|Deletions are marked like this.||Additions are marked like this.|
|Line 59:||Line 59:|
| ''bugs like Debian 327780 crop up regularly anyway though''
-- there are bugs on other archs as well, such as http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=210809 (ppc) or http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=223767 (i386) and there's no criticism on the binutils for those ports on their wiki pages. So why here? ''Higher severity, affecting more packages, and no evidence for a long time that the bugs were being worked on. However, recently upstream seems to actually be getting such bugs fixed, which certainly qualifies as upstream support.'' The requirement was to have upstream binutils support, which m68k has, not if this upstream support produces bugfree software. (ij)
-> so can we remove these comments now?
The purpose of this page is to demonstrate that m68k meets the [http://release.debian.org/etch_arch_criteria.html architecture recertification criteria for etch].
If some of these requirements are irrelevant or implausible for the port, please add a waiver request indicating why and how this isn't a problem for the arch in bold.
The architecture is publicly available without NDAs via:
The following machines are available to developers:
crest.debian.org - maintained by DSA, connectivity via 10Mbit Ethernet.
kullervo.debian.org - maintained by DSA, connectivity via 10Mbit Ethernet -- kullervo was not available for aba on 2005-10-29
The Debian port is maintained by the following developers, who are familiar with arch-specific issues for this port
Wouter Verhelst <email@example.com>
Stephen R. Marenka <firstname.lastname@example.org>
Christian T. Steigies <email@example.com>
Adam Conrad <firstname.lastname@example.org>
Michael Schmitz <email@example.com>
The port is being actively used at the following sites:
Please see http://www.buildd.net/cgi/archvote.phtml -- voluntary registration of some users since the requirement was made public. Lists 51 users as of 2005/10/10, after the page had been somewhat advertised for about a week. 95 users are listed on 2005/11/18.
Also, http://www.linux-m68k.org/Registry/Registration.html -- voluntary registration of many Linux/m68k users since the m68k Linux port was started (some of those may not be active anymore).
Furthermore, http://lists.debian.org/stats/ lists 258 users as of Sunday October 16 2005 -- 18:54 +02:00.
Popcon indicates 5 users of the architecture as at 2005/11/18.
The installer is being maintained by Stephen Marenka and Wouter Verhelst and the Sarge version is currently working effectively (the Etch version needs some more work, see ?DebianInstallerM68kTodo).
Upstream support is provided by:
glibc: Andreas Schwab <firstname.lastname@example.org>
gcc: Andreas Schwab <email@example.com>
kernel: firstname.lastname@example.org -- central mailinglist for m68k kernel development; rather active, with ~5-10 regular posters and (on average) a few posts each day.
boot loader: see Bootloaders on http://wiki.debian.org/DebianInstallerM68kTodo
binutils: Ben Elliston <email@example.com>
Archive coverage and Autobuilder support
The port is currently at 94.5% (2005/11/20) up to date according to [http://buildd.debian.org/stats/graph2-week.png the buildd stats] (95% cut off)
The port has currently built 96.5% (2005/11/20) of the archive according to [http://buildd.debian.org/stats/graph-week.png the buildd stats] (90% cut off)
The port builds from unmodified Debian source.
For an overview of current autobuilders, please see http://unstable.buildd.net/index-m68k.html. 15 active machines as of this writing. Quickstep might be considered as redundancy. It usually builds experimental, but can do unstable, too.
Also note that the release team has agreed on grandfathering in the m68k port wrt this requirement.
The security team have noted the following problems in supporting the architecture:
The Debian admin team have noted the following problems in supporting the architecture:
The release team have noted the following problems in supporting the architecture:
- Not keeping up ATM.
- -- backlog cleared since ~2 weeks (as of 2005/11/18) (ij)