On Thu, Sep 26, 2002 at 09:31:44PM -0400, Ilia A. wrote: > > PHP is full of tweaks and hacks specifically to augment and make easier the > > job of the people using it. > > The only simular hack I can imagine you are referring to is safe_mode, which > is a very ugly thing that was only implemented because at the time it was not > possible to implement simular functionality on most web servers.
i'm thinking more in line with the extensions. like --enable-ucd-snmp-hack for example. the documents tell the user to modify the ucd-snmp code, but if the user is doesn't do that, they can simply add that to ./configure and away you go. > > if php-dev is not interested in such suggestions, i'll back off and not > > bother you again. > > sorry if my providing a complete patch to add this functionality took you > > away from more important things. > > I am hardly in a position to say what all the readers and developers on this > list think of this patch. I merely try to explain why I believe this > particular patch is not appropriate for standard PHP distribution. sorry, i might have become a little overheated there. i've run into this kind of problem before in other environments and it gets me going. i'm just a user of the code, and i've come up with what i think is a beneficial augmentation. i spent some time investigating the structure of the code, and trying to ensure that my patch meets with the existing standard of coding (ie diagnostics, memory allocation/deallocation, etc, etc). i don't have a vision of what PHP should or shouldn't do at the development level. i'm just trying to solve production problems. i feel that maybe the feeling of this group is too much purist, and too little production. (i realize that's painting with a wide brush, sorry) it would be wonderful if we could spin off children of the web/php process into their own uid/gid, but that is not easily done (without cgi). it would also be wonderful if pgsql had intimate knowledge of the innards of the calling process, but that is not easily done. i thought about using environment variables to go httpd.conf -> php -> pgsql, but those are too easily over-ridden by the user's code. even if pgsql had a mechanism to restrict which dataset a process was attaching to, we would still need to pass the list of restrictions from httpd.conf, through php, to pgsql. as best i can tell, this is the most efficient and safe way to do this. -- [ Jim Mercer [EMAIL PROTECTED] +1 416 410-5633 ] [ I want to live forever, or die trying. ] -- PHP Development Mailing List <http://www.php.net/> To unsubscribe, visit: http://www.php.net/unsub.php
