Re: Re: cvs: pearweb /public_html/trackback trackback-admin.php

Helgi Þormar wrote: > On Fri, 21 Oct 2005 22:43:55 -0400, Greg Beaver wrote: > > >>I think the best thing to do is to test prior to commit, or to post a >>patch and ask for testing if it's not possible to test. Committing a >>broken patch is far worse than not committing a possible fix. >> >>Committing broken code to the live website that represents the PEAR >>project is a really big deal, which brings me to a another point: why is >>this even *possible*? We all know the answer, but I think it is not a >>good one :). > > > This is more the question about security then about if the code broke or > not, I'd think that for Pierre to do this all by him self on a private > server or even if we helped would have taken way too much time compare to > the security issues we had before register globals was turned off, so I > rather want to have a half broken site that we fix on the fly (i.e. when > we spot the issues, be it in the code or on the site it self) then have > massive security problems that could lead to worse PR and just plain > problems if someone exploits these problems. Let me absolutely clear: there have been a large number of "fixing broken commit" commit messages over the past 2-3 months in pearweb. This is a pattern not specifically related to any one commit, otherwise I would not mention it. Even so, register_globals has been disabled in PHP by default since PHP version 4.2, and the security problems with register_globals have been known for several years. We have yet to have had an attack, as far as I know. Why the sudden need to change the code without testing it? This is a bogus argument. Committing code without testing it is asking for both fatal errors and whole new security vulnerabilities. If it is too much to ask that we test code before committing, then I suggest we take a SERIOUS look at our priorities. Greg