From mboxrd@z Thu Jan 1 00:00:00 1970 From: Thomas Petazzoni Date: Sat, 31 Jan 2015 23:06:20 +0100 Subject: [Buildroot] Best-Practice Suggestions for developing package patches in buildroot In-Reply-To: References: Message-ID: <20150131230620.2236108e@free-electrons.com> List-Id: MIME-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: 7bit To: buildroot@busybox.net Dear Bryce Schober, On Thu, 29 Jan 2015 10:09:55 -0800, Bryce Schober wrote: > Do you use something like quilt to manage patches while developing them? > > Do you fork the upstream repository a switch the package to use the fork, > then back out your modifications into patches? > > Do you just copy the full commands generated by buildroot and use re-use > them outside of buildroot's high-level make commands? Either quilt or git, but more and more I use git. I generally clone the upstream repository somewhere completely separate from Buildroot, then create a branch called 'buildroot' with the starting point being the tag indicating the release of the software currently in use by Buildroot. Then, I import as separate Git commits each of the individual patches that Buildroot has for this package (if any). I do my work, and then use 'git format-patch' to format the patches and copy them back in Buildroot. Quite ironically, I never use _OVERRIDE_SRCDIR: I simply copy the patches back to the Buildroot package directory, and restart the package build process from scratch. Yes, this is inefficient, but one could pretend that it forces you to think twice before testing a stupid change :-) Also, I tend to very often start by hacking directly in output/build/-/, and once I have a good idea of the change that needs to be done, I do the Git work flow described above. Hope this helps, Thomas -- Thomas Petazzoni, CTO, Free Electrons Embedded Linux, Kernel and Android engineering http://free-electrons.com