NetBSD: Statusreport für NetBSD 2.0

[asg]

James Chacon har auf "netbsd-announce@" eine Zusammenfassung geschrieben wie weit das NetBSD Team noch von einem 2.0 RELEASE entfernt ist.

Hier seine Email
(auch zu lesen unter http://mail-index.netbsd.org/netbsd-...7/26/0000.html):

Below is a summary of where we're at on the 2.0 release. As you'll notice,
according to this schedule we would be in RC1, but due to the number of
outstanding bugs we're not quite there yet.

So, rather than reformulate another schedule I wanted to publish this to
give folks an idea of where we're at and once RC1 tags approx how long it'll
take before the release is done. I personally expect RC1 to tag within
the next 2 weeks and we'll publish that here when it happens as well.

Also attached are the list of critical PR's from testing that have been
noted as critical. IPF 4.1.3 still remains to be merged into the branch at
which point the remaining bugs will hopefully be fixed (or easy to fix).

------

Original Branch date: Mar 28 2004

Current schedule: - Slipped from original. Re-re-worked one is below:

*Mar 29nd - July 17th - Beta period - Testing feedback and pullups
July 18th - RC1 tag - Only pullups past this point are for showstoppers. Any
pullups will cause a new RC cycle so be aware of this
when asking for something.
**July 19th - Aug 2nd - RC1 testing
Aug 3rd - RC2 tag or final release tag - Assuming nothing was wrong w. RC1
If RC2 then final release pushes out
2 weeks, etc.
Aug 17th - Announce release after ftp uploaded, iso's created and mirrors
synced.

*The decision to stop beta and go to RC1 is based on the overall quality of
the release, not a specific timeframe. 4 weeks was a rough estimate on how
long this was going to take originally. Based on bugs this was extended
multiple times.

**The 2 weeks listed here assumes an RC candidate that has no issues. After 2
weeks of feedback/testing if it's good we'll go to the next stage. If there
are any issues a new RC will be cut and the cycle started over for it.

Pullups processed: 690 since original branch,
80 in the last 3 weeks.
12 currently open

Bugs/PR's specific to 2.0: (that we've been informed of)

Current bugs: 21 (not closed) 26 (closed)

24969: /usr/sbin/ipfs from ipfilter 4.1.1 does not work
24987: Weird NFS permissions problem -- aka 23716
25071: libXft cores when fonts.cache-1 timestamp old
25087: ipfilter 4.1 NAT seems to cause fragmentation problems
25185: netbsd-2-0 build (with -x) breaks on fc-cache without $HOME (1)
25223: ftp proxy in ipnat causes panic
25227: panic: m_copydata - in ipf logging during initial ifconfig
25526: standard ipsec sysctl infrastructure not coherent after recent
fastipsec changes
25699: sh(1) hangs opening a named pipe as stdin for background process
25761: kill -HUP <pid-of-ipmon> freezes system ~weekly
25790: ipnat cannot load rules on sparc64
25810: ipf4 ftp proxy has problems with long 221- lines
25844: NAT router with current and 2.0_BETA suddenly reset connection
25909: INSTALL notes need updating before 2.0 release
25918: XF4.4 ATI regression
25987: ipf not config source compatible from 1.6 to 2.0
25992: some protocol names in ipf.conf don't work
25993: ipf parser accepts invalid flags in rules
26121: Sun keyboard does not work on Sun Ultra 60
26160: objcopy creates unbootable kernel sometimes
26328: 2.0_BETA kernel crash

1 - Not critical before the release but it should get fixed.

Closed bugs:

22452: programs broken b/c libgcc1 functions are in libc, not libgcc
23486: -current gdb no longer works with emacs
24961: ipfilter 4.1.1 has problems parsing ipf.conf
24981: ipfilter in 2.0 branch panics the system (in feedback from patch)
24989: ipfilter 4.1.1 does not behave according to rules in ipf.conf
25094: gcc generates wrong debugging symbols per default
25103: IP Filter 4.4.1 breaks some connections when NATing.
25113: fdisk(8) unusable for boot selector
25138: sysinst upgrade broken: fsck_ffs bails out
25146: bootselector botched after sysinst upgrade
25207: tar cores on certain archive contents
25226: serious compat problem when mount_msdos calls mount(2)
25235: improper boot block installation on 2.0beta
25608: 2.0: ipf still crashes system
25773: ipnat rdr rules not properly matched against ipf.conf rules
25868: Bringing up an interface freezes the system when IPFilter is enabled
25897: src/regress/lib/libpthread/sem fails
25898: src/regress/lib/librt/sem fails
25904: the MAXBSIZE->MAXPHYS changes broke usb printing
25942: mips doesn't work (cache problems)
25986: ipf not binary compatible from 1.6 to 2.0
25988: ipf reports wrong line number in error messages
25989: ipf error messages misleading
25991: ipnat.conf rules don't allow port/protocol names
25999: ipnat is corrupting "bimap" translations in 2.0_BETA and -current
26145: "zcat -f" misbehaves

[CW]

Wenn ich mir das Ganze genauer durchlese, dann sieht es wohl so aus, als ob es nicht so viele Probleme mit dem System (Kernel) gibt, sondern viel mehr mit dem alten (und ausgelutschten) ipf. Aber zum Glück ist ja noch itojun bei NetBSD dabei.

Er hat ja den pf importiert.

Und den Mailing-Listen zufolge ist Darren Reed (der Entwickler von ipf) mächtig sauer darüber (was er ja nicht direkt zugibt ... aber wenn man zwischen den Zeilen liest *he he*)

[asg]

Komisch diese Schadenfreude der OBSD User gegenüber NetBSD und IPF. Ist mir nun schon mehrfach untergekommen....

[CW]

Zitat von asg

Komisch diese Schadenfreude der OBSD User gegenüber NetBSD und IPF. Ist mir nun schon mehrfach untergekommen....

Das ist nicht ganz richtig.

Nix gegen NetBSD (schon gar nicht aus meinem Munde), nur was gegen ipf (technisch gesehen ist er veraltet ... oder steht hier jemand auf Konfigurationsfiles ohne Makros und die ganze group & head-Sch...e? ).

Außerdem war die damalige Diskussion zwischen Darren Reed und der OBSD-Crew alles andere als gesund.

Jedoch brachte sie uns allen (auch euch FBSD-lern) den geilen PF

NetBSD sollte auf jeden Fall pf erhalten, WEIL es ein ausgezeichnetes System ist.