| Summary: | script: closing stdin causes immediate exit of 'script' (but the child process runs in background) | ||
|---|---|---|---|
| Product: | Red Hat Enterprise Linux 7 | Reporter: | Pavel Raiskup <praiskup> |
| Component: | util-linux | Assignee: | Karel Zak <kzak> |
| Status: | CLOSED WONTFIX | QA Contact: | qe-baseos-daemons |
| Severity: | high | Docs Contact: | |
| Priority: | unspecified | ||
| Version: | 7.2 | CC: | fweimer, jpazdziora, kdudka, praiskup |
| Target Milestone: | rc | ||
| Target Release: | --- | ||
| Hardware: | Unspecified | ||
| OS: | Unspecified | ||
| Whiteboard: | |||
| Fixed In Version: | Doc Type: | If docs needed, set a value | |
| Doc Text: | Story Points: | --- | |
| Clone Of: | Environment: | ||
| Last Closed: | 2020-12-15 07:46:01 UTC | Type: | Bug |
| Regression: | --- | Mount Type: | --- |
| Documentation: | --- | CRM: | |
| Verified Versions: | Category: | --- | |
| oVirt Team: | --- | RHEL 7.3 requirements from Atomic Host: | |
| Cloudforms Team: | --- | Target Upstream Version: | |
| Bug Depends On: | |||
| Bug Blocks: | 1465901 | ||
I have doubts that 0>&- is the right syntax to closes stdout. See http://www.tldp.org/LDP/abs/html/io-redirection.html 0<&-, <&- Close stdin. 1>&-, >&- Close stdout. RHEL-7: $ script -e -c 'time sleep 1' 1>&- script: write error: Bad file descriptor the option -q hides the error message: $ time script -q -e -c 'time sleep 1' 1>&- real 0m1.011s user 0m0.003s sys 0m0.009s So it works for me. You are right, sorry, there's typo in $Summary. Also this (from original report): script -q -e -c 'sleep 3 ; echo why' /dev/null < /dev/null Script for some reason loses control over it's child processes when one plays a bit with stdin. Seems to be fixed upstream. If there's no way to avoid closing stdin due to environment constraints,
can we mention here in bugzilla proper work-around?
Or do we have other command as an alternative to this feature ("simulate"
fully _working_ terminal with fd <= 2 opened)?
(In reply to Pavel Raiskup from comment #8) > If there's no way to avoid closing stdin due to environment constraints, > can we mention here in bugzilla proper work-around? Sorry, trying to better describe the motivation: It is possible that user's script is run in environment (shell) where stdin is closed (by shell); for such situations it would be nice to have at least work-around. Is this the right way? #! /bin/sh command_with_closed_stdin () { script -q -e -c 'sleep 1 ; test -t 2 && echo why' /dev/null </dev/ptmx } command_with_closed_stdin 0>&- Being the person who initially inquired about this topic, merely running the program I care about with < /dev/ptmx without script did the trick for me. Yes, 'COMMAND 0>/dev/ptmx 1>/dev/ptmx 2>/dev/ptmx' makes the 'COMMAND' work, but I'm not sure it is trivial record what's written to output. Is it possible without script? This is often used for testing purposes, if 'TEST' tries to check that 'COMMAND' behaves differently with terminal attached and without attached (e.g. tar's testsuite used to have such test, dropped by 30fe67226a7cfc52 because of limited portability). 'TESTSUITE' running 'TEST' might be run with and without terminal, but results for 'TEST' should be the same. Maybe we can rebase script command for RHEL7.5 if the issue is really critical (important use-case). After evaluating this issue, there are no plans to address it further or fix it in an upcoming release. Therefore, it is being closed. If plans change such that this issue will be fixed in an upcoming release, then the bug can be reopened. |
For example: $ time script -q -e -c 'time sleep 1' 0>&- real 0m0.004s user 0m0.000s sys 0m0.003s $ real 0m1.001s user 0m0.000s sys 0m0.001s Doesn't happen on Fedora 24. At least because there is '-e' used, this looks like a bug rather than undefined behavior.