Hello Frans, thank you for your detailed look at the summary and at the ppc64 port page.
On 06-Jun-19 11:07, Frans Pop wrote: > On Monday 19 June 2006 08:29, Andreas Jochens wrote: > > Please give me hint if I should file the missing patches now > > or if I should wait for a decision at the d-i meeting. > > Thank you for the overview. Please wait a bit before filing any new bugs > as we should first revert the decision made earlier. > > I've taken a look at your patch directory and AFAICT there are two other > bugs open that will prevent you from building d-i "from the archive": > busybox and gnupg. Fortunately the busybox issue has already been fixed in unstable. The patch that you found in the patch directory is no longer necessary because the issue has been fixed in a different way by Bastian Blank in busybox version 1.1.3-1. gnupg still needs the '--disable-asm' patch (BTS #343434) to build on ppc64. > I will put your request on the agenda for the next meeting and will let > you know afterwards. OK, thanks. > There is still the structural question of "why add support for an arch > that basically nobody in the project sees any future for" with some even > actively saying it's a suboptimal solution. I know that Sven Luther has voiced his negative opinion about the native ppc64 port on various occasions. Also Bastian Blank seems to be at least very sceptical about a native ppc64 port. However, all developers were generally very helpful with the native ppc64 port. More than 100 patches have already been applied to the archive and the native ppc64 port is working quite well now. The native ppc64 port is already being used for some serious things, e.g. by Matthias Klose and other upstream python developers for the development and testing of the new python version 2.5 (see http://www.python.org/dev/buildbot/trunk/PPC64%20Debian%20trunk). At the moment, the native ppc64 port is the only solution which provides a full native 64-bit environment on ppc64. It already works and it needs minimal changes to the existing Debian archive. I did not yet see any other working solution which provides a reasonably complete Debian 64-bit environment on ppc64. Up to now I have just heard rough ideas like 'biarch' or 'multiarch' which do not yet have working implementations and which would need massive changes to the existing Debian archive. Moreover, even for multiarch it would be the easiest solution to have a native ppc64 port which provides the 64-bit binary packages because with that setup multiarch could be implemented for ppc64/powerpc in exactly the same way as for amd64/i386. > The two patches that need most to be discussed are: > - debian-installer: doing this in a branch instead could be an option I do not expect frequent ppc64 specific changes for debian-installer. It would really help to have the ppc64 specific config and pkg-lists files in the official debian-installer sources. > - linux-kernel-di-powerpc-2.6: looks like it would work, but also fairly > invasive for a package that already has a huge control file The usual way would probably be to have a separate linux-kernel-di-ppc64-2.6 package. This would of course also be possible, but it would lead to a duplication of code which could be avoided by using the existing linux-kernel-di-powerpc-2.6 package with a small patch. Anyway, thanks again for looking at the ppc64 related issues. Regards Andreas Jochens -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]