Re: Re: [pear-webmaster] Re: [PEAR-CORE] Re: cvs: pearweb /public_html/trackback trackback-admin.php

Pierre Joye wrote: > On 10/24/05, Greg Beaver <greg@chiaraquartet.net> wrote: > >>Pierre wrote: >> >>>On Fri, 21 Oct 2005 22:43:55 -0400 >>>cellog@php.net (Greg Beaver) wrote: >>> >>> >>> >>>>1) if you haven't tested the code, don't commit it >>>>2) if you can't test it, post a patch to pear-webmaster >>>>3) if you've tested it, commit it >>>>4) if it's obvious you haven't tested it, you get a warning. On the >>>>second commit of obviously untested code, your cvs karma is revoked >>>>for 4 weeks, minimum. >>> >>> >>>For your information, it was < 1 day for both "problems" together (both >>>were not critical for our infrastructure). The last change the problem >>>reporters did blocked pearweb for 5 days, who needs to loose anything? >>> >>>Short version, move on unless you have something usefull to say. >> >>I've been unable to get a great internet connection, but now I have one, >>so I will send a quick message. >> >>Pierre, I would like to point out that I said "test before committing" >>which has nothing to do with register_globals. (Anatoly, are you >>listening?) >> >>"committing" refers to CVS. I also replied to a CVS commit message. I >>would like to point out this message from October 12, 2005 as an example >>of a blatant commit without any testing whatsoever. I am referring to >>commits of this nature [names and such removed to protect the guilty]: > > > Demagogy, pollitics (or BS in short). > > I did this commit. Do you know _WHY_ I did it urgently? Any clue? > > Because some people thinks they can update PHP like they change their > shirt, or change apache or any other part of the server. > > Yes I prepared and tested a fix for this problem (globals being > ignored in 4.4 and some other apc), my mistake, some silly CS fixes > right before commit and leave for an important meeting. > > So, and excuse my tone, shut up with that now and continue to work on > something more usefull and constructive and spare me such pathetic > ideas. Pierre: all I'm asking is that every developer test code prior to committing it. I think a better procedure in the case of not enough time would be to capture the output of "cvs diff -u" and send it to the mailing list for another with karma to commit. I would welcome constructive coding from you or anyone else in the future, always have. Let's be honest: nobody is claiming to be an angel. I have committed code without testing in the past, I think it is fair to say that all (or most) of us have done so. The thing is, it has never turned out to be a good idea in a live website, and I would suggest we discourage it. In fact, *all* changes to the live website need to be testable, just as they are for the PEAR package. This includes PHP version, register_globals, PEAR version on the website, etc. Part of what will make changing things testable is to re-factor to de-couple the entire website from pear-prepend/pear-database. This means that PEPr/bugs/trackbacks could be re-factored slightly, to use a custom file containing database access code. In this way, it would be possible to swap in a different core for the main website, or make other changes of that nature. There is too much code that reads "where does that function come from?" and requires a full-text search of cvs.php.net/pearweb. This makes any changes risky and difficult to verify/test. I simply don't think it is useful to point fingers and dwell on the past, only to remedy problems we've had for the future. Please keep the tone professional, there is no good cause for obscenity and insults, this is completely unacceptable behavior on a public mailing list. Thanks, Greg