work: # Put the above commands into a script, say check.sh # Run with: /bin/sh < check.sh # Or : /bin/sh ./check.sh # Or : /usr/bin/env ./check.sh # Output is always not ok: not_nt nt Most likely, this is because your /bin/sh is not bash, but probably dash, and dash has picked a missing file as being treated as always an error. That does not make it a bug in bash, though, but a difference in behavior of your /bin/sh. -- Eric Blake, Principal Software Engineer Red Hat, Inc. +1-919-301-3266 Virtualization: qemu.org | libvirt.org
Re: Control characters in declare output
On 10/31/2018 11:01 PM, Rob Foehl wrote:
Prompted (pun intended) by the recent thread on detecting missing newlines
in command output, I'd had another look at my own version, and discovered
a potential issue with control characters being written as-is in declare
output. Minimal (harmless) reproducer:
╶➤ x () { echo $'\e[31m'"oops"$'\e[0m'; }
╶➤ declare -f x
x ()
{
echo ''"oops"''
}
Emits the string in red in a terminal. Any instances with control
sequences that do anything more invasive with the terminal cause more
BTW, to keep that red from turning your terminal red, I used:
read _CRST <<<"$(tput sgr0)" #Reset
read _CRED <<<"$(tput setaf 1)" #RED
read _CGREEN <<<"$(tput setaf 2)" #GREEN
read _CBLD <<<"$(tput bold)" #Bold
And in usage:
[[ $UID -eq 0 ]] && {
_prompt_open="$_CBLD$_CRED"
_prompt="#"
_prompt_close="$_CRST"
}
That way I can display the file on a screen without the control
characters
actually changing the color of the terminal.
(just in case you might be looking for a way around that behavior).
RE: bash will not link against ncursesw and readline in /usr/local
> > Using bash-4.4.18 > > Intel core i7 laptop running 32-bit or 64-bit linux Using gcc-8.2.0 > > > > The configure script does not find libncursesw on a system where > > only the wide version of ncurses exists - even when readine is linked > > against ncursesw. > > > I haven't seen a distro where ncursesw is installed without a link to ncurses. > Which distribution are you using? > The 64-bit version of tinycorelinux - all non-base packages are installed to /usr/local. Since ncursesw is now the default, I'm trying to compile against that. > I could add a check for ncursesw, but that's the kind of thing the distro > usually does. > If ncursesw is now the default, maybe it would make sense to check for that rather than a symlink? > > The configure scripts does not find libreadline when it is compiled to > > /usr/local and when using the configure switch > > "--with-installed-readline=/usr/local" > > > I don't have any trouble finding readline in /usr/local/lib/libreadline.so > after > installing it, editing /etc/ld.so.conf, and running ldconfig. I tried with > readline-8.0-beta > and bash-5.0-beta, so at least it will be working when those hit release > status. > It appears that the readline check relies on the ncurses check being successful. If I configure without an ncurses symlink the check for readline fails. If I add an ncurses symlink the check for readline succeeds. checking for tgetent... no checking for tgetent in -ltermcap... no checking for tgetent in -ltinfo... no checking for tgetent in -lcurses... no checking for tgetent in -lncurses... no checking which library has the termcap functions... using gnutermcap checking version of installed readline library... configure: WARNING: Could not test version of installed readline library. configure: WARNING: installed readline library is too old to be linked with bash configure: WARNING: using private bash version configure:5194: result: no configure:5213: checking which library has the termcap functions configure:5216: result: using gnutermcap configure:5242: checking version of installed readline library configure:5296: gcc -flto -fuse-linker-plugin -mtune=generic -Os -pipe -o conftest -g -O2 -Wno-parentheses -Wno-format-security -I/usr/local/include -L./lib/termcap -L/usr/local/lib conftest.c -lreadline ./lib/termcap/libtermcap.a >&5 gcc: error: ./lib/termcap/libtermcap.a: No such file or directory $ cd /usr/local/lib $ sudo ln -s libncursesw.so.6.1 libncurses.so $ sudo ldconfig checking for tgetent... no checking for tgetent in -ltermcap... no checking for tgetent in -ltinfo... no checking for tgetent in -lcurses... no checking for tgetent in -lncurses... yes checking which library has the termcap functions... using libncurses checking version of installed readline library... 7.0
