(Replying 2 years later because you forgot to Cc: me on this :-) On Sat, 13 Dec 2003, Roberto Lumbreras wrote:
> tags 217776 + wontfix > stop > > I don't know why a binary included in the upstream .tgz is a bug, the > file is deleted and sources recompiled to generate it... It is a bug because binary executables belong to binary packages, not source packages. Also, the source is supposed to be clean. If you unpack the source and do "debian/rules clean" and some file is deleted, such file does not belong to the source to begin with. If this sounds too much as a "matter of principles" reason, we can consider what could happen in practice: You are running the risk of the executable not being regenerated and being shipped in the .deb "as is", which would be a disaster, since it would be a wrong binary for at least 10 different architectures. In this case, you are just being lucky, as eql_enslave.c is modified by the Debian .diff.gz, which forces a rebuild, but imagine what could happen if: a) the author releases eql 1.3. b) source still contains an executable. c) Debian .orig.tar.gz is still the same as upstream. d) you don't need to patch eql_enslave.c this time. e) you use essentially the same debian/* files to package it. Then, it would be a real disaster. In some sense, this is like a time bomb. The fact that it did not explode yet just means you are a lucky man. > neither I know how to fix it smoothly, so I'm tagging it as wontfix ;-) You can repackage the source and add some tag to the version, as it is usually done when there is need to remove any material which is not dfsg-free. For example, you could call it eql_1.2.debian.orig.tar.gz or eql_1.2.clean.orig.tar.gz or maybe eql_1.2.0.orig.tar.gz. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
