From mboxrd@z Thu Jan 1 00:00:00 1970 From: Kay Sievers Date: Thu, 08 Apr 2004 14:16:24 +0000 Subject: Re: Test script for udevd binary, Request for Comments! Message-Id: <20040408141624.GB30979@vrfy.org> List-Id: References: <3ACA40606221794F80A5670F0AF15F84037B9188@PDSMSX403.ccr.corp.intel.com> In-Reply-To: <3ACA40606221794F80A5670F0AF15F84037B9188@PDSMSX403.ccr.corp.intel.com> MIME-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: 7bit To: linux-hotplug@vger.kernel.org On Wed, Apr 07, 2004 at 06:53:21PM +0800, Yin, Hu wrote: > Hi, All, > > I have written a test script for udevd binary. As you know there have been > some good test scripts for udev binary but we didn't write the corresponding > test scripts for the new udev's binaries since udev is split into several > binary program. Moreover, I think the work is necessary, especially for > udevd and udevsend binaries. > > Now I focus on the validation of udev as a Intel's intern student, so wrote > this test script for udevd and send it to all. I know this script is not > very good and enough for udevd's test but I believe we can do better with > the help from all of you. So please just take a look at this script and give > me some advises and suggestions in order that we can improve it together. The timout is 10 seconds now. It is not acceptable for a test like this, to work on the "real" $udev_root. It may render your system unusable! I prefer a test for udevsend/udevd only, not calling the real udev. You may change udevd to look on startup in the environment for the key UDEV_BIN and take this value instead of the real udev. Then you replace the real udev by a call to a small test program, which maybe writes a log file to be analyzed by your test script. This way you test the daemon only, not udev and udevd together and you are able to check if the logic for holding back events for the same device and execute different devices in parallel, works too. And a second time: Two events with the same sequence number are not a case we need to handle differently than a missing sequence number. A test for it is nice, but a timeout is the expected behavior. What do you think? thanks, Kay ------------------------------------------------------- This SF.Net email is sponsored by: IBM Linux Tutorials Free Linux tutorial presented by Daniel Robbins, President and CEO of GenToo technologies. Learn everything from fundamentals to system administration.http://ads.osdn.com/?ad_id70&alloc_id638&op=click _______________________________________________ Linux-hotplug-devel mailing list http://linux-hotplug.sourceforge.net Linux-hotplug-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/linux-hotplug-devel