On Mon, Oct 29, 2001 at 02:41:43PM -0500, Rodent of Unusual Size wrote: > Marc Slemko wrote: > > > > It is the start of a very ugly trend.
I fear I have to agree with Marc here. > Okey, I'm convinced. It would be better to have this > done by httpd itself in some way in which apachectl > can detect the failure -- like doing the setuid() and > setgid() before detach(). (Having apachectl check the > log for the start message is just as ugly as my original > idea because of the locating and parsing the logfile > issue.) IMHO this is an issue for "make install" and neither for apachectl nor for httpd's startup phase. "Make install" inserts "Group #-1" into the config, and could just as well do a quick check over "#-1", "nobody" and "nogroup". Anything the admin will configure after that is up to her, and she better know where the error_log file is. With that in mind, I agree on the source of checkgid being in the source tree, but I disagree a little (-0.5) with it being installed as a support tool proper. On a different target machine (in the case of binbuild.sh), groups may be configured differently, but that should IMO be handled by the administrator there. Just my $.02 Martin -- <[EMAIL PROTECTED]> | Fujitsu Siemens Fon: +49-89-636-46021, FAX: +49-89-636-47655 | 81730 Munich, Germany
