Re: release policy changes for etch

2005-06-30 Thread Frank Lichtenheld
On Thu, Jun 30, 2005 at 11:01:17AM -0300, Gustavo Noronha Silva wrote: > I have a question about this requirement: > > Packages must not install programs in the default PATH with > different functionality with the same file name, even if they > Conflict:. > > Totem and Glade, fo

Re: release policy changes for etch

2005-06-30 Thread Gustavo Noronha Silva
Em Ter, 2005-06-28 às 00:43 +0200, Andreas Barth escreveu: > Hi, All, Hey, > As always, the full canoncial list of the RC policy for etch is available > at http://release.debian.org/etch_rc_policy.txt . I feel there's too much redundancy there when taking debian policy into account, is that redu

Re: release policy changes

2005-06-13 Thread Thomas Bushnell BSG
Russ Allbery <[EMAIL PROTECTED]> writes: >> If the ABI changes but not the SONAME then the only sane thing to do is >> to encode the ABI in the package name, e.g. libstc++-6c1002. Just like >> for any other lib doing the upcoming c++ abi transition. > > We agree on this. So just to make sure we'r

Re: release policy changes

2005-06-13 Thread Thomas Bushnell BSG
Russ Allbery <[EMAIL PROTECTED]> writes: > I think you're still too caught up on libstdc++, the library. The library > is not the C++ ABI, just like how libgcc is not used by all C programs. > The C++ ABI involves things like how classes are laid out and how function > names are mangled. So what

Re: release policy changes

2005-06-13 Thread Russ Allbery
Goswin von Brederlow <[EMAIL PROTECTED]> writes: > Russ Allbery <[EMAIL PROTECTED]> writes: >> I think you're confusing the C++ ABI with the SONAME of libstdc++. >> They're not necessarily the same thing, although right now they tend to >> change at the same time. Having the SONAME of libstdc++ c

Re: release policy changes

2005-06-13 Thread Goswin von Brederlow
Russ Allbery <[EMAIL PROTECTED]> writes: > Goswin von Brederlow <[EMAIL PROTECTED]> writes: > >> [EMAIL PROTECTED]:~% apt-cache show mozilla-dev >> Package: mozilla-dev >> Architecture: amd64 >> Source: mozilla >> Version: 2:1.7.8-1 >> Depends: mozilla-browser (= 2:1.7.8-1), libnspr-dev (= 2:1.7.8

Re: release policy changes

2005-06-13 Thread Russ Allbery
Goswin von Brederlow <[EMAIL PROTECTED]> writes: > [EMAIL PROTECTED]:~% apt-cache show mozilla-dev > Package: mozilla-dev > Architecture: amd64 > Source: mozilla > Version: 2:1.7.8-1 > Depends: mozilla-browser (= 2:1.7.8-1), libnspr-dev (= 2:1.7.8-1), libxt-dev, > libc6 (>= 2.3.2.ds1-21), libgcc1

Re: release policy changes

2005-06-13 Thread Goswin von Brederlow
Matthias Klose <[EMAIL PROTECTED]> writes: > How can you tell, for which C++ ABI packages like mozilla-dev and > festival-dev are built? IMO that is the technical reason you are > asking for. It doesn't matter if an application or a library is linked > to libstdc++. Remember that C++ ABI != libstd

Re: release policy changes

2005-06-09 Thread Florian Weimer
* Andreas Barth: >> Wouldn't it be sufficient to defer to generic policy on this one? PIC >> library/DSO support is mostly a C/C++-specific domain, and not >> everything that Debian ships is written in C/C++ or some scripting >> language. > > Do we have such a policy available? Policy section 10

Re: release policy changes

2005-06-09 Thread Andreas Barth
* Florian Weimer ([EMAIL PROTECTED]) [050609 15:16]: > * Andreas Barth: > > About: > > Shared libraries must be compiled with -fPIC, and normally static > > libraries must not be. If you need to provide static libraries > > compiled with -fPIC, call it "_pic.a". > > > > There was some

Re: release policy changes

2005-06-09 Thread Florian Weimer
* Andreas Barth: > One addition I would like very much to see is: > A library that is included in a package in Debian must be linked to > dynamically; for static-only executables like sash also static linking > to that other library package is accepted. Importing and using the > source code of

Re: release policy changes

2005-06-09 Thread Santiago Vila
On Thu, 9 Jun 2005, Jeroen van Wolffelaar wrote: > On Thu, Jun 09, 2005 at 12:35:17AM +0200, Matthias Klose wrote: > > one more thing: C++ dependencies are currently "hidden" within > > build-essential. Currently you cannot see this dependency from the > > outside, if the code is not dynamically

Re: release policy changes

2005-06-09 Thread Mark Brown
On Thu, Jun 09, 2005 at 12:09:26AM +0200, Andreas Barth wrote: > Rationale: > Some libraries are provided multiple in Debian, IIRC e.g. libz. That is > bad from general QA (as usually this is just an old version, and normale > bug fixes don't go in), and especially bad if there is a security updat

Re: release policy changes

2005-06-09 Thread Andreas Barth
* Thomas Bushnell BSG ([EMAIL PROTECTED]) [050609 00:38]: > Andreas Barth <[EMAIL PROTECTED]> writes: > > One addition I would like very much to see is: > > A library that is included in a package in Debian must be linked to > > dynamically; for static-only executables like sash also static lin

Re: release policy changes

2005-06-08 Thread Matthias Klose
Thomas Bushnell BSG writes: > Matthias Klose <[EMAIL PROTECTED]> writes: > > > one more thing: C++ dependencies are currently "hidden" within > > build-essential. Currently you cannot see this dependency from the > > outside, if the code is not dynamically linked to libstdc++ (or > > doesn't use

Re: release policy changes

2005-06-08 Thread Thomas Bushnell BSG
Joerg Jaspert <[EMAIL PROTECTED]> writes: > #311724 > > cdbs is an example for this madness (B-D) here, maybe others do it too. Oh, blech. Yes, I agree this should be prohibited. Never occurred to me to do something like that. I think a suitable substitute would be, instead of generating new B

Re: release policy changes

2005-06-08 Thread Joerg Jaspert
On 10315 March 1977, Thomas Bushnell wrote: >> Rationale: >> Well, the contents of both are decisions of the maintainer. It is IMHO >> very bad if a package starts to change build dependencies during a NMU >> or an security upload, and even worse if the maintainer is "adjusted" in >> the binary pa

Re: release policy changes

2005-06-08 Thread Thomas Bushnell BSG
Matthias Klose <[EMAIL PROTECTED]> writes: > How can you tell, for which C++ ABI packages like mozilla-dev and > festival-dev are built? IMO that is the technical reason you are > asking for. It doesn't matter if an application or a library is linked > to libstdc++. Remember that C++ ABI != libstd

Re: release policy changes

2005-06-08 Thread Jeroen van Wolffelaar
On Thu, Jun 09, 2005 at 12:35:17AM +0200, Matthias Klose wrote: > one more thing: C++ dependencies are currently "hidden" within > build-essential. Currently you cannot see this dependency from the > outside, if the code is not dynamically linked to libstdc++ (or > doesn't use libstdc++ at all).

Re: release policy changes

2005-06-08 Thread Thomas Bushnell BSG
Matthias Klose <[EMAIL PROTECTED]> writes: > one more thing: C++ dependencies are currently "hidden" within > build-essential. Currently you cannot see this dependency from the > outside, if the code is not dynamically linked to libstdc++ (or > doesn't use libstdc++ at all). I'm proposing to req

Re: release policy changes

2005-06-08 Thread Matthias Klose
Andreas Barth writes: > One addition I would like very much to see is: > A library that is included in a package in Debian must be linked to > dynamically; for static-only executables like sash also static linking > to that other library package is accepted. Importing and using the > source cod

Re: release policy changes

2005-06-08 Thread Thomas Bushnell BSG
Andreas Barth <[EMAIL PROTECTED]> writes: > One addition I would like very much to see is: > A library that is included in a package in Debian must be linked to > dynamically; for static-only executables like sash also static linking > to that other library package is accepted. Importing and us

release policy changes

2005-06-08 Thread Andreas Barth
Hi, I looked through our release policy. There are some changes I think we should or need to make. One addition I would like very much to see is: A library that is included in a package in Debian must be linked to dynamically; for static-only executables like sash also static linking to that