Your message dated Wed, 22 May 2013 22:30:26 +0000 (UTC) with message-id <[email protected]> and subject line Re: Bug#709382: mksh: broken Built-Using handling has caused the Debian Bug report #709382, regarding mksh: broken Built-Using handling to be marked as done.
This means that you claim that the problem has been dealt with. If this is not the case it is now your responsibility to reopen the Bug report if necessary, and/or fix the problem forthwith. (NB: If you are a system administrator and have no idea what this message is talking about, this may indicate a serious mail system misconfiguration somewhere. Please contact [email protected] immediately.) -- 709382: http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=709382 Debian Bug Tracking System Contact [email protected] with problems
--- Begin Message ---Package: mksh Version: 46-1 Severity: serious Dear Maintainer, The handling of built-using is wrong. It is not meant to encode the compiler used, nor binutils or kernel headers should be recorded there It is specifically for building against -source packages and for hacks like ia32libs where binaries are copied into a source package. Not for 'everything'. What you effectively are doing is asking for a mksh rebuild on each upload of the kernel, of gcc and what else you have put there. Imagine if we all did that. Buildd power. or mirror size. or both. /Sune -- System Information: Debian Release: jessie/sid APT prefers unstable APT policy: (500, 'unstable'), (1, 'experimental') Architecture: amd64 (x86_64) Foreign Architectures: i386 Kernel: Linux 3.8-1-amd64 (SMP w/8 CPU cores) Locale: LANG=en_US.UTF-8, LC_CTYPE=en_US.UTF-8 (charmap=UTF-8) Shell: /bin/sh linked to /bin/dash Versions of packages mksh depends on: ii libc6 2.17-3 ii libgcc1 1:4.8.0-7 mksh recommends no packages. Versions of packages mksh suggests: ii ed 1.8-1 -- no debconf information
--- End Message ---
--- Begin Message ---Sune Vuorela dixit: >The handling of built-using is wrong. It is not meant to encode the >compiler used, nor binutils or kernel headers should be recorded there Policy 3.9.4 §7.8 says: Some binary packages incorporate parts of other packages when built but do not have to depend on those packages. Examples include linking with static libraries or incorporating source code from another >It is specifically for building against -source packages and for hacks >like ia32libs where binaries are copied into a source package. Not for >'everything'. In this specific case, there are one to two statically linked programs there. In some cases, they link statically against a GPL licenced library. So my current interpretation of the text from Policy above says that Built-Using is indeed required there. >What you effectively are doing is asking for a mksh rebuild on each No, just that dak keeps the source versions around for longer. A final rebuild near the end of the freeze should be enough, if it is indeed needed at all. (If dak just keeps the relevant source packages at hand, and they end up on the source CDs, I believe all requirements are met, and IIRC reading that this, not rebuilding, is how things are handled on Debian side; the only requirement is that, upon a binary *entering* the archive, the source packages in that precise version must be known to dak, i.e. not already superseded (by newer version, NBFAS or removal); once a package B-Us them they will not be removed. I’m closing this as not a bug. Please feel free to file a bug against the Policy wording in the meantime; as things are now, the wording specifically includes statically linked binaries. The composition of B-U in mksh is as follows: • mksh-static is always built statically; either against klibc (plus linux-libc-dev and libgcc), or against dietlibc (plus libgcc), or against eglibc (plus libgcc) • lksh is built statically if klibc or dietlibc are available, with the same “plus” as above • the build script records what was actually put into the binaries, gets the appropriate source package relationships from that and puts it into Built-Using In the dietlibc case at the very least (since it’s GPL; would have to look closer at others), the resulting binary is fully covered by the requirement of the GPL that its precise and complete sources be available. I don’t mind changing this *at all*, but I can only do that (justifiedly) if it’s not against what I believe to interpret Policy correctly (especially since it specifically lists static linkage), or if there’s a CTTE resolution asking me to change it. I’ve had more troubles with this B-U ever since using it in experi‐ mental, so… really, no argument from me, just following Policy (with some background in licencing and toolchains). bye, //mirabilos -- I believe no one can invent an algorithm. One just happens to hit upon it when God enlightens him. Or only God invents algorithms, we merely copy them. If you don't believe in God, just consider God as Nature if you won't deny existence. -- Coywolf Qi Hunt
--- End Message ---
