Hi, thanx for the answer. I wrote an bug report, but while I was searching the source code for the right place to create an patch I was astonished that I couldn't reproduce the error in a newly compiled version, though I didn't change anything in the source code.
Further research resulted in the following (reproducable) behavior. As long as the httpd is not running, the -t option behaves as expected an results in 1 as return code, as long as there are errors in the configuration. When the httpd is running, the return code is always 0, no matter how the configuration looks like. First I thougt that might be the result of an suexec patch of mine and removed the bug report, but some more testing showed that the patch didn't cause this behavior. Now I'm unsure whether this is realy a bug, or the behavior as wanted. Personally I would say it should be regarded as bug, as I think it's quite usefull to check the configuration automatically before the server is restarted. In an production shared hosting environment this could avoid downtimes for hundreds of websites because of an unconcentrated, stressed administrator. What do you think? Is this an bug, or not? Ciao Richard ----- Ursprüngliche Nachricht ----- Von: Jeff Trawick Gesendet: 10.04.12 17:03 Uhr An: users@httpd.apache.org Betreff: Re: [users@httpd] httpd 2.4.1 -t option return code results always in 0 On Tue, Apr 10, 2012 at 9:57 AM, Richard Westebbe <richard.weste...@gmx.de> wrote: > Hi, > > it looks like I stumbled across an bug in httpd 2.4.1, when I tried to > modify the init script to make a configtest before a restart. > > I evaluate the return code of the -t option, to decide whether I should > restart the web server or not. Regarding to the manpage > this option should result in 0 when there's no error in the config, > otherwise it should result in 1. > > With httpd 2.4.1 the result is 0 in every case, no matter how the config > looks like. With httpd 2.2 it works like expected. I would say > that's bug, or does anybody know something about it? > > Is there a patch available? I'll gonna have a look into the source code to > see whether I can write a patch for it on my own, but if there is > an official patch I would prefer to use that. I didn't see an existing bug or fix for this. File a bug -- with or without patch -- at https://issues.apache.org/bugzilla/ -------------- ------------------------------------------------------- To unsubscribe, e-mail: users-unsubscr...@httpd.apache.org For additional commands, e-mail: users-h...@httpd.apache.org
