Blockers in Bugfix-Releases (Re: [Release Planning 3.6] Samba 3.6.6 on May 31 (was May 24)?)

[Karolin Seeger]

On Sun, Jun 10, 2012 at 10:35:44AM +1000, Andrew Bartlett wrote:
> I'm not against shipping stable releases, and I am for constantly
> improving the software we ship.  But for me that means that if a release
> is an improvement, then we should make it.

In general I agree, but I think there are certain key features that should
work.
 
> Developers should be tapped on the shoulder, reminded about bugs etc,
> but that should be it.  A release should not be delayed unless it would
> contain a regression from the previous point release (see below on
> regressions)
> 
> > > If an issue is severe, then it should deserve attention on it's own, and
> > > missing the release entirely is our penalty. 
> > > 
> > > > Of course a release should not be delayed due to minor issues. If that
> > > > ever happened, please let me know.
> > > 
> > > As per our other discussion, I think we have terrible trouble defining
> > > minor/major.  For each of us, particularly after spending days chasing
> > > difficult issue, our particular bug is usually a major issue.
> > 
> > I am trying to ask the developers "hey, what exactly does not work for
> > users without your patch(es), use children's language". From my point of view,
> > that's one of the advantages when the release manager is not a developer.
> > If basic features do not work, it's a major issue. If e.g.a VFS module has a
> > problem, it's certainly not a major issue.
> 
> I don't think we should use the language of 'major issue'.  It isn't the
> right test.  We should say "hey, what exactly does not work for users
> without your patch(es), compared with the previous point release, use
> children's language"
> 
> > > I would much rather use 'regression' as the definition - that is much
> > > easier to pin down. 
> > 
> > My problem with the 'regression' is the following. A regression since
> > when? 3.0? If in 3.6.6 joining of XP clients does not work, it's a
> > regression compared to 3.0.x, but if it has never worked since 3.6.0, it's
> > not a regression compared to 3.6.0, but it's definately a major issue from my
> > point of view 
> > and I would try to grab one developer to get that fixed before the next
> > bugfix release.
> 
> This is actually a prefect example.  The issue is serious, but by
> blocking a release waiting for a fix (or confirmation of a fix) for this
> issue, we have denied users access to a stable release for non-DC
> environments (or environments with the 'right' length hostnames - it was
> an odd/even issue).

That's not true. The delay was caused by the possible security issue which
turned out to be not so bad later on. Of course, I couldn't write that on
this list publicly. Catching the joining issue was a nice side effect.
And I am very glad that this one is fixed as it kept users away from 3.6!
 
> This is important, particularly as we move to Samba 4.0 with a much
> larger codebase.  The failure to join to an S3 DC has been with us since
> before 3.6.0, so a release without that fix would by definition be no
> worse than 3.6.0 or any subsequent point release.  Therefore we should
> make the release. 
> 
> The same *will* happen with Samba 4.0.  The new AD DC will not be
> released perfect - no software ever is - but if we apply this test, then
> bugs in the AD DC (and it's complex code, there is a good chance of some
> quite serious issues despite our good results so far) will hold up file
> server users from getting releases.  This would not be a good thing!
> 
> So far, due to these delays, our 3.6 users are still waiting for fixes
> for:
> 
> +o   Michael Adam <obnox at samba.org>
> +    * BUG 8738: SMB2 server will not release unused shares.
> +    * BUG 8749: Sign non guest sessions in SessionSetup.
> +    * BUG 8921: Fix race writing registry values.
> +
>  
>  o   Jeremy Allison <jra at samba.org>
> +    * BUG 8723: Add pthread-based aio VFS module.
> +    * BUG 8784: When calculating the share security mask, take
> priviliges into
> +      account for the connecting user.
> +    * BUG 8837: Fix crash in smbd when deleting directory and veto
> files are
> +      enabled.
> +    * BUG 8857: Setting traverse rights fails to enable directory
> traversal when
> +      acl_xattr in use.
> +    * BUG 8897: Make winbind_krb5_locator not only returning one IP
> address.
> +
> +
> +o   Christian Ambach <ambi at samba.org>
> +    * BUG 8406: Fix a return code check in Winbind.
> +    * BUG 8807: Fix crash in dcerpc_lsa_lookup_sids_noalloc() crashes
> when
> +      groups has more than 1000 groups.
> +
> +
> +o   Andrew Bartlett <abartlet at samba.org>
> +    * BUG 8599: Only use SamLogonEx when we can get unencrypted session
> keys.
> +    * BUG 8727: Fix smbclients with posix large reads.
> +
> +
> +o   Björn Baumbach <bb at sernet.de>
> +    * BUG 7564: Fix default name resolve order in the manpage.
> +    * BUG 8554, 8612, 8748: Add new printers to registry.
> +    * BUG 8789: Remove whitespace in example samba.ldif.
> +
> +
> +o   Alejandro Escanero Blanco <aescanero at gmail.com>
> +    * BUG 8798: The primary rid should be in the groups rid array.
> +
> +
> +o   Ira Cooper <samba at ira.wakeful.net>
> +    * BUG 8729: Fix getpass regressions on Solaris/Illumos.
> +    * BUG 8743: Fix configure.developer builds on Solaris.
> +
> +
> +o   David Disseldorp <ddiss at samba.org>
> +    * BUG 8762: Fix crash in printer_list_set_printer().
> +
> +
> +o   Olaf Flebbe <o.flebbe at science-computing.de>
> +    * BUG 8859: Fix assertion in reg_parse.
> +
> +
> +o   Björn Jacke <bj at sernet.de>
> +    * BUG 8869: Remove outdated netscape ds 5 schema file.
> +
> +
> +o   Steve Langasek <steve.langasek at ubuntu.com>
> +    * BUG 8920: Fix null dereference in pdb_interface.
> +
> +
> +o   Volker Lendecke <vl at samba.org>
> +    * BUG 8567: Fix segfault in dom_sid_compare.
> +    * BUG 8733: Delete streams on directories (streams_depot).
> +    * BUG 8760: Add SERVERID_UNIQUE_ID_NOT_TO_VERIFY.
> +    * BUG 8836: Fix segfaults on "smbcontrol close-share" in aio_fork.
> +    * BUG 8861: Fix a segfault with debug level 3 on Solaris.
> +    * BUG 8904: Fix Winbind crash triggered by 'wbinfo --lookup-sids
> ""'.
> +
> +
> +o   Stefan Metzmacher <metze at samba.org>
> +    * BUG 8139: Ignore SMBecho errors (the server may not support it).
> +    * BUG 8527: db_ctdb_traverse fails to traverse records created
> within the
> +      current transaction.
> +    * BUG 8739: Fill the sids array of the info in
> +      wbcAuthUserInfo_to_netr_SamInfo3().
> +    * BUG 8749: Sign non guest sessions in SessionSetup.
> +
> +
> +o   Matthieu Patou <mat at matws.net>
> +    * BUG 8599: Set the can_do_validation6 also for trusted domain.
> +    * BUG 8734: Don't try to do clever thing if the username is not
> found while
> +      authenticating through Winbind.
> +    * BUG 8771: Winbind takes up to 20 minutes to change from DC 1 to
> DC 2.
> +
> +
> +o   SATOH Fumiyasu <fumiyas at osstech.co.jp>
> +    * BUG 8826: Prepend '/' to filename argument (docs).
> +
> +
> +o   Richard Sharpe <realrichardsharpe at gmail.com>
> +    * BUG 8768: Honor SeTakeOwnershipPrivilege when file opened with
> SEC_STD_WRITE_OWNER.
> +    * BUG 8797: Correctly handle DENY ACEs when privileges apply.
> +    * BUG 8822: Fix building out-of-tree modules.
> +
> +
> +o   Simo Sorce <idra at samba.org>
> +    * BUG 8915: Fix pam_winbind build against newer iniparser library.
> +
> +
> +o   Joseph Tam <jtam.home at gmail.com>
> +    * BUG 8877: Syslog broken owing to mistyping of
> debug_settings.syslog.
> +
> +
> +o   Ralph Wuerthner <ralph.wuerthner at de.ibm.com>
> +    * BUG 8845: Move print_backend_init() behind init_system_info().

To be completed. And I know this list very well as I created it.
 
> Should we deny our users all these fixes, because we want to get in one
> more fix?  Instead of saying 'these serious issues remain unresolved' in
> a WHATSNEW of a release fixing all these other things, we are saying
> 'these serious issues remain unresolved, so you can't have a release at
> all!'

No. Again, it's a special case this time. Due to the 3 (almost 4) security
releases, the list of fixed bugs is longer than usual, but it's still very
short compared with older 3.0 releases!
 
> Beyond individual users (who could go to GIT if they really wanted)
> there are other costs to constantly slipping.  For distributions, in the
> thread on changing the stable branch rules Christian PERRIER
> <bubulle at debian.org> said:
> > 
> > Next Debian stable should be frozen "Quite Soon". We hope we'll be
> > able
> > to get 3.6.6 in before that (which is why the delay in that release
> > saddens me, though I understand the reasons).

Sticking to the last planned release date would have meant to ship a
bugfix release with a known security issue (which it was to me at that
point). That would have been inacceptable from my point of view.
 
> So, particularly if we keep slipping, the risk is that we deny all these
> fixes to Debian stable users (they won't backport this many patches),
> because of a serious issue hitting only the DC case that could actually
> be backported.  

I am sorry, but it just did not work out this time.
Hopefully I coul clarify the circumstances a bit.
 
> So, what I'm saying is that unless a release would be measurably worse
> than the minor release which it replaces, it should be shipped
> regardless.  We can then follow up with an ever better release next time
> - quite soon if the fix warrants it.   

Karolin

-- 
Samba			http://www.samba.org
SerNet			http://www.sernet.de
sambaXP			http://www.sambaxp.org