Package: osdsh Version: 0.7.0-9 Severity: normal Tags: patch While upgrading osdsh from a local version to 0.7.0-9, I noticed that osdctl started to hang again when executed directly after osdsh. Even with patch 09-process-running-checks.dpatch applied, there remains a race condition.
With the current design, the osdsh parent process forks a child process, writes the child's pid to a file and exits. If osdctl is started directly in succession, the fifo might not exist, as the child process might not have created it yet. With my local version, I had tested a real fix for this problem (and subsequently forgotten about the patch). The only way to know for sure that the fifo exists after osdsh has exited is to simply create the fifo *before* forking the child process. Regards, Peter -- System Information: Debian Release: lenny/sid APT prefers unstable APT policy: (500, 'unstable'), (500, 'stable'), (400, 'testing'), (1, 'experimental') Architecture: amd64 (x86_64) Kernel: Linux 2.6.22-maia (PREEMPT) Locale: LANG=en_US.UTF-8, LC_CTYPE=en_US.UTF-8 (charmap=UTF-8) Shell: /bin/sh linked to /bin/bash Versions of packages osdsh depends on: ii libapm1 3.2.2-8.1 Library for interacting with APM d ii libc6 2.6-1 GNU C Library: Shared libraries ii libxosd2 2.2.14-1.3 X On-Screen Display library - runt ii tk8.4 [wish] 8.4.12-1 Tk toolkit for Tcl and X11, v8.4 - osdsh recommends no packages. -- no debconf information
10-osdsh-create-fifo-before-fork.dpatch
Description: application/shellscript
