From mboxrd@z Thu Jan 1 00:00:00 1970 From: Yann E. MORIN Date: Thu, 26 Mar 2020 21:56:58 +0100 Subject: [Buildroot] [PATCH v2] package/libapparmor: new package In-Reply-To: References: <20200326180115.30643-1-angelo@amarulasolutions.com> <20200326185659.GT22325@scaer> Message-ID: <20200326205658.GY22325@scaer> List-Id: MIME-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: 7bit To: buildroot@busybox.net Angelo, All, On 2020-03-26 21:34 +0100, Angelo Compagnucci spake thusly: > On Thu, Mar 26, 2020 at 7:57 PM Yann E. MORIN wrote: > > On 2020-03-26 19:01 +0100, Angelo Compagnucci spake thusly: > > > From: Angelo Compagnucci > > > This patch adds libapparmor and it's related tools. > > > The patch is quite complicated by the layout of the source tree: [--SNIP--] > > I've looked at the .mk, and I don't like it. [--SNIP--] > > Why don't you provide multiple packages: > > - libapparmor [--SNIP--] > > - apparmor-utils, with just the parser (and binutils?) sub-dirs > > - pam > > - apache > > - python > > - profiles > > - rules caching > I don't know. I've tried that approach at in the end it was a mess. > Some of the steps to build the swig python are embedded into the > makefile, so we need to call configure and make even for a package > that instead could have been a simple python one. Well, as far as I can see, that's exactly what your patch does: it installs libapparmor, and then as post-staging hooks, it then builds the rest of the package. This is exactly what having two packages would provide. Now, specifically about the python bindings: maybe they should be built from the libapparmor package rather than the utils one, sure, if it makes more sense... > You mean having a patch series that will add bit by bit to the package? Yes. As you say yourself, the package is a mess as it is. By splitting it in a series that adds each pieces one by one, it will: - allow you to provide a detailed commit log with full explanations about the required uglyness, - allow reviewers to understand that problem and better asses the uglyness, and see if it is indeed needed. Also, "it was a mess" is not descriptive enough to dismiss the multi-package attempt (where 'multi' may well be just '2'). Regards, Yann E. MORIN. -- .-----------------.--------------------.------------------.--------------------. | Yann E. MORIN | Real-Time Embedded | /"\ ASCII RIBBON | Erics' conspiracy: | | +33 662 376 056 | Software Designer | \ / CAMPAIGN | ___ | | +33 561 099 427 `------------.-------: X AGAINST | \e/ There is no | | http://ymorin.is-a-geek.org/ | _/*\_ | / \ HTML MAIL | v conspiracy. | '------------------------------^-------^------------------^--------------------'