From mboxrd@z Thu Jan 1 00:00:00 1970 From: =?UTF-8?Q?Andr=c3=a9_Hentschel?= Date: Tue, 31 May 2016 23:56:51 +0200 Subject: [Buildroot] [PATCH] [PATCH v2] p7zip: New package In-Reply-To: <20160530152042.5acb6bb0@free-electrons.com> References: <1464539299-3853-1-git-send-email-nerv@dawncrow.de> <20160529171412.GA3367@free.fr> <574B256A.4010502@dawncrow.de> <20160530152042.5acb6bb0@free-electrons.com> Message-ID: <574E08A3.3070204@dawncrow.de> List-Id: MIME-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: 7bit To: buildroot@busybox.net Am 30.05.2016 um 15:20 schrieb Thomas Petazzoni: > Hello, > > On Sun, 29 May 2016 19:22:50 +0200, Andr? Hentschel wrote: > >>>> +config BR2_PACKAGE_P7ZIP >>>> + bool "p7zip" >>>> + depends on BR2_INSTALL_LIBSTDCPP >>>> + depends on BR2_i386 || BR2_x86_64 || \ >>>> + BR2_aarch64 || BR2_arm || BR2_armeb >>> >>> Please explain why this is limited to those architectures. >>> >>> You can put it in the commit log, or you may add a terse comment just >>> above the depends-on line. If the explanations is not straightforward, >>> prefer the commit log. >>> >>> (Yes, I read the previous reply that it was broken on NIOSII. Limiting >>> to a small set of architectures wihtout an explanation is not enough.) >>> >>> Thanks! :-) >> >> Can I start with a small set of archs until it is tested on more? >> (with a comment then anyway of course) > > I will disagree with Yann here. We typically do not limit the > architectures based simply on what has been tested or not. > > A dependency on specific architectures should only be added if the > package is either 1/ contains some architecture-specific code or 2/ is > known to not build/work properly on some architectures. > > In any case, the fact that you could only test on i386/AArch64/ARM is > not a proper reason to restrict the package to those architectures only. I simply used my autobuild machine now to test on almost all archs, bfin is the only troublemaker. > Building p7zip for blackfin using > > http://autobuild.buildroot.net/toolchains/configs/bfin-uclinux.config > > as defconfig fails with > > ../../../../CPP/myWindows/mySplitCommandLine.cpp:99:8: error: #error ENDIANNESS > > Regards, Bernd I tried it, the compiler doesn't expose ENDIANess defines... reason unknown...