Le Mon, Feb 11, 2008 at 01:40:09PM +0100, David Paleino a écrit : > Il giorno Mon, 11 Feb 2008 20:51:25 +0900 > Charles Plessy <[EMAIL PROTECTED]> ha scritto: > > > (By the way, I am swiching a lot of packages to CDBS; if it is an issue, > > I am of course open to discussion. I am usually doing this only if the > > resulting debian/rules files is really minimal). > > (I don't really like CDBS, it's a "black box" to me. I like to have full > control on what debian/rules does, and including huge segments > from /usr/share/cdbs/ is probably the worst solution, when you can write > something *really* simple by hand. But it's IMHO :)
Hi all, the "compat" example in this thread is a good indication of the good reasons to be cautious with CDBS. This said, many of the programs we package are very simple, and I think that the advantage of CDBS, beyond having debian/rules being compacted in a dozen of lines, is that the code is factorized and that future complexifications of the build process, such as the introduction of novel build options like "nodoc", will be applied without effort for us. I think that we all agree that the most complex packages should not use a CDBS makefile with plenty of variables set to change the default behavior. I do not find an easy quantitative way to formalise this in our policy. Please feel free to propose one. At the very worse, something like "If you use CDBS and the makefile is more than 12 lines, please do seriously consider rewriting it using debhelper directly". Have a nice day, -- Charles Plessy http://charles.plessy.org Wakō, Saitama, Japan -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
