From mboxrd@z Thu Jan 1 00:00:00 1970 From: Ludovic Desroches Date: Thu, 07 Apr 2011 11:32:27 +0200 Subject: [Buildroot] Patching method Message-ID: <4D9D84AB.3080101@atmel.com> List-Id: MIME-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: 7bit To: buildroot@busybox.net Hello, I was building a new file system and I have noticed something about patching that seems a little dangerous : >>> liblockfile 1.08 Patching package/liblockfile toolchain/patch-kernel.sh /home/ldesroches/buildroot/output/build/liblockfile-1.08 /home/ldesroches/buildroot/dl liblockfile_1.08-4.debian.tar.bz2 ... + echo 'Applying liblockfile_1.08-4.debian.tar.bz2 using bzip2: ' Applying liblockfile_1.08-4.debian.tar.bz2 using bzip2: + echo liblockfile_1.08-4.debian.tar.bz2 + cat + bunzip2 -dc /home/ldesroches/buildroot/dl/liblockfile_1.08-4.debian.tar.bz2 + patch -p1 -E -d /home/ldesroches/buildroot/output/build/liblockfile-1.08 patching file dotlockfile.c patching file lockfile.c missing header for unified diff at line 320 of patch patching file COPYRIGHT patching file dotlockfile.1 patching file Makefile.in patch unexpectedly ends in middle of line ... As you can see, we have an archive and we don't extract it, we simply unzip the compressed file and send the tar file to the patch command. The content of this archive is a tree with files and directories (one of them is the patch directory) that's why I think this method is a little dangerous. The patches are not applied in the same order as defined into the series file because the patch order is not kept into the archive. Am I wrong ? Do you think it can be a problem ? Regards Ludovic Desroches